AuroraMiddleware/gtk2
1
0
Fork 0
forked from AuroraMiddleware/gtk
Code Issues Pull Requests Projects Releases Wiki Activity
9fa1b4f65f
gtk2/gtk/gtkcssprovider.c

2944 lines
91 KiB
C
Raw Normal View History

Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
/* GTK - The GIMP Toolkit
* Copyright (C) 2010 Carlos Garnacho <carlosg@gnome.org>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
Change FSF Address
2012-02-27 13:01:10 +00:00
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
*/
#include "config.h"
#include <string.h>
#include <stdlib.h>
Split out private style context api into a private header
2011-01-23 23:29:07 +00:00
GtkCssProvider: Add a parser for 9slice typed properties. The value it parses is similar to the border-image CSS3 property, so strings like this will be accepted: border-image: url (foo.png) 4 3 4 3 repeat repeat; the image path is relative to the parsed CSS file dirname if no absolute path is provided.
2010-09-24 20:05:20 +00:00
#include <gdk-pixbuf/gdk-pixbuf.h>
Use the correct cairo-gobject.h include path.
2010-11-12 20:43:13 +00:00
#include <cairo-gobject.h>
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
Privately export css_provider_get_theme_dir
2011-01-26 16:17:34 +00:00
#include "gtkcssproviderprivate.h"
cssprovider: Keep a bitmask for the properties That way we can easily tell wich properties are in use.
2011-12-23 13:23:52 +00:00
#include "gtkbitmaskprivate.h"
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
#include "gtkcssarrayvalueprivate.h"
styleproperty: require property in _gtk_style_property_parse_value() Other code uses _gtk_css_style_parse_value() instead now.
2012-01-01 23:14:23 +00:00
#include "gtkcssstylefuncsprivate.h"
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
#include "gtkcssparserprivate.h"
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
#include "gtkcsssectionprivate.h"
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
#include "gtkcssselectorprivate.h"
styleproperty: Get rid of _gtk_style_property_is_shorthand() Use GTK_IS_CSS_SHORTHAND_PROPERTY() instead now that we have it.
2011-12-31 12:50:03 +00:00
#include "gtkcssshorthandpropertyprivate.h"
css: Overhaul value parsing Value parsing only sometimes emitted errors. Sometimes it didn't emit errors but ignored the value, sometimes it took a default, sometimes it converted it to something it deemed suitable. While refactoring, I moved the whole GValue <=> char * conversion routines to a separate file, to make navigating the core css provider easier.
2011-04-08 14:08:28 +00:00
#include "gtksymboliccolor.h"
Split out private style context api into a private header
2011-01-23 23:29:07 +00:00
#include "gtkstyleprovider.h"
#include "gtkstylecontextprivate.h"
css: Refactor code to do property lookups earlier We want to ook up the property in the CSS parser, so we can do fancy things with it. We currently don't but we want to later.
2011-05-16 21:37:29 +00:00
#include "gtkstylepropertiesprivate.h"
css: Rename gtkcssstrfuncs file to gtkstyleproperty The idea is to handle the style property registering etc in this file instead of the styleproperties file. The next commits will move those files.
2011-05-21 01:30:36 +00:00
#include "gtkstylepropertyprivate.h"
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
#include "gtkstyleproviderprivate.h"
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
#include "gtkbindings.h"
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
#include "gtkmarshalers.h"
win32: include gtkprivate.h for win32 GTK_DATA_PREFIX
2011-01-02 11:11:34 +00:00
#include "gtkprivate.h"
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
#include "gtkintl.h"
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* SECTION:gtkcssprovider
* @Short_description: CSS-like styling for widgets
* @Title: GtkCssProvider
* @See_also: #GtkStyleContext, #GtkStyleProvider
*
Some header cleanups Move GtkGradient to their own files, also move GtkBorder out of gtkstyle.h, so that header can be all deprecated.
2010-11-26 01:39:53 +00:00
* GtkCssProvider is an object implementing the #GtkStyleProvider interface.
* It is able to parse <ulink url="http://www.w3.org/TR/CSS2">CSS</ulink>-like
* input in order to style widgets.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
*
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <refsect2 id="gtkcssprovider-files">
* <title>Default files</title>
* <para>
* An application can cause GTK+ to parse a specific CSS style sheet by
* calling gtk_css_provider_load_from_file() and adding the provider with
* gtk_style_context_add_provider() or gtk_style_context_add_provider_for_screen().
* In addition, certain files will be read when GTK+ is initialized. First,
Remove reference to GTK2_RC_FILES from the docs Pointed out in bug 637471.
2010-12-20 15:19:55 +00:00
* the file <filename><envar>$XDG_CONFIG_HOME</envar>/gtk-3.0/gtk.css</filename>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* is loaded if it exists. Then, GTK+ tries to load
Remove reference to GTK2_RC_FILES from the docs Pointed out in bug 637471.
2010-12-20 15:19:55 +00:00
* <filename><envar>$HOME</envar>/.themes/<replaceable>theme-name</replaceable>/gtk-3.0/gtk.css</filename>,
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* falling back to
Remove reference to GTK2_RC_FILES from the docs Pointed out in bug 637471.
2010-12-20 15:19:55 +00:00
* <filename><replaceable>datadir</replaceable>/share/themes/<replaceable>theme-name</replaceable>/gtk-3.0/gtk.css</filename>,
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* where <replaceable>theme-name</replaceable> is the name of the current theme
Fix a tag mismatch in the docs
2010-12-21 23:15:43 +00:00
* (see the #GtkSettings:gtk-theme-name setting) and <replaceable>datadir</replaceable>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* is the prefix configured when GTK+ was compiled, unless overridden by the
* <envar>GTK_DATA_PREFIX</envar> environment variable.
* </para>
* </refsect2>
* <refsect2 id="gtkcssprovider-stylesheets">
* <title>Style sheets</title>
* <para>
* The basic structure of the style sheets understood by this provider is
* a series of statements, which are either rule sets or '@-rules', separated
* by whitespace.
* </para>
* <para>
* A rule set consists of a selector and a declaration block, which is
* a series of declarations enclosed in curly braces ({ and }). The
* declarations are separated by semicolons (;). Multiple selectors can
* share the same declaration block, by putting all the separators in
* front of the block, separated by commas.
* </para>
* <example><title>A rule set with two selectors</title>
* <programlisting language="text">
* GtkButton, GtkEntry {
* color: &num;ff00ea;
* font: Comic Sans 12
* }
* </programlisting>
* </example>
* </refsect2>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <refsect2 id="gtkcssprovider-selectors">
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <title>Selectors</title>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* Selectors work very similar to the way they do in CSS, with widget class
* names taking the role of element names, and widget names taking the role
* of IDs. When used in a selector, widget names must be prefixed with a
* '&num;' character. The '*' character represents the so-called universal
* selector, which matches any widget.
* </para>
* <para>
* To express more complicated situations, selectors can be combined in
* various ways:
* <itemizedlist>
* <listitem><para>To require that a widget satisfies several conditions,
* combine several selectors into one by concatenating them. E.g.
* <literal>GtkButton&num;button1</literal> matches a GtkButton widget
* with the name button1.</para></listitem>
* <listitem><para>To only match a widget when it occurs inside some other
* widget, write the two selectors after each other, separated by whitespace.
* E.g. <literal>GtkToolBar GtkButton</literal> matches GtkButton widgets
* that occur inside a GtkToolBar.</para></listitem>
* <listitem><para>In the previous example, the GtkButton is matched even
* if it occurs deeply nested inside the toolbar. To restrict the match
* to direct children of the parent widget, insert a '>' character between
* the two selectors. E.g. <literal>GtkNotebook > GtkLabel</literal> matches
* GtkLabel widgets that are direct children of a GtkNotebook.</para></listitem>
* </itemizedlist>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <example>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <title>Widget classes and names in selectors</title>
* <programlisting language="text">
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* /&ast; Theme labels that are descendants of a window &ast;/
* GtkWindow GtkLabel {
A little closer to css Declarations are ;-separated, but there is no ; required after the last declaration. We harmlessly accept either.
2010-11-18 23:36:08 +00:00
* background-color: &num;898989
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme notebooks, and anything that's within these &ast;/
* GtkNotebook {
A little closer to css Declarations are ;-separated, but there is no ; required after the last declaration. We harmlessly accept either.
2010-11-18 23:36:08 +00:00
* background-color: &num;a939f0
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme combo boxes, and entries that
* are direct children of a notebook &ast;/
* GtkComboBox,
* GtkNotebook > GtkEntry {
A little closer to css Declarations are ;-separated, but there is no ; required after the last declaration. We harmlessly accept either.
2010-11-18 23:36:08 +00:00
* color: @fg_color;
* background-color: &num;1209a2
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme any widget within a GtkBin &ast;/
* GtkBin * {
docs: There is no font-name style property It's "font"
2011-06-01 23:55:45 +00:00
* font: Sans 20
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
*
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* /&ast; Theme a label named title-label &ast;/
* GtkLabel&num;title-label {
docs: There is no font-name style property It's "font"
2011-06-01 23:55:45 +00:00
* font: Sans 15
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme any widget named main-entry &ast;/
* &num;main-entry {
A little closer to css Declarations are ;-separated, but there is no ; required after the last declaration. We harmlessly accept either.
2010-11-18 23:36:08 +00:00
* background-color: &num;f0a810
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
* </programlisting>
* </example>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* Widgets may also define style classes, which can be used for matching.
* When used in a selector, style classes must be prefixed with a '.'
* character.
* </para>
* <para>
* Refer to the documentation of individual widgets to learn which
Avoid compiler warnings
2010-11-26 04:46:15 +00:00
* style classes they define and see <xref linkend="gtkstylecontext-classes"/>
* for a list of all style classes used by GTK+ widgets.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
Mention type vs region ambiguity in the docs
2011-01-19 17:55:48 +00:00
* <para>
* Note that there is some ambiguity in the selector syntax when it comes
* to differentiation widget class names from regions. GTK+ currently treats
* a string as a widget class name if it contains any uppercase characters
* (which should work for more widgets with names like GtkLabel).
* </para>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <example>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <title>Style classes in selectors</title>
* <programlisting language="text">
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* /&ast; Theme all widgets defining the class entry &ast;/
* .entry {
Rely only on fg/bg colors. "foreground-color" has been renamed to just "color" as in regular CSS, and the "text-color" and "base-color" properties have been removed. The default CSS has been changed to have widgets cope with this change.
2010-10-29 03:33:38 +00:00
* color: &num;39f1f9;
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme spinbuttons' entry &ast;/
* GtkSpinButton.entry {
A little closer to css Declarations are ;-separated, but there is no ; required after the last declaration. We harmlessly accept either.
2010-11-18 23:36:08 +00:00
* color: &num;900185
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
* </programlisting>
* </example>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* In complicated widgets like e.g. a GtkNotebook, it may be desirable
* to style different parts of the widget differently. To make this
* possible, container widgets may define regions, whose names
* may be used for matching in selectors.
* </para>
* <para>
* Some containers allow to further differentiate between regions by
* applying so-called pseudo-classes to the region. For example, the
* tab region in GtkNotebook allows to single out the first or last
* tab by using the :first-child or :last-child pseudo-class.
* When used in selectors, pseudo-classes must be prefixed with a
* ':' character.
* </para>
* <para>
* Refer to the documentation of individual widgets to learn which
Add list of classes and regions to the docs
2010-11-26 04:49:07 +00:00
* regions and pseudo-classes they define and see
* <xref linkend="gtkstylecontext-classes"/> for a list of all regions
* used by GTK+ widgets.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <example>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <title>Regions in selectors</title>
* <programlisting language="text">
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* /&ast; Theme any label within a notebook &ast;/
* GtkNotebook GtkLabel {
Rely only on fg/bg colors. "foreground-color" has been renamed to just "color" as in regular CSS, and the "text-color" and "base-color" properties have been removed. The default CSS has been changed to have widgets cope with this change.
2010-10-29 03:33:38 +00:00
* color: &num;f90192;
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme labels within notebook tabs &ast;/
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* GtkNotebook tab GtkLabel {
Rely only on fg/bg colors. "foreground-color" has been renamed to just "color" as in regular CSS, and the "text-color" and "base-color" properties have been removed. The default CSS has been changed to have widgets cope with this change.
2010-10-29 03:33:38 +00:00
* color: &num;703910;
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
*
* /&ast; Theme labels in the any first notebook
* tab, both selectors are equivalent &ast;/
* GtkNotebook tab:nth-child(first) GtkLabel,
* GtkNotebook tab:first-child GtkLabel {
Rely only on fg/bg colors. "foreground-color" has been renamed to just "color" as in regular CSS, and the "text-color" and "base-color" properties have been removed. The default CSS has been changed to have widgets cope with this change.
2010-10-29 03:33:38 +00:00
* color: &num;89d012;
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* }
* </programlisting>
* </example>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* Another use of pseudo-classes is to match widgets depending on their
* state. This is conceptually similar to the :hover, :active or :focus
* pseudo-classes in CSS. The available pseudo-classes for widget states
* are :active, :prelight (or :hover), :insensitive, :selected, :focused
* and :inconsistent.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <example>
* <title>Styling specific widget states</title>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <programlisting language="text">
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* /&ast; Theme active (pressed) buttons &ast;/
* GtkButton:active {
* background-color: &num;0274d9;
* }
*
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* /&ast; Theme buttons with the mouse pointer on it,
* both are equivalent &ast;/
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* GtkButton:hover,
* GtkButton:prelight {
* background-color: &num;3085a9;
* }
*
* /&ast; Theme insensitive widgets, both are equivalent &ast;/
* :insensitive,
* *:insensitive {
* background-color: &num;320a91;
* }
*
* /&ast; Theme selection colors in entries &ast;/
* GtkEntry:selected {
* background-color: &num;56f9a0;
* }
*
* /&ast; Theme focused labels &ast;/
* GtkLabel:focused {
* background-color: &num;b4940f;
* }
*
* /&ast; Theme inconsistent checkbuttons &ast;/
* GtkCheckButton:inconsistent {
* background-color: &num;20395a;
* }
* </programlisting>
* </example>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* Widget state pseudoclasses may only apply to the last element
* in a selector.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* To determine the effective style for a widget, all the matching rule
* sets are merged. As in CSS, rules apply by specificity, so the rules
* whose selectors more closely match a widget path will take precedence
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* over the others.
* </para>
* </refsect2>
* <refsect2 id="gtkcssprovider-rules">
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <title>&commat; Rules</title>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* GTK+'s CSS supports the &commat;import rule, in order to load another
* CSS style sheet in addition to the currently parsed one.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <example>
* <title>Using the &commat;import rule</title>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <programlisting language="text">
A little closer to css Declarations are ;-separated, but there is no ; required after the last declaration. We harmlessly accept either.
2010-11-18 23:36:08 +00:00
* &commat;import url ("path/to/common.css");
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </programlisting>
* </example>
Add a missing id to the docs
2011-01-31 12:37:36 +00:00
* <para id="css-binding-set">
Document @binding-set and the gtk-key-bindings css parser property
2011-01-28 09:25:17 +00:00
* In order to extend key bindings affecting different widgets, GTK+
* supports the &commat;binding-set rule to parse a set of bind/unbind
doc fixes
2011-01-31 04:34:52 +00:00
* directives, see #GtkBindingSet for the supported syntax. Note that
* the binding sets defined in this way must be associated with rule sets
* by setting the gtk-key-bindings style property.
* </para>
* <para>
* Customized key bindings are typically defined in a separate
* <filename>gtk-keys.css</filename> CSS file and GTK+ loads this file
* according to the current key theme, which is defined by the
* #GtkSettings:gtk-key-theme-name setting.
Document @binding-set and the gtk-key-bindings css parser property
2011-01-28 09:25:17 +00:00
* </para>
* <example>
* <title>Using the &commat;binding rule</title>
* <programlisting language="text">
* &commat;binding-set binding-set1 {
* bind "&lt;alt&gt;Left" { "move-cursor" (visual-positions, -3, 0) };
* unbind "End";
* };
*
Make documentation match behaviour
2011-01-28 23:08:57 +00:00
* &commat;binding-set binding-set2 {
Document @binding-set and the gtk-key-bindings css parser property
2011-01-28 09:25:17 +00:00
* bind "&lt;alt&gt;Right" { "move-cursor" (visual-positions, 3, 0) };
Make documentation match behaviour
2011-01-28 23:08:57 +00:00
* bind "&lt;alt&gt;KP_space" { "delete-from-cursor" (whitespace, 1)
doc fixes
2011-01-31 04:34:52 +00:00
* "insert-at-cursor" (" ") };
Document @binding-set and the gtk-key-bindings css parser property
2011-01-28 09:25:17 +00:00
* };
*
* GtkEntry {
Make documentation match behaviour
2011-01-28 23:08:57 +00:00
* gtk-key-bindings: binding-set1, binding-set2;
Document @binding-set and the gtk-key-bindings css parser property
2011-01-28 09:25:17 +00:00
* }
* </programlisting>
* </example>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* GTK+ also supports an additional &commat;define-color rule, in order
* to define a color name which may be used instead of color numeric
* representations. Also see the #GtkSettings:gtk-color-scheme setting
* for a way to override the values of these named colors.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <example>
* <title>Defining colors</title>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <programlisting language="text">
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* &commat;define-color bg_color &num;f9a039;
*
* &ast; {
* background-color: &commat;bg_color;
* }
* </programlisting>
* </example>
* </refsect2>
* <refsect2 id="gtkcssprovider-symbolic-colors">
* <title>Symbolic colors</title>
* <para>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* Besides being able to define color names, the CSS parser is also able
* to read different color expressions, which can also be nested, providing
* a rich language to define colors which are derived from a set of base
* colors.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </para>
* <example>
* <title>Using symbolic colors</title>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <programlisting language="text">
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* &commat;define-color entry-color shade (&commat;bg_color, 0.7);
*
* GtkEntry {
* background-color: @entry-color;
* }
*
* GtkEntry:focused {
* background-color: mix (&commat;entry-color,
* shade (&num;fff, 0.5),
* 0.8);
* }
* </programlisting>
* </example>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <para>
* The various ways to express colors in GTK+ CSS are:
* </para>
* <informaltable>
* <tgroup cols="3">
* <thead>
* <row>
* <entry>Syntax</entry>
* <entry>Explanation</entry>
* <entry>Examples</entry>
* </row>
* </thead>
* <tbody>
* <row>
* <entry>rgb(@r, @g, @b)</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>An opaque color; @r, @g, @b can be either integers between
* 0 and 255 or percentages</entry>
* <entry><literallayout>rgb(128, 10, 54)
* rgb(20%, 30%, 0%)</literallayout></entry>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* </row>
* <row>
* <entry>rgba(@r, @g, @b, @a)</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>A translucent color; @r, @g, @b are as in the previous row,
* @a is a floating point number between 0 and 1</entry>
* <entry><literallayout>rgba(255, 255, 0, 0.5)</literallayout></entry>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* </row>
* <row>
* <entry>&num;@xxyyzz</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>An opaque color; @xx, @yy, @zz are hexadecimal numbers
* specifying @r, @g, @b variants with between 1 and 4
* hexadecimal digits per component are allowed</entry>
* <entry><literallayout>&num;ff12ab
* &num;f0c</literallayout></entry>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* </row>
* <row>
* <entry>&commat;name</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>Reference to a color that has been defined with
* &commat;define-color
* </entry>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <entry>&commat;bg_color</entry>
* </row>
* <row>
* <entry>mix(@color1, @color2, @f)</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>A linear combination of @color1 and @color2. @f is a
* floating point number between 0 and 1.</entry>
* <entry><literallayout>mix(&num;ff1e0a, &commat;bg_color, 0.8)</literallayout></entry>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* </row>
* <row>
* <entry>shade(@color, @f)</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>A lighter or darker variant of @color. @f is a
* floating point number.
* </entry>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* <entry>shade(&commat;fg_color, 0.5)</entry>
* </row>
* <row>
* <entry>lighter(@color)</entry>
* <entry>A lighter variant of @color</entry>
* </row>
* <row>
* <entry>darker(@color)</entry>
* <entry>A darker variant of @color</entry>
* </row>
docs: Add documentation for alpha() in CSS
2012-02-10 22:55:18 +00:00
* <row>
* <entry>alpha(@color, @f)</entry>
* <entry>Modifies passed color's alpha by a factor @f. @f is a
* floating point number. @f < 1.0 results in a more transparent
* color while @f > 1.0 results in a more opaque color.
* </entry>
* <entry>alhpa(blue, 0.5)</entry>
* </row>
Documentation improvements This patch adds more detail to the explanation of CSS, and also adds documentation about which files are read. We also make :focus an alternative to :focused (since :focus is a standard CSS pseudo-class).
2010-11-20 05:22:29 +00:00
* </tbody>
* </tgroup>
* </informaltable>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </refsect2>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <refsect2 id="gtkcssprovider-gradients">
* <title>Gradients</title>
* <para>
* Linear or radial Gradients can be used as background images.
* </para>
* <para>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* A linear gradient along the line from (@start_x, @start_y) to
* (@end_x, @end_y) is specified using the syntax
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <literallayout>-gtk-gradient (linear,
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* @start_x @start_y, @end_x @end_y,
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* color-stop (@position, @color),
* ...)</literallayout>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* where @start_x and @end_x can be either a floating point number between
* 0 and 1 or one of the special values 'left', 'right' or 'center', @start_y
* and @end_y can be either a floating point number between 0 and 1 or one
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* of the special values 'top', 'bottom' or 'center', @position is a floating
* point number between 0 and 1 and @color is a color expression (see above).
* The color-stop can be repeated multiple times to add more than one color
* stop. 'from (@color)' and 'to (@color)' can be used as abbreviations for
* color stops with position 0 and 1, respectively.
* </para>
* <example>
* <title>A linear gradient</title>
* <inlinegraphic fileref="gradient1.png" format="PNG"/>
* <para>This gradient was specified with
* <literallayout>-gtk-gradient (linear,
* left top, right bottom,
* from(&commat;yellow), to(&commat;blue))</literallayout></para>
* </example>
* <example>
* <title>Another linear gradient</title>
* <inlinegraphic fileref="gradient2.png" format="PNG"/>
* <para>This gradient was specified with
* <literallayout>-gtk-gradient (linear,
* 0 0, 0 1,
* color-stop(0, &commat;yellow),
* color-stop(0.2, &commat;blue),
* color-stop(1, &num;0f0))</literallayout></para>
* </example>
* <para>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* A radial gradient along the two circles defined by (@start_x, @start_y,
* @start_radius) and (@end_x, @end_y, @end_radius) is specified using the
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* syntax
* <literallayout>-gtk-gradient (radial,
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* @start_x @start_y, @start_radius,
* @end_x @end_y, @end_radius,
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* color-stop (@position, @color),
* ...)</literallayout>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* where @start_radius and @end_radius are floating point numbers and
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* the other parameters are as before.
* </para>
* <example>
* <title>A radial gradient</title>
* <inlinegraphic fileref="gradient3.png" format="PNG"/>
* <para>This gradient was specified with
* <literallayout>-gtk-gradient (radial,
* center center, 0,
* center center, 1,
* from(&commat;yellow), to(&commat;green))</literallayout></para>
* </example>
* <example>
* <title>Another radial gradient</title>
* <inlinegraphic fileref="gradient4.png" format="PNG"/>
* <para>This gradient was specified with
Document border images better
2010-11-22 16:15:47 +00:00
* <literallayout>-gtk-gradient (radial,
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* 0.4 0.4, 0.1,
* 0.6 0.6, 0.7,
* color-stop (0, &num;f00),
* color-stop (0.1, &num;a0f),
* color-stop (0.2, &commat;yellow),
* color-stop (1, &commat;green))</literallayout></para>
* </example>
* </refsect2>
styleproperties: add a "text-shadow" style property And document its use in the CSS parser gtk-doc. https://bugzilla.gnome.org/show_bug.cgi?id=649314
2011-05-20 20:11:08 +00:00
* <refsect2 id="gtkcssprovider-shadows">
* <title>Text shadow</title>
* <para>
* A shadow list can be applied to text or symbolic icons, using the CSS3
* text-shadow syntax, as defined in
* <ulink url="http://www.w3.org/TR/css3-text/#text-shadow">the CSS3 specification</ulink>.
* </para>
* <para>
* A text shadow is specified using the syntax
* <literallayout>text-shadow: @horizontal_offset @vertical_offset [ @blur_radius ] @color</literallayout>
* The offset of the shadow is specified with the @horizontal_offset and @vertical_offset
* parameters. The optional blur radius is parsed, but it is currently not rendered by
* the GTK+ theming engine.
* </para>
* <para>
* To set multiple shadows on an element, you can specify a comma-separated list
* of shadow elements in the text-shadow property. Shadows are always rendered
* front-back, i.e. the first shadow specified is on top of the others. Shadows
* can thus overlay each other, but they can never overlay the text itself,
* which is always rendered on top of the shadow layer.
* </para>
* </refsect2>
cssprovider: add a doc paragraph about box-shadow
2011-06-07 16:57:08 +00:00
* <refsect2>
* <title>Box shadow</title>
* <para>
* Themes can apply shadows on framed elements using the CSS3 box-shadow syntax,
* as defined in
* <ulink url="http://www.w3.org/TR/css3-background/#the-box-shadow">the CSS3 specification</ulink>.
* </para>
* <para>
* A box shadow is specified using the syntax
* <literallayout>box-shadow: [ @inset ] @horizontal_offset @vertical_offset [ @blur_radius ] [ @spread ] @color</literallayout>
* A positive offset will draw a shadow that is offset to the right (down) of the box,
* a negative offset to the left (top). The optional spread parameter defines an additional
* distance to expand the shadow shape in all directions, by the specified radius.
* The optional blur radius parameter is parsed, but it is currently not rendered by
* the GTK+ theming engine.
* The inset parameter defines whether the drop shadow should be rendered inside or outside
* the box canvas. Only inset box-shadows are currently supported by the GTK+ theming engine,
* non-inset elements are currently ignored.
* </para>
* <para>
* To set multiple box-shadows on an element, you can specify a comma-separated list
* of shadow elements in the box-shadow property. Shadows are always rendered
* front-back, i.e. the first shadow specified is on top of the others, so they may
* overlap other boxes or other shadows.
* </para>
* </refsect2>
docs: expand the docs for border-image https://bugzilla.gnome.org/show_bug.cgi?id=651194
2011-05-31 13:32:00 +00:00
* <refsect2 id="gtkcssprovider-border-image">
Tweak wording
2010-11-22 16:18:27 +00:00
* <title>Border images</title>
Document border images better
2010-11-22 16:15:47 +00:00
* <para>
docs: expand the docs for border-image https://bugzilla.gnome.org/show_bug.cgi?id=651194
2011-05-31 13:32:00 +00:00
* Images and gradients can also be used in slices for the purpose of creating
* scalable borders.
* For more information, see the CSS3 documentation for the border-image property,
* which can be found <ulink url="http://www.w3.org/TR/css3-background/#border-images">here</ulink>.
Document border images better
2010-11-22 16:15:47 +00:00
* </para>
* <inlinegraphic fileref="slices.png" format="PNG"/>
* <para>
docs: expand the docs for border-image https://bugzilla.gnome.org/show_bug.cgi?id=651194
2011-05-31 13:32:00 +00:00
* The parameters of the slicing process are controlled by
border-image: add support for border-image-width too It's useful to set a slice size != border-width, as backgrounds are clipped to border-width too. As slices can be half-transparent and overlap the background, this would not fill the border box properly if we only use a single property for specifying the width. Also, this brings us even closer to CSS3.
2011-06-09 23:59:14 +00:00
* four separate properties. Note that you can use the
docs: expand the docs for border-image https://bugzilla.gnome.org/show_bug.cgi?id=651194
2011-05-31 13:32:00 +00:00
* <literallayout>border-image</literallayout> shorthand property
* to set values for the three properties at the same time.
* </para>
* <para>
* <literallayout>border-image-source: url(@path)
* (or border-image-source: -gtk-gradient(...))</literallayout>:
* Specifies the source of the border image, and it can either
* be an URL or a gradient (see above).
* </para>
* <para>
* <literallayout>border-image-slice: @top @right @bottom @left</literallayout>
* The sizes specified by the @top, @right, @bottom and @left parameters
* are the offsets, in pixels, from the relevant edge where the image
* should be "cut off" to build the slices used for the rendering
* of the border.
* </para>
* <para>
border-image: add support for border-image-width too It's useful to set a slice size != border-width, as backgrounds are clipped to border-width too. As slices can be half-transparent and overlap the background, this would not fill the border box properly if we only use a single property for specifying the width. Also, this brings us even closer to CSS3.
2011-06-09 23:59:14 +00:00
* <literallayout>border-image-width: @top @right @bottom @left</literallayout>
* The sizes specified by the @top, @right, @bottom and @left parameters
* are inward distances from the border box edge, used to specify the
* rendered size of each slice determined by border-image-slice.
* If this property is not specified, the values of border-width will
* be used as a fallback.
* </para>
* <para>
docs: expand the docs for border-image https://bugzilla.gnome.org/show_bug.cgi?id=651194
2011-05-31 13:32:00 +00:00
* <literallayout>border-image-repeat: [stretch|repeat|round|space] ?
* [stretch|repeat|round|space]</literallayout>
* Specifies how the image slices should be rendered in the area
* outlined by border-width.
* The default (stretch) is to resize the slice to fill in the whole
* allocated area.
* If the value of this property is 'repeat', the image slice
* will be tiled to fill the area.
* If the value of this property is 'round', the image slice will
* be tiled to fill the area, and scaled to fit it exactly
* a whole number of times.
* If the value of this property is 'space', the image slice will
* be tiled to fill the area, and if it doesn't fit it exactly a whole
* number of times, the extra space is distributed as padding around
* the slices.
* If two options are specified, the first one affects
Document border images better
2010-11-22 16:15:47 +00:00
* the horizontal behaviour and the second one the vertical behaviour.
* If only one option is specified, it affects both.
* </para>
* <example>
* <title>A border image</title>
* <inlinegraphic fileref="border1.png" format="PNG"/>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* <para>This border image was specified with
* <literallayout>url("gradient1.png") 10 10 10 10</literallayout>
* </para>
Document border images better
2010-11-22 16:15:47 +00:00
* </example>
* <example>
* <title>A repeating border image</title>
* <inlinegraphic fileref="border2.png" format="PNG"/>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* <para>This border image was specified with
* <literallayout>url("gradient1.png") 10 10 10 10 repeat</literallayout>
* </para>
Document border images better
2010-11-22 16:15:47 +00:00
* </example>
* <example>
* <title>A stretched border image</title>
* <inlinegraphic fileref="border3.png" format="PNG"/>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* <para>This border image was specified with
* <literallayout>url("gradient1.png") 10 10 10 10 stretch</literallayout>
* </para>
Document border images better
2010-11-22 16:15:47 +00:00
* </example>
* </refsect2>
More documentation reshuffling
2010-11-22 20:12:42 +00:00
* <refsect2 id="gtkcssprovider-transitions">
* <para>Styles can specify transitions that will be used to create a gradual
* change in the appearance when a widget state changes. The following
* syntax is used to specify transitions:
* <literallayout>@duration [s|ms] [linear|ease|ease-in|ease-out|ease-in-out] [loop]?</literallayout>
* The @duration is the amount of time that the animation will take for
* a complete cycle from start to end. If the loop option is given, the
* animation will be repated until the state changes again.
* The option after the duration determines the transition function from a
* small set of predefined functions.
* <figure><title>Linear transition</title>
* <graphic fileref="linear.png" format="PNG"/>
* </figure>
* <figure><title>Ease transition</title>
* <graphic fileref="ease.png" format="PNG"/>
* </figure>
* <figure><title>Ease-in-out transition</title>
* <graphic fileref="ease-in-out.png" format="PNG"/>
* </figure>
* <figure><title>Ease-in transition</title>
* <graphic fileref="ease-in.png" format="PNG"/>
* </figure>
* <figure><title>Ease-out transition</title>
* <graphic fileref="ease-out.png" format="PNG"/>
* </figure>
* </para>
* </refsect2>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <refsect2 id="gtkcssprovider-properties">
* <title>Supported properties</title>
* <para>
* Properties are the part that differ the most to common CSS,
* not all properties are supported (some are planned to be
* supported eventually, some others are meaningless or don't
* map intuitively in a widget based environment).
* </para>
* <para>
* The currently supported properties are:
* </para>
* <informaltable>
* <tgroup cols="4">
* <thead>
* <row>
* <entry>Property name</entry>
* <entry>Syntax</entry>
* <entry>Maps to</entry>
* <entry>Examples</entry>
* </row>
* </thead>
* <tbody>
* <row>
* <entry>engine</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>engine-name</entry>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>#GtkThemingEngine</entry>
Mention gvfs in the gtk_show_uri() docs https://bugzilla.gnome.org/show_bug.cgi?id=622125
2011-01-04 16:25:40 +00:00
* <entry>engine: clearlooks;
* engine: none; /&ast; use the default (i.e. builtin) engine) &ast;/ </entry>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </row>
* <row>
* <entry>background-color</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry morerows="2">color (see above)</entry>
styleproperty: Add border-*-color style properties And make border-color a shorthand
2011-06-13 01:45:18 +00:00
* <entry morerows="7">#GdkRGBA</entry>
* <entry morerows="7"><literallayout>background-color: &num;fff;
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* color: &amp;color1;
* background-color: shade (&amp;color1, 0.5);
* color: mix (&amp;color1, &num;f0f, 0.8);</literallayout>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </entry>
* </row>
* <row>
Rely only on fg/bg colors. "foreground-color" has been renamed to just "color" as in regular CSS, and the "text-color" and "base-color" properties have been removed. The default CSS has been changed to have widgets cope with this change.
2010-10-29 03:33:38 +00:00
* <entry>color</entry>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </row>
* <row>
styleproperty: Add border-*-color style properties And make border-color a shorthand
2011-06-13 01:45:18 +00:00
* <entry>border-top-color</entry>
* <entry morerows="4">transparent|color (see above)</entry>
* </row>
* <row>
* <entry>border-right-color</entry>
* </row>
* <row>
* <entry>border-bottom-color</entry>
* </row>
* <row>
* <entry>border-left-color</entry>
* </row>
* <row>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>border-color</entry>
styleproperty: Add border-*-color style properties And make border-color a shorthand
2011-06-13 01:45:18 +00:00
* <entry>[transparent|color]{1,4}</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* </row>
* <row>
css: Make font property a shorthand ... and implement the CSS font properties: - font-size - font-style - font-family - font-weight - font-variant This is the second try at this. The first was backed out previously due to bugginess. Let's hope this one survives a bit longer. Also makes the font-family CSS test work again.
2011-05-29 05:03:59 +00:00
* <entry>font-family</entry>
* <entry>@family [, @family]*</entry>
* <entry>#gchararray</entry>
* <entry>font-family: Sans, Arial;</entry>
* </row>
* <row>
* <entry>font-style</entry>
* <entry>[normal|oblique|italic]</entry>
* <entry>#PANGO_TYPE_STYLE</entry>
* <entry>font-style: italic;</entry>
* </row>
* <row>
* <entry>font-variant</entry>
* <entry>[normal|small-caps]</entry>
* <entry>#PANGO_TYPE_VARIANT</entry>
* <entry>font-variant: normal;</entry>
* </row>
* <row>
* <entry>font-weight</entry>
* <entry>[normal|bold|bolder|lighter|100|200|300|400|500|600|700|800|900]</entry>
* <entry>#PANGO_TYPE_WEIGHT</entry>
* <entry>font-weight: bold;</entry>
* </row>
* <row>
* <entry>font-size</entry>
* <entry>Font size in point</entry>
* <entry>#gint</entry>
* <entry>font-size: 13;</entry>
* </row>
* <row>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>font</entry>
Document border images better
2010-11-22 16:15:47 +00:00
* <entry>@family [@style] [@size]</entry>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>#PangoFontDescription</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>font: Sans 15;</entry>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </row>
* <row>
style: Make margin a shorthand ... and add the real properties: - margin-top - margin-left - margin-bottom - margin-right
2011-05-22 03:55:12 +00:00
* <entry>margin-top</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>margin-top: 0;</entry>
* </row>
* <row>
* <entry>margin-left</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>margin-left: 1;</entry>
* </row>
* <row>
* <entry>margin-bottom</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>margin-bottom: 2;</entry>
* </row>
* <row>
* <entry>margin-right</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>margin-right: 4;</entry>
* </row>
* <row>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>margin</entry>
Document border images better
2010-11-22 16:15:47 +00:00
* <entry morerows="1"><literallayout>@width
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* @vertical_width @horizontal_width
* @top_width @horizontal_width @bottom_width
* @top_width @right_width @bottom_width @left_width</literallayout>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </entry>
* <entry morerows="1">#GtkBorder</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry morerows="1"><literallayout>margin: 5;
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* margin: 5 10;
* margin: 5 10 3;
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* margin: 5 10 3 5;</literallayout>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </entry>
* </row>
* <row>
style: Make padding a shorthand ... and add the real properties: - padding-top - padding-left - padding-bottom - padding-right
2011-05-22 03:42:17 +00:00
* <entry>padding-top</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>padding-top: 5;</entry>
* </row>
* <row>
* <entry>padding-left</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>padding-left: 5;</entry>
* </row>
* <row>
* <entry>padding-bottom</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>padding-bottom: 5;</entry>
* </row>
* <row>
* <entry>padding-right</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>padding-right: 5;</entry>
* </row>
* <row>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>padding</entry>
* </row>
* <row>
* <entry>background-image</entry>
Allow images for background-image In addition to -gtk-gradient(...), allow url(path), and create a pattern from the image. Still to do: allow stretch/repeat options.
2010-11-26 05:43:16 +00:00
* <entry><literallayout>gradient (see above) or
* url(@path)</literallayout></entry>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>#cairo_pattern_t</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry><literallayout>-gtk-gradient (linear,
GtkCssProvider: Docs fix for -gtk-gradient() syntax.
2010-11-13 12:27:37 +00:00
* left top, right top,
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* from (&num;fff), to (&num;000));
* -gtk-gradient (linear, 0.0 0.5, 0.5 1.0,
* from (&num;fff),
* color-stop (0.5, &num;f00),
* to (&num;000));
* -gtk-gradient (radial,
* center center, 0.2,
* center center, 0.8,
* color-stop (0.0, &num;fff),
Allow images for background-image In addition to -gtk-gradient(...), allow url(path), and create a pattern from the image. Still to do: allow stretch/repeat options.
2010-11-26 05:43:16 +00:00
* color-stop (1.0, &num;000));
* url ('background.png');</literallayout>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </entry>
* </row>
* <row>
css: start background-repeat By default, a background image is stretched. Instead, it is worth to have a tiled background. This patch allows background surfaces to be repeated or not, and should be compatible with future extensions and CSS. https://bugzilla.gnome.org/show_bug.cgi?id=663522
2011-11-06 18:33:05 +00:00
* <entry>background-repeat</entry>
* <entry>[repeat|no-repeat]</entry>
* <entry>internal</entry>
* <entry><literallayout>background-repeat: no-repeat;</literallayout>
* If not specified, the style doesn't respect the CSS3
* specification, since the background will be
* stretched to fill the area.
* </entry>
* </row>
* <row>
style: Convert border-width to a shorthand Add the corresponding properties it refers to, too: - border-top-width - border-left-width - border-bottom-width - border-right-width
2011-05-21 22:10:43 +00:00
* <entry>border-top-width</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>border-top-width: 5;</entry>
* </row>
* <row>
* <entry>border-left-width</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>border-left-width: 5;</entry>
* </row>
* <row>
* <entry>border-bottom-width</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry>integer</entry>
* <entry>#gint</entry>
style: Convert border-width to a shorthand Add the corresponding properties it refers to, too: - border-top-width - border-left-width - border-bottom-width - border-right-width
2011-05-21 22:10:43 +00:00
* <entry>border-bottom-width: 5;</entry>
* </row>
* <row>
* <entry>border-right-width</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>border-right-width: 5;</entry>
* </row>
* <row>
* <entry>border-width</entry>
* <entry morerows="1">#GtkBorder</entry>
* <entry morerows="1"><literallayout>border-width: 1;
* border-width: 1 2;
* border-width: 1 2 3;
* border-width: 1 2 3 5;</literallayout>
* </entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* </row>
* <row>
* <entry>border-radius</entry>
* <entry>integer</entry>
* <entry>#gint</entry>
* <entry>border-radius: 5;</entry>
* </row>
* <row>
* <entry>border-style</entry>
* <entry>[none|solid|inset|outset]</entry>
* <entry>#GtkBorderStyle</entry>
* <entry>border-style: solid;</entry>
* </row>
* <row>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <entry>border-image</entry>
Document border images better
2010-11-22 16:15:47 +00:00
* <entry><literallayout>border image (see above)</literallayout></entry>
Declare transition and border-image properties to be writeonly We don't make the types of their values public, so getting their values is useless and will leak.
2010-11-23 13:24:49 +00:00
* <entry>internal use only</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry><literallayout>border-image: url("/path/to/image.png") 3 4 3 4 stretch;
* border-image: url("/path/to/image.png") 3 4 4 3 repeat stretch;</literallayout>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </entry>
* </row>
docs: Add missing <row>
2011-05-21 01:41:52 +00:00
* <row>
styleproperties: add a "text-shadow" style property And document its use in the CSS parser gtk-doc. https://bugzilla.gnome.org/show_bug.cgi?id=649314
2011-05-20 20:11:08 +00:00
* <entry>text-shadow</entry>
* <entry>shadow list (see above)</entry>
docs: fix a number of typos and obsolete references
2012-07-02 06:19:06 +00:00
* <entry>internal use only</entry>
styleproperties: add a "text-shadow" style property And document its use in the CSS parser gtk-doc. https://bugzilla.gnome.org/show_bug.cgi?id=649314
2011-05-20 20:11:08 +00:00
* <entry><literallayout>text-shadow: 1 1 0 blue, -4 -4 red;</literallayout></entry>
* </row>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* <row>
* <entry>transition</entry>
More documentation reshuffling
2010-11-22 20:12:42 +00:00
* <entry>transition (see above)</entry>
Declare transition and border-image properties to be writeonly We don't make the types of their values public, so getting their values is useless and will leak.
2010-11-23 13:24:49 +00:00
* <entry>internal use only</entry>
Add some images to the docs Add graphs for the transition functions and gradient examples.
2010-11-22 06:41:45 +00:00
* <entry><literallayout>transition: 150ms ease-in-out;
* transition: 1s linear loop;</literallayout>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </entry>
* </row>
Document @binding-set and the gtk-key-bindings css parser property
2011-01-28 09:25:17 +00:00
* <row>
* <entry>gtk-key-bindings</entry>
* <entry>binding set name list</entry>
* <entry>internal use only</entry>
* <entry><literallayout>gtk-bindings: binding1, binding2, ...;</literallayout>
* </entry>
* </row>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </tbody>
* </tgroup>
* </informaltable>
Improve some gtk_render_ docs With pictures !!
2010-11-23 06:05:11 +00:00
* <para>
* GtkThemingEngines can register their own, engine-specific style properties
* with the function gtk_theming_engine_register_property(). These properties
* can be set in CSS like other properties, using a name of the form
* <literallayout>-<replaceable>namespace</replaceable>-<replaceable>name</replaceable></literallayout>, where <replaceable>namespace</replaceable> is typically
* the name of the theming engine, and <replaceable>name</replaceable> is the
* name of the property. Style properties that have been registered by widgets
* using gtk_widget_class_install_style_property() can also be set in this
* way, using the widget class name for <replaceable>namespace</replaceable>.
* </para>
* <example>
* <title>Using engine-specific style properties</title>
* <programlisting>
* * {
* engine: clearlooks;
* border-radius: 4;
* -GtkPaned-handle-size: 6;
* -clearlooks-colorize-scrollbar: false;
* }
* </programlisting>
* </example>
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* </refsect2>
*/
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
typedef struct GtkCssRuleset GtkCssRuleset;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
typedef struct _GtkCssScanner GtkCssScanner;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
typedef struct _PropertyValue PropertyValue;
typedef struct _WidgetPropertyValue WidgetPropertyValue;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
typedef enum ParserScope ParserScope;
GtkCssProvider: Parse named regions.
2010-04-20 23:46:52 +00:00
typedef enum ParserSymbol ParserSymbol;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
struct _PropertyValue {
GtkCssStyleProperty *property;
css: Use GtkCssValues instead of GValue in the css machinery Also, in places where we're computing a new CssValue based on an old one, make sure that if nothing changes we're returning a reference to the old one, rather than creating a new identical instance.
2012-03-06 13:16:32 +00:00
GtkCssValue *value;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
GtkCssSection *section;
};
struct _WidgetPropertyValue {
WidgetPropertyValue *next;
cssprovider: Store widget style value as string There's no need to turn it into a GtkCssValue, it's always a string.
2012-03-30 00:45:01 +00:00
char *name;
char *value;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
GtkCssSection *section;
};
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
struct GtkCssRuleset
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
GtkCssSelector *selector;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
WidgetPropertyValue *widget_style;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
PropertyValue *styles;
cssprovider: Keep a bitmask for the properties That way we can easily tell wich properties are in use.
2011-12-23 13:23:52 +00:00
GtkBitmask *set_styles;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
guint n_styles;
guint owns_styles : 1;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
guint owns_widget_style : 1;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
};
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
struct _GtkCssScanner
cssprovider: Add a private structure to the scanner So far, it's unused
2011-04-10 01:39:52 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssProvider *provider;
GtkCssParser *parser;
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
GtkCssSection *section;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *parent;
cssprovider: track state in the scanner This way, we don't have to do magic copies when we create a new scanner for a subfile.
2011-04-10 01:51:43 +00:00
GSList *state;
cssprovider: Add a private structure to the scanner So far, it's unused
2011-04-10 01:39:52 +00:00
};
cssprovider: Move private struct definition to header consistency, improved debugging, etc
2011-04-06 14:10:25 +00:00
struct _GtkCssProviderPrivate
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
GScanner *scanner;
GtkCssProvider: Improve error handling and reporting. Nicer error reports are provided, and a GError can be spread from anywhere in parsing, so over time more precise and meaningful messages can be produced.
2010-12-03 20:56:39 +00:00
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
GHashTable *symbolic_colors;
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
GArray *rulesets;
Support a gtk.gresource file in themes This lets themes install a .gresource for efficient loading of resources. This resource file will (if it exists) be automatically loaded (mmaped) when the theme is used and unloaded when not used.
2012-02-03 14:50:49 +00:00
GResource *resource;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
};
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
enum {
PARSING_ERROR,
LAST_SIGNAL
};
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
static gboolean gtk_keep_css_sections = FALSE;
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
static guint css_provider_signals[LAST_SIGNAL] = { 0 };
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
static void gtk_css_provider_finalize (GObject *object);
static void gtk_css_style_provider_iface_init (GtkStyleProviderIface *iface);
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
static void gtk_css_style_provider_private_iface_init (GtkStyleProviderPrivateInterface *iface);
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
static void widget_property_value_list_free (WidgetPropertyValue *head);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Make @import use the new load_internal() func
2011-04-11 23:10:13 +00:00
static gboolean
gtk_css_provider_load_internal (GtkCssProvider *css_provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner,
cssprovider: Make @import use the new load_internal() func
2011-04-11 23:10:13 +00:00
GFile *file,
const char *data,
GError **error);
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
Some error reporting When we take a GError argument, better use it.
2010-11-18 03:29:39 +00:00
GQuark
gtk_css_provider_error_quark (void)
{
return g_quark_from_static_string ("gtk-css-provider-error-quark");
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
G_DEFINE_TYPE_EXTENDED (GtkCssProvider, gtk_css_provider, G_TYPE_OBJECT, 0,
G_IMPLEMENT_INTERFACE (GTK_TYPE_STYLE_PROVIDER,
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
gtk_css_style_provider_iface_init)
G_IMPLEMENT_INTERFACE (GTK_TYPE_STYLE_PROVIDER_PRIVATE,
gtk_css_style_provider_private_iface_init));
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
static void
gtk_css_provider_parsing_error (GtkCssProvider *provider,
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
GtkCssSection *section,
const GError *error)
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
{
cssprovider: Emit a g_warning() when a css file is erroneous ... and no other error handler is active. See the source code comment for the reasoning.
2011-04-10 23:51:49 +00:00
/* Only emit a warning when we have no error handlers. This is our
* default handlers. And in this case erroneous CSS files are a bug
* and should be fixed.
* Note that these warnings can also be triggered by a broken theme
* that people installed from some weird location on the internets.
*/
if (!g_signal_has_handler_pending (provider,
css_provider_signals[PARSING_ERROR],
0,
TRUE))
{
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
GFileInfo *info;
GFile *file;
const char *path;
file = gtk_css_section_get_file (section);
if (file)
{
gtkcssprovider: fix a crash in gtk_css_provider_parsing_error "info" was being declared in two different scopes, and the one that was freed wasn't the one that was set.
2011-09-09 13:32:54 +00:00
info = g_file_query_info (file, G_FILE_ATTRIBUTE_STANDARD_DISPLAY_NAME, 0, NULL, NULL);
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
if (info)
path = g_file_info_get_display_name (info);
else
path = "<broken file>";
}
else
{
info = NULL;
path = "<data>";
}
g_warning ("Theme parsing error: %s:%u:%u: %s",
path,
gtk_css_section_get_end_line (section) + 1,
gtk_css_section_get_end_position (section),
error->message);
if (info)
g_object_unref (info);
cssprovider: Emit a g_warning() when a css file is erroneous ... and no other error handler is active. See the source code comment for the reasoning.
2011-04-10 23:51:49 +00:00
}
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
static void
gtk_css_provider_class_init (GtkCssProviderClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
if (g_getenv ("GTK_CSS_DEBUG"))
gtk_keep_css_sections = TRUE;
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
/**
* GtkCssProvider::parsing-error:
* @provider: the provider that had a parsing error
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
* @section: section the error happened in
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
* @error: The parsing error
*
* Signals that a parsing error occured. the @path, @line and @position
* describe the actual location of the error as accurately as possible.
*
* Parsing errors are never fatal, so the parsing will resume after
* the error. Errors may however cause parts of the given
* data or even all of it to not be parsed at all. So it is a useful idea
* to check that the parsing succeeds by connecting to this signal.
*
* Note that this signal may be emitted at any time as the css provider
* may opt to defer parsing parts or all of the input to a later time
* than when a loading function was called.
*/
css_provider_signals[PARSING_ERROR] =
g_signal_new (I_("parsing-error"),
G_TYPE_FROM_CLASS (object_class),
G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GtkCssProviderClass, parsing_error),
NULL, NULL,
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
_gtk_marshal_VOID__BOXED_BOXED,
G_TYPE_NONE, 2, GTK_TYPE_CSS_SECTION, G_TYPE_ERROR);
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
object_class->finalize = gtk_css_provider_finalize;
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
klass->parsing_error = gtk_css_provider_parsing_error;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
g_type_class_add_private (object_class, sizeof (GtkCssProviderPrivate));
}
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
static void
gtk_css_ruleset_init_copy (GtkCssRuleset *new,
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
GtkCssRuleset *ruleset,
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
GtkCssSelector *selector)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
memcpy (new, ruleset, sizeof (GtkCssRuleset));
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
new->selector = selector;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
/* First copy takes over ownership */
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
if (ruleset->owns_styles)
ruleset->owns_styles = FALSE;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
if (ruleset->owns_widget_style)
ruleset->owns_widget_style = FALSE;
cssprovider: Keep a bitmask for the properties That way we can easily tell wich properties are in use.
2011-12-23 13:23:52 +00:00
if (new->set_styles)
new->set_styles = _gtk_bitmask_copy (new->set_styles);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
static void
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
gtk_css_ruleset_clear (GtkCssRuleset *ruleset)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
if (ruleset->owns_styles)
{
guint i;
for (i = 0; i < ruleset->n_styles; i++)
{
css: Use GtkCssValues instead of GValue in the css machinery Also, in places where we're computing a new CssValue based on an old one, make sure that if nothing changes we're returning a reference to the old one, rather than creating a new identical instance.
2012-03-06 13:16:32 +00:00
_gtk_css_value_unref (ruleset->styles[i].value);
ruleset->styles[i].value = NULL;
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
if (ruleset->styles[i].section)
gtk_css_section_unref (ruleset->styles[i].section);
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
}
g_free (ruleset->styles);
}
cssprovider: Keep a bitmask for the properties That way we can easily tell wich properties are in use.
2011-12-23 13:23:52 +00:00
if (ruleset->set_styles)
_gtk_bitmask_free (ruleset->set_styles);
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
if (ruleset->owns_widget_style)
widget_property_value_list_free (ruleset->widget_style);
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
if (ruleset->selector)
_gtk_css_selector_free (ruleset->selector);
Plugged leak in gtkcssprovider.c SelectorStyleInfo structs were never freed.
2011-01-16 11:52:43 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
memset (ruleset, 0, sizeof (GtkCssRuleset));
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
static WidgetPropertyValue *
widget_property_value_new (char *name, GtkCssSection *section)
{
WidgetPropertyValue *value;
value = g_slice_new0 (WidgetPropertyValue);
value->name = name;
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
if (gtk_keep_css_sections)
value->section = gtk_css_section_ref (section);
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
return value;
}
static void
widget_property_value_free (WidgetPropertyValue *value)
{
cssprovider: Store widget style value as string There's no need to turn it into a GtkCssValue, it's always a string.
2012-03-30 00:45:01 +00:00
g_free (value->value);
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
g_free (value->name);
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
if (value->section)
gtk_css_section_unref (value->section);
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
g_slice_free (WidgetPropertyValue, value);
}
static void
widget_property_value_list_free (WidgetPropertyValue *head)
{
WidgetPropertyValue *l, *next;
for (l = head; l != NULL; l = next)
{
next = l->next;
widget_property_value_free (l);
}
}
static WidgetPropertyValue *
widget_property_value_list_remove_name (WidgetPropertyValue *head, const char *name)
{
WidgetPropertyValue *l, **last;
last = &head;
for (l = head; l != NULL; l = l->next)
{
if (strcmp (l->name, name) == 0)
{
*last = l->next;
widget_property_value_free (l);
break;
}
last = &l->next;
}
return head;
}
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
static void
gtk_css_ruleset_add_style (GtkCssRuleset *ruleset,
char *name,
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
WidgetPropertyValue *value)
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
{
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
value->next = widget_property_value_list_remove_name (ruleset->widget_style, name);
ruleset->widget_style = value;
ruleset->owns_widget_style = TRUE;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
cssprovider: Add a new setter for rulesets We now have set-by-pspec (which is used for normal style properties) and set-by-name (which is used by widget properties).
2011-05-18 02:30:51 +00:00
static void
css: Fix fallout .. from carelessly merging 85b1b9b77816198040d9437ea2e1231c47e41e68
2012-03-01 04:28:41 +00:00
gtk_css_ruleset_add (GtkCssRuleset *ruleset,
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
GtkCssStyleProperty *property,
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
GtkCssValue *value,
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
GtkCssSection *section)
cssprovider: Add a new setter for rulesets We now have set-by-pspec (which is used for normal style properties) and set-by-name (which is used by widget properties).
2011-05-18 02:30:51 +00:00
{
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
guint i;
g_return_if_fail (ruleset->owns_styles || ruleset->n_styles == 0);
css: Fix fallout .. from carelessly merging 85b1b9b77816198040d9437ea2e1231c47e41e68
2012-03-01 04:28:41 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
if (ruleset->set_styles == NULL)
ruleset->set_styles = _gtk_bitmask_new ();
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
bitmask: Make setters return a new value (Actually, it's not a real 'new' value yet, but will be soon. This is the first step to make bitmasks immutable.
2012-02-08 18:19:59 +00:00
ruleset->set_styles = _gtk_bitmask_set (ruleset->set_styles,
_gtk_css_style_property_get_id (property),
TRUE);
style: Add support for shorthand properties Shorthand properties are basically the same a in CSS. For storage in style properties or the CSS provider, they are unpacked into the real values, so it is possible to partially override them. No properties are yet converted to the new world yet, this is just the code for supporting them.
2011-05-21 21:47:19 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
ruleset->owns_styles = TRUE;
for (i = 0; i < ruleset->n_styles; i++)
{
if (ruleset->styles[i].property == property)
{
css: Use GtkCssValues instead of GValue in the css machinery Also, in places where we're computing a new CssValue based on an old one, make sure that if nothing changes we're returning a reference to the old one, rather than creating a new identical instance.
2012-03-06 13:16:32 +00:00
_gtk_css_value_unref (ruleset->styles[i].value);
ruleset->styles[i].value = NULL;
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
if (ruleset->styles[i].section)
gtk_css_section_unref (ruleset->styles[i].section);
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
break;
}
}
if (i == ruleset->n_styles)
{
ruleset->n_styles++;
ruleset->styles = g_realloc (ruleset->styles, ruleset->n_styles * sizeof (PropertyValue));
css: Use GtkCssValues instead of GValue in the css machinery Also, in places where we're computing a new CssValue based on an old one, make sure that if nothing changes we're returning a reference to the old one, rather than creating a new identical instance.
2012-03-06 13:16:32 +00:00
ruleset->styles[i].value = NULL;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
ruleset->styles[i].property = property;
}
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
ruleset->styles[i].value = value;
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
if (gtk_keep_css_sections)
ruleset->styles[i].section = gtk_css_section_ref (section);
else
ruleset->styles[i].section = NULL;
cssprovider: Add a new setter for rulesets We now have set-by-pspec (which is used for normal style properties) and set-by-name (which is used by widget properties).
2011-05-18 02:30:51 +00:00
}
cssprovider: Refactor code to have a gtk_css_ruleset_matches()
2011-05-17 20:11:50 +00:00
static gboolean
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
gtk_css_ruleset_matches (GtkCssRuleset *ruleset,
const GtkCssMatcher *matcher)
cssprovider: Refactor code to have a gtk_css_ruleset_matches()
2011-05-17 20:11:50 +00:00
{
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
return _gtk_css_selector_matches (ruleset->selector, matcher);
cssprovider: Refactor code to have a gtk_css_ruleset_matches()
2011-05-17 20:11:50 +00:00
}
syleprovider: Add a vfunc to get the changes This way we can check what changes are even interesting for our matcher.
2012-03-17 20:09:51 +00:00
static GtkCssChange
gtk_css_ruleset_get_change (GtkCssRuleset *ruleset)
{
return _gtk_css_selector_get_change (ruleset->selector);
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static void
gtk_css_scanner_destroy (GtkCssScanner *scanner)
{
cssprovider: Fix leak Thanks to the heroic valgrinding efforts of Olivier Sessnik, another leak could be plugged.
2011-12-23 16:13:42 +00:00
if (scanner->section)
gtk_css_section_unref (scanner->section);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
g_object_unref (scanner->provider);
_gtk_css_parser_free (scanner->parser);
g_slice_free (GtkCssScanner, scanner);
cssprovider: track state in the scanner This way, we don't have to do magic copies when we create a new scanner for a subfile.
2011-04-10 01:51:43 +00:00
}
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
static void
gtk_css_provider_emit_error (GtkCssProvider *provider,
GtkCssScanner *scanner,
const GError *error)
{
g_signal_emit (provider, css_provider_signals[PARSING_ERROR], 0,
GtkCssProvider: don't segfault when CSS file is not found * Fix null dereference in gtk_css_provider_take_error() * And another in gtk_css_provider_emit_error() https://bugzilla.gnome.org/show_bug.cgi?id=664537
2011-11-22 07:13:04 +00:00
scanner != NULL ? scanner->section : NULL, error);
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
}
cssprovider: Add a custom gtk_css_scanner_destroy() function So far, it does nothing
2011-04-10 01:35:19 +00:00
static void
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
gtk_css_scanner_parser_error (GtkCssParser *parser,
const GError *error,
gpointer user_data)
cssprovider: Add a custom gtk_css_scanner_destroy() function So far, it does nothing
2011-04-10 01:35:19 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner = user_data;
cssprovider: Store current properties and selectors in scanner They don't belong in the css provider
2011-04-10 23:28:00 +00:00
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
gtk_css_provider_emit_error (scanner->provider,
scanner,
error);
cssprovider: Add a custom gtk_css_scanner_destroy() function So far, it does nothing
2011-04-10 01:35:19 +00:00
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static GtkCssScanner *
gtk_css_scanner_new (GtkCssProvider *provider,
GtkCssScanner *parent,
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
GtkCssSection *section,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GFile *file,
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
const gchar *text)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
scanner = g_slice_new0 (GtkCssScanner);
cssprovider: Add a private structure to the scanner So far, it's unused
2011-04-10 01:39:52 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
g_object_ref (provider);
scanner->provider = provider;
scanner->parent = parent;
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
if (section)
scanner->section = gtk_css_section_ref (section);
cssprovider: Allow storing the parent stylesheet
2011-04-12 00:06:08 +00:00
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
scanner->parser = _gtk_css_parser_new (text,
cssparser: Keep track of the file
2012-04-18 14:54:04 +00:00
file,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
gtk_css_scanner_parser_error,
scanner);
cssprovider: Move properties to scanner constructor So we can set them in one place.
2011-04-11 14:48:00 +00:00
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
return scanner;
}
cssprovider: Add recursion checks for @import This is pretty important, because otherwise recursions cause crashes. And if you accidentally change your theme to one that crashes on load, all your gonna SEGV and then on reboot, gdm tries to load the theme...
2011-04-12 00:50:25 +00:00
static gboolean
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
gtk_css_scanner_would_recurse (GtkCssScanner *scanner,
GFile *file)
cssprovider: Add recursion checks for @import This is pretty important, because otherwise recursions cause crashes. And if you accidentally change your theme to one that crashes on load, all your gonna SEGV and then on reboot, gdm tries to load the theme...
2011-04-12 00:50:25 +00:00
{
while (scanner)
{
cssprovider: Remove unused members from scanner Now that the parser takes care of the file, we don't need to keep track of it ourselves anymore.
2012-04-18 20:19:30 +00:00
GFile *parser_file = _gtk_css_parser_get_file (scanner->parser);
if (parser_file && g_file_equal (parser_file, file))
cssprovider: Add recursion checks for @import This is pretty important, because otherwise recursions cause crashes. And if you accidentally change your theme to one that crashes on load, all your gonna SEGV and then on reboot, gdm tries to load the theme...
2011-04-12 00:50:25 +00:00
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
scanner = scanner->parent;
cssprovider: Add recursion checks for @import This is pretty important, because otherwise recursions cause crashes. And if you accidentally change your theme to one that crashes on load, all your gonna SEGV and then on reboot, gdm tries to load the theme...
2011-04-12 00:50:25 +00:00
}
return FALSE;
}
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
static void
gtk_css_scanner_push_section (GtkCssScanner *scanner,
GtkCssSectionType section_type)
{
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
GtkCssSection *section;
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
section = _gtk_css_section_new (scanner->section,
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
section_type,
csssection: Remove file argument from constructor The parser has the file available.
2012-04-18 20:15:32 +00:00
scanner->parser);
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
if (scanner->section)
gtk_css_section_unref (scanner->section);
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
scanner->section = section;
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
}
static void
gtk_css_scanner_pop_section (GtkCssScanner *scanner,
GtkCssSectionType check_type)
{
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
GtkCssSection *parent;
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
g_assert (gtk_css_section_get_section_type (scanner->section) == check_type);
parent = gtk_css_section_get_parent (scanner->section);
if (parent)
gtk_css_section_ref (parent);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
_gtk_css_section_end (scanner->section);
gtk_css_section_unref (scanner->section);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
scanner->section = parent;
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
}
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
static void
gtk_css_provider_init (GtkCssProvider *css_provider)
{
GtkCssProviderPrivate *priv;
priv = css_provider->priv = G_TYPE_INSTANCE_GET_PRIVATE (css_provider,
GTK_TYPE_CSS_PROVIDER,
GtkCssProviderPrivate);
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
priv->rulesets = g_array_new (FALSE, FALSE, sizeof (GtkCssRuleset));
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
priv->symbolic_colors = g_hash_table_new_full (g_str_hash, g_str_equal,
(GDestroyNotify) g_free,
(GDestroyNotify) gtk_symbolic_color_unref);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
static void
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
css_provider_dump_symbolic_colors (GtkCssProvider *css_provider,
GtkStyleProperties *props)
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
{
GtkCssProviderPrivate *priv;
GHashTableIter iter;
gpointer key, value;
GtkCssProvider: Add priv pointer to object struct.
2010-07-03 00:19:27 +00:00
priv = css_provider->priv;
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
g_hash_table_iter_init (&iter, priv->symbolic_colors);
while (g_hash_table_iter_next (&iter, &key, &value))
{
const gchar *name;
GtkSymbolicColor *color;
name = key;
color = value;
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
gtk_style_properties_map_color (props, name, color);
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
}
}
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
static GtkStyleProperties *
GtkCssProvider: refactor out selector matching.
2010-06-18 14:52:34 +00:00
gtk_css_provider_get_style (GtkStyleProvider *provider,
GtkWidgetPath *path)
{
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
GtkCssMatcher matcher;
GtkCssProvider: Add priv pointer to object struct.
2010-07-03 00:19:27 +00:00
GtkCssProvider *css_provider;
cssprovider: Simplify code Don't go through an intermediate array when matching styles.
2011-05-16 17:54:57 +00:00
GtkCssProviderPrivate *priv;
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
GtkStyleProperties *props;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
guint i, j;
GtkCssProvider: refactor out selector matching.
2010-06-18 14:52:34 +00:00
GtkCssProvider: Add priv pointer to object struct.
2010-07-03 00:19:27 +00:00
css_provider = GTK_CSS_PROVIDER (provider);
cssprovider: Simplify code Don't go through an intermediate array when matching styles.
2011-05-16 17:54:57 +00:00
priv = css_provider->priv;
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
props = gtk_style_properties_new ();
GtkCssProvider: refactor out selector matching.
2010-06-18 14:52:34 +00:00
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
css_provider_dump_symbolic_colors (css_provider, props);
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
if (_gtk_css_matcher_init (&matcher, path, 0))
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
for (i = 0; i < priv->rulesets->len; i++)
{
GtkCssRuleset *ruleset;
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
ruleset = &g_array_index (priv->rulesets, GtkCssRuleset, i);
cssprovider: Simplify code Don't go through an intermediate array when matching styles.
2011-05-16 17:54:57 +00:00
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
if (ruleset->styles == NULL)
continue;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
if (!gtk_css_ruleset_matches (ruleset, &matcher))
continue;
css: Refactor code to do property lookups earlier We want to ook up the property in the CSS parser, so we can do fancy things with it. We currently don't but we want to later.
2011-05-16 21:37:29 +00:00
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
for (j = 0; j < ruleset->n_styles; j++)
_gtk_style_properties_set_property_by_property (props,
GTK_CSS_STYLE_PROPERTY (ruleset->styles[i].property),
_gtk_css_selector_get_state_flags (ruleset->selector),
ruleset->styles[i].value);
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
s/GtkStyleSet/GtkStyleProperties/ The former name had oddities such as gtk_style_set_set() and such, so the new name is nicer in that regard and not much longer.
2010-10-29 04:47:32 +00:00
return props;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
static gboolean
gtk_css_provider_get_style_property (GtkStyleProvider *provider,
GtkWidgetPath *path,
GtkStyleProvider: Add GtkStateFlags parameter to get_style_property(). Widget style properties can now have different values depending on the current state.
2010-11-30 01:14:00 +00:00
GtkStateFlags state,
GtkStyleProvider: Pass a GParamSpec in get_style_property(). This is so we can know the owner type of the property, and matching with the stored strings in GtkCssProvider is direct.
2010-11-17 01:42:27 +00:00
GParamSpec *pspec,
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
GValue *value)
{
cssprovider: Speed up gtk_widget_style_get() property lookups Previously we got the list of all matching rules and then iterated it to find the first one that had the property. Now we look while matching rules, so we don't lookup rules that we don't need.
2011-05-16 15:55:50 +00:00
GtkCssProvider *css_provider = GTK_CSS_PROVIDER (provider);
GtkCssProviderPrivate *priv = css_provider->priv;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
WidgetPropertyValue *val;
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
GtkCssMatcher matcher;
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
gboolean found = FALSE;
gchar *prop_name;
gint i;
cssmatcher: Handle case of empty widget path This is tested by the stylecontext test, but doesn't appear in practice.
2012-04-30 19:28:57 +00:00
if (!_gtk_css_matcher_init (&matcher, path, state))
return FALSE;
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
prop_name = g_strdup_printf ("-%s-%s",
GtkStyleProvider: Pass a GParamSpec in get_style_property(). This is so we can know the owner type of the property, and matching with the stored strings in GtkCssProvider is direct.
2010-11-17 01:42:27 +00:00
g_type_name (pspec->owner_type),
pspec->name);
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
for (i = priv->rulesets->len - 1; i >= 0; i--)
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
{
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
GtkCssRuleset *ruleset;
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
ruleset = &g_array_index (priv->rulesets, GtkCssRuleset, i);
cssprovider: Speed up gtk_widget_style_get() property lookups Previously we got the list of all matching rules and then iterated it to find the first one that had the property. Now we look while matching rules, so we don't lookup rules that we don't need.
2011-05-16 15:55:50 +00:00
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
if (ruleset->widget_style == NULL)
continue;
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
if (!gtk_css_ruleset_matches (ruleset, &matcher))
cssprovider: Speed up gtk_widget_style_get() property lookups Previously we got the list of all matching rules and then iterated it to find the first one that had the property. Now we look while matching rules, so we don't lookup rules that we don't need.
2011-05-16 15:55:50 +00:00
continue;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
for (val = ruleset->widget_style; val != NULL; val = val->next)
{
if (strcmp (val->name, prop_name) == 0)
{
GtkCssScanner *scanner;
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
scanner = gtk_css_scanner_new (css_provider,
NULL,
val->section,
css: Only keep around GtkCssSections if GTK_CSS_DEBUG env var set These are used purely for CSS debugging, and was using over 100k just starting up gtk3-demo.
2012-03-29 11:57:22 +00:00
val->section != NULL ? gtk_css_section_get_file (val->section) : NULL,
cssprovider: Store widget style value as string There's no need to turn it into a GtkCssValue, it's always a string.
2012-03-30 00:45:01 +00:00
val->value);
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
found = _gtk_css_style_parse_value (value,
cssstylefuncs: Remove base argument The GtkCssParser keeps track of the base file now.
2012-04-18 19:52:36 +00:00
scanner->parser);
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
gtk_css_scanner_destroy (scanner);
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
break;
}
}
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
if (found)
break;
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
}
g_free (prop_name);
return found;
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
static void
gtk_css_style_provider_iface_init (GtkStyleProviderIface *iface)
{
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
iface->get_style = gtk_css_provider_get_style;
iface->get_style_property = gtk_css_provider_get_style_property;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
static GtkSymbolicColor *
gtk_css_style_provider_get_color (GtkStyleProviderPrivate *provider,
const char *name)
{
GtkCssProvider *css_provider = GTK_CSS_PROVIDER (provider);
return g_hash_table_lookup (css_provider->priv->symbolic_colors, name);
}
static void
gtk_css_style_provider_lookup (GtkStyleProviderPrivate *provider,
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
const GtkCssMatcher *matcher,
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
GtkCssLookup *lookup)
{
GtkCssProvider *css_provider;
GtkCssProviderPrivate *priv;
int i;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
guint j;
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
css_provider = GTK_CSS_PROVIDER (provider);
priv = css_provider->priv;
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
for (i = priv->rulesets->len - 1; i >= 0; i--)
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
{
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
GtkCssRuleset *ruleset;
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
ruleset = &g_array_index (priv->rulesets, GtkCssRuleset, i);
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
if (ruleset->styles == NULL)
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
continue;
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
css: Add generic support for 'inherit' and 'initial' CSS3 says they work for every property, so here we go.
2011-12-29 15:18:12 +00:00
if (!_gtk_bitmask_intersects (_gtk_css_lookup_get_missing (lookup),
ruleset->set_styles))
continue;
css: Make the style provider take a matcher as an input argument
2012-03-17 01:36:11 +00:00
if (!gtk_css_ruleset_matches (ruleset, matcher))
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
continue;
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
for (j = 0; j < ruleset->n_styles; j++)
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
{
css: Fix variable And you'd think I'd test my code...
2012-03-01 12:46:11 +00:00
GtkCssStyleProperty *prop = ruleset->styles[j].property;
css: Move a bunch of functions ... from GtkStyleProperty to GtkCssStyleProperty.
2011-12-31 18:45:17 +00:00
guint id = _gtk_css_style_property_get_id (prop);
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
css: Move a bunch of functions ... from GtkStyleProperty to GtkCssStyleProperty.
2011-12-31 18:45:17 +00:00
if (!_gtk_css_lookup_is_missing (lookup, id))
cssprovider: Don't do inheritance anymore Its' done in GtkCssLookup with the help of GtkStyleContext::parent now.
2011-12-29 13:15:39 +00:00
continue;
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
_gtk_css_lookup_set (lookup,
id,
css: Fix variable And you'd think I'd test my code...
2012-03-01 12:46:11 +00:00
ruleset->styles[j].section,
css: Use GtkCssValues instead of GValue in the css machinery Also, in places where we're computing a new CssValue based on an old one, make sure that if nothing changes we're returning a reference to the old one, rather than creating a new identical instance.
2012-03-06 13:16:32 +00:00
ruleset->styles[j].value);
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
}
}
}
syleprovider: Add a vfunc to get the changes This way we can check what changes are even interesting for our matcher.
2012-03-17 20:09:51 +00:00
static GtkCssChange
gtk_css_style_provider_get_change (GtkStyleProviderPrivate *provider,
const GtkCssMatcher *matcher)
{
GtkCssProvider *css_provider;
GtkCssProviderPrivate *priv;
GtkCssChange change = 0;
int i;
css_provider = GTK_CSS_PROVIDER (provider);
priv = css_provider->priv;
for (i = priv->rulesets->len - 1; i >= 0; i--)
{
GtkCssRuleset *ruleset;
ruleset = &g_array_index (priv->rulesets, GtkCssRuleset, i);
if (ruleset->styles == NULL)
continue;
if (!gtk_css_ruleset_matches (ruleset, matcher))
continue;
change |= gtk_css_ruleset_get_change (ruleset);
}
return change;
}
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
static void
gtk_css_style_provider_private_iface_init (GtkStyleProviderPrivateInterface *iface)
{
iface->get_color = gtk_css_style_provider_get_color;
iface->lookup = gtk_css_style_provider_lookup;
syleprovider: Add a vfunc to get the changes This way we can check what changes are even interesting for our matcher.
2012-03-17 20:09:51 +00:00
iface->get_change = gtk_css_style_provider_get_change;
css: Redo StyleProvider interface We now use the GtkStleProviderPrivate interface, which hopefully is faster and more conformant to CSS. Long term, it definitely should be both. I would have liked to split this up into multiple commits, but couldn't find a way.
2011-12-28 21:59:55 +00:00
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
static void
gtk_css_provider_finalize (GObject *object)
{
GtkCssProvider: Add priv pointer to object struct.
2010-07-03 00:19:27 +00:00
GtkCssProvider *css_provider;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
GtkCssProviderPrivate *priv;
gtkcssprovider: plug several leaks In finalize(), clear all rulesets. In parse_declaration(), Free the GValue under unhandled error situations. In gtk_css_provider_load_internal(), Do not leak the file contents.
2011-05-30 17:26:28 +00:00
guint i;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
GtkCssProvider: Add priv pointer to object struct.
2010-07-03 00:19:27 +00:00
css_provider = GTK_CSS_PROVIDER (object);
priv = css_provider->priv;
GtkCssProvider: set scanner->input_name.
2010-06-19 14:04:30 +00:00
gtkcssprovider: plug several leaks In finalize(), clear all rulesets. In parse_declaration(), Free the GValue under unhandled error situations. In gtk_css_provider_load_internal(), Do not leak the file contents.
2011-05-30 17:26:28 +00:00
for (i = 0; i < priv->rulesets->len; i++)
gtk_css_ruleset_clear (&g_array_index (priv->rulesets, GtkCssRuleset, i));
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
g_array_free (priv->rulesets, TRUE);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
Some error reporting When we take a GError argument, better use it.
2010-11-18 03:29:39 +00:00
if (priv->symbolic_colors)
g_hash_table_destroy (priv->symbolic_colors);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
Support a gtk.gresource file in themes This lets themes install a .gresource for efficient loading of resources. This resource file will (if it exists) be automatically loaded (mmaped) when the theme is used and unloaded when not used.
2012-02-03 14:50:49 +00:00
if (priv->resource)
{
g_resources_unregister (priv->resource);
g_resource_unref (priv->resource);
priv->resource = NULL;
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
G_OBJECT_CLASS (gtk_css_provider_parent_class)->finalize (object);
}
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* gtk_css_provider_new:
*
* Returns a newly created #GtkCssProvider.
*
* Returns: A new #GtkCssProvider
**/
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
GtkCssProvider *
gtk_css_provider_new (void)
{
return g_object_new (GTK_TYPE_CSS_PROVIDER, NULL);
}
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
static void
gtk_css_provider_take_error (GtkCssProvider *provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner,
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
GError *error)
{
GtkCssProvider: don't segfault when CSS file is not found * Fix null dereference in gtk_css_provider_take_error() * And another in gtk_css_provider_emit_error() https://bugzilla.gnome.org/show_bug.cgi?id=664537
2011-11-22 07:13:04 +00:00
gtk_css_provider_emit_error (provider,
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
scanner,
error);
g_error_free (error);
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
}
cssprovider: Rework declaration parsing loop The new loop can cope better with invalid identifiers and manages to resume parsing properly.
2011-04-09 16:18:02 +00:00
static void
gtk_css_provider_error_literal (GtkCssProvider *provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner,
cssprovider: Rework declaration parsing loop The new loop can cope better with invalid identifiers and manages to resume parsing properly.
2011-04-09 16:18:02 +00:00
GQuark domain,
gint code,
const char *message)
{
gtk_css_provider_take_error (provider,
cssprovider: Pass scanner argument explicitly to error functions
2011-04-10 13:22:07 +00:00
scanner,
cssprovider: Rework declaration parsing loop The new loop can cope better with invalid identifiers and manages to resume parsing properly.
2011-04-09 16:18:02 +00:00
g_error_new_literal (domain, code, message));
}
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
static void
gtk_css_provider_error (GtkCssProvider *provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner,
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
GQuark domain,
gint code,
const char *format,
cssprovider: Pass scanner argument explicitly to error functions
2011-04-10 13:22:07 +00:00
...) G_GNUC_PRINTF (5, 6);
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
static void
gtk_css_provider_error (GtkCssProvider *provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner,
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
GQuark domain,
gint code,
const char *format,
...)
{
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
GError *error;
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
va_list args;
va_start (args, format);
API: cssprovider: Make error a signal
2011-04-07 15:26:47 +00:00
error = g_error_new_valist (domain, code, format, args);
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
va_end (args);
cssprovider: Pass scanner argument explicitly to error functions
2011-04-10 13:22:07 +00:00
gtk_css_provider_take_error (provider, scanner, error);
cssprovider: Handle errors completely via new error functions
2011-04-07 12:03:08 +00:00
}
cssprovider: Add a function for signalling errors
2011-04-06 13:38:06 +00:00
static void
gtk_css_provider_invalid_token (GtkCssProvider *provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner,
cssprovider: Add a function for signalling errors
2011-04-06 13:38:06 +00:00
const char *expected)
{
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
gtk_css_provider_error (provider,
cssprovider: Pass scanner argument explicitly to error functions
2011-04-10 13:22:07 +00:00
scanner,
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
GTK_CSS_PROVIDER_ERROR,
cssprovider: use INVALID_SYNTAX for unexpected token error
2011-04-07 22:18:38 +00:00
GTK_CSS_PROVIDER_ERROR_SYNTAX,
cssprovider: Provide better API to set errors
2011-04-07 11:37:03 +00:00
"expected a valid %s", expected);
cssprovider: Add a function for signalling errors
2011-04-06 13:38:06 +00:00
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static void
cssprovider: Store current properties and selectors in scanner They don't belong in the css provider
2011-04-10 23:28:00 +00:00
css_provider_commit (GtkCssProvider *css_provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GSList *selectors,
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
GtkCssRuleset *ruleset)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
GtkCssProviderPrivate *priv;
GSList *l;
GtkCssProvider: Add priv pointer to object struct.
2010-07-03 00:19:27 +00:00
priv = css_provider->priv;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
if (ruleset->styles == NULL && ruleset->widget_style == NULL)
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
{
g_slist_free_full (selectors, (GDestroyNotify) _gtk_css_selector_free);
return;
}
cssprovider: Ignore empty rules
2011-04-07 00:15:02 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
for (l = selectors; l; l = l->next)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
GtkCssRuleset new;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
gtk_css_ruleset_init_copy (&new, ruleset, l->data);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
g_array_append_val (priv->rulesets, new);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
g_slist_free (selectors);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static void
gtk_css_provider_reset (GtkCssProvider *css_provider)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssProviderPrivate *priv;
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
guint i;
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
priv = css_provider->priv;
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
css: Remove any loaded resource in gtk_css_provider_reset
2012-02-03 15:01:51 +00:00
if (priv->resource)
{
g_resources_unregister (priv->resource);
g_resource_unref (priv->resource);
priv->resource = NULL;
}
cssprovider: Remove colors when resetting
2011-05-27 23:01:42 +00:00
g_hash_table_remove_all (priv->symbolic_colors);
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
for (i = 0; i < priv->rulesets->len; i++)
gtk_css_ruleset_clear (&g_array_index (priv->rulesets, GtkCssRuleset, i));
g_array_set_size (priv->rulesets, 0);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static void
gtk_css_provider_propagate_error (GtkCssProvider *provider,
Change gtk_css_provider_propagate_error's signature. Update to new parsing-error signal. https://bugzilla.gnome.org/show_bug.cgi?id=657882
2011-08-31 22:24:53 +00:00
GtkCssSection *section,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
const GError *error,
GError **propagate_to)
{
Change gtk_css_provider_propagate_error's signature. Update to new parsing-error signal. https://bugzilla.gnome.org/show_bug.cgi?id=657882
2011-08-31 22:24:53 +00:00
GFileInfo *info;
GFile *file;
const char *path;
file = gtk_css_section_get_file (section);
if (file)
{
info = g_file_query_info (file,G_FILE_ATTRIBUTE_STANDARD_DISPLAY_NAME, 0, NULL, NULL);
if (info)
path = g_file_info_get_display_name (info);
else
path = "<broken file>";
}
else
{
info = NULL;
path = "<unknown>";
}
cssprovider: Don't propagate deprecation messages to passed-in error It's an informative message, don't stop parsing the CSS because of it.
2011-05-18 16:35:37 +00:00
/* don't fail for deprecations */
if (g_error_matches (error, GTK_CSS_PROVIDER_ERROR, GTK_CSS_PROVIDER_ERROR_DEPRECATED))
{
Change gtk_css_provider_propagate_error's signature. Update to new parsing-error signal. https://bugzilla.gnome.org/show_bug.cgi?id=657882
2011-08-31 22:24:53 +00:00
g_warning ("Theme parsing error: %s:%u:%u: %s", path,
gtk_css_section_get_end_line (section) + 1,
gtk_css_section_get_end_position (section), error->message);
cssprovider: Don't propagate deprecation messages to passed-in error It's an informative message, don't stop parsing the CSS because of it.
2011-05-18 16:35:37 +00:00
return;
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
/* we already set an error. And we'd like to keep the first one */
if (*propagate_to)
return;
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
*propagate_to = g_error_copy (error);
Change gtk_css_provider_propagate_error's signature. Update to new parsing-error signal. https://bugzilla.gnome.org/show_bug.cgi?id=657882
2011-08-31 22:24:53 +00:00
g_prefix_error (propagate_to, "%s:%u:%u: ", path,
gtk_css_section_get_end_line (section) + 1,
gtk_css_section_get_end_position (section));
if (info)
g_object_unref (info);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
static gboolean
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_import (GtkCssScanner *scanner)
{
GFile *file;
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_IMPORT);
if (!_gtk_css_parser_try (scanner->parser, "@import", TRUE))
{
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_IMPORT);
return FALSE;
}
cssprovider: @import <string> is valid CSS The parser rewrite insisted on @import url(<string>) syntax, but CSS xplicitly allows omitting the url string.
2011-07-19 23:19:59 +00:00
if (_gtk_css_parser_is_string (scanner->parser))
provider: use _gtk_css_parse_read_url() in parse_import() Instead of _gtk_css_parser_read_uri(), which is going away.
2012-01-13 23:08:12 +00:00
{
char *uri;
uri = _gtk_css_parser_read_string (scanner->parser);
cssprovider: Remove unused members from scanner Now that the parser takes care of the file, we don't need to keep track of it ourselves anymore.
2012-04-18 20:19:30 +00:00
file = _gtk_css_parser_get_file_for_path (scanner->parser, uri);
provider: use _gtk_css_parse_read_url() in parse_import() Instead of _gtk_css_parser_read_uri(), which is going away.
2012-01-13 23:08:12 +00:00
g_free (uri);
}
cssprovider: @import <string> is valid CSS The parser rewrite insisted on @import url(<string>) syntax, but CSS xplicitly allows omitting the url string.
2011-07-19 23:19:59 +00:00
else
provider: use _gtk_css_parse_read_url() in parse_import() Instead of _gtk_css_parser_read_uri(), which is going away.
2012-01-13 23:08:12 +00:00
{
cssparser: Change APIs that work with files Now that we keep track of the parser's file, we don't require it to be passed anymore.
2012-04-18 17:43:24 +00:00
file = _gtk_css_parser_read_url (scanner->parser);
provider: use _gtk_css_parse_read_url() in parse_import() Instead of _gtk_css_parser_read_uri(), which is going away.
2012-01-13 23:08:12 +00:00
}
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
provider: use _gtk_css_parse_read_url() in parse_import() Instead of _gtk_css_parser_read_uri(), which is going away.
2012-01-13 23:08:12 +00:00
if (file == NULL)
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_IMPORT);
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
cssprovider: Improve error handling for import errors We don't want to include the whitespace after an import statement. Because people tend to add newlines after @import, we report the wrong lines for failed imports. And that's bad.
2012-01-16 17:44:43 +00:00
if (!_gtk_css_parser_try (scanner->parser, ";", FALSE))
css: Don't import CSS too early First, finish parsing the import rule including the semicolon. Then try to actually import. test: css/parser/at-invalid-01.css
2012-01-15 18:20:34 +00:00
{
gtk_css_provider_invalid_token (scanner->provider, scanner, "semicolon");
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
}
else if (gtk_css_scanner_would_recurse (scanner, file))
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
char *path = g_file_get_path (file);
gtk_css_provider_error (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_IMPORT,
"Loading '%s' would recurse",
path);
g_free (path);
}
else
{
gtk_css_provider_load_internal (scanner->provider,
scanner,
file,
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
NULL,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
NULL);
}
GtkCssProvider: Parse color mappings. These are of the form: @color-a: #fff; @color-b: mix (@color-a, #000, 0.5) @color-c: shade (@color-b, 0.7) In the stylesheet, arbitrary nesting of these commands is handled.
2010-06-25 18:27:15 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
g_object_unref (file);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_IMPORT);
cssprovider: Improve error handling for import errors We don't want to include the whitespace after an import statement. Because people tend to add newlines after @import, we report the wrong lines for failed imports. And that's bad.
2012-01-16 17:44:43 +00:00
_gtk_css_parser_skip_whitespace (scanner->parser);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
GtkCssProvider: Transform custom colors to a @define-color rule. Given there are other rules such as @import (which will be supported), keep a sane namespace here, so for (re)defining a color name, in the CSS file it will look like: @define-color color-name #fff; @define-color other-color mix (@color-name, #f00, 0.4);
2010-10-08 16:19:34 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
static gboolean
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_color_definition (GtkCssScanner *scanner)
{
symboliccolor: Treat it as a CssValue .. now that it is one.
2012-04-07 04:18:03 +00:00
GtkCssValue *symbolic;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
char *name;
GtkCssProvider: Transform custom colors to a @define-color rule. Given there are other rules such as @import (which will be supported), keep a sane namespace here, so for (re)defining a color name, in the CSS file it will look like: @define-color color-name #fff; @define-color other-color mix (@color-name, #f00, 0.4);
2010-10-08 16:19:34 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_COLOR_DEFINITION);
if (!_gtk_css_parser_try (scanner->parser, "@define-color", TRUE))
{
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_COLOR_DEFINITION);
return FALSE;
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
name = _gtk_css_parser_try_name (scanner->parser, TRUE);
if (name == NULL)
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Not a valid color name");
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_COLOR_DEFINITION);
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
symboliccolor: Treat it as a CssValue .. now that it is one.
2012-04-07 04:18:03 +00:00
symbolic = _gtk_css_symbolic_value_new (scanner->parser);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (symbolic == NULL)
{
g_free (name);
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_COLOR_DEFINITION);
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, ";", TRUE))
{
g_free (name);
symboliccolor: Treat it as a CssValue .. now that it is one.
2012-04-07 04:18:03 +00:00
_gtk_css_value_unref (symbolic);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Missing semicolon at end of color definition");
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_COLOR_DEFINITION);
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
Work harder at parsing import statements Also, turn g_warning() into g_message() for now, so that tests don't abort. Eventually, this needs to all go into GErrors.
2010-11-18 06:59:27 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
g_hash_table_insert (scanner->provider->priv->symbolic_colors, name, symbolic);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_COLOR_DEFINITION);
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
css: Overhaul value parsing Value parsing only sometimes emitted errors. Sometimes it didn't emit errors but ignored the value, sometimes it took a default, sometimes it converted it to something it deemed suitable. While refactoring, I moved the whole GValue <=> char * conversion routines to a separate file, to make navigating the core css provider easier.
2011-04-08 14:08:28 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
static gboolean
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_binding_set (GtkCssScanner *scanner)
{
GtkBindingSet *binding_set;
char *name;
css: Overhaul value parsing Value parsing only sometimes emitted errors. Sometimes it didn't emit errors but ignored the value, sometimes it took a default, sometimes it converted it to something it deemed suitable. While refactoring, I moved the whole GValue <=> char * conversion routines to a separate file, to make navigating the core css provider easier.
2011-04-08 14:08:28 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_BINDING_SET);
if (!_gtk_css_parser_try (scanner->parser, "@binding-set", TRUE))
{
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_BINDING_SET);
return FALSE;
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
name = _gtk_css_parser_try_ident (scanner->parser, TRUE);
if (name == NULL)
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Expected name for binding set");
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
goto skip_semicolon;
}
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
binding_set = gtk_binding_set_find (name);
if (!binding_set)
{
binding_set = gtk_binding_set_new (name);
binding_set->parsed = TRUE;
}
g_free (name);
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, "{", TRUE))
{
gtk_css_provider_error_literal (scanner->provider,
cssprovider: Add recursion checks for @import This is pretty important, because otherwise recursions cause crashes. And if you accidentally change your theme to one that crashes on load, all your gonna SEGV and then on reboot, gdm tries to load the theme...
2011-04-12 00:50:25 +00:00
scanner,
GTK_CSS_PROVIDER_ERROR,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Expected '{' for binding set");
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
goto skip_semicolon;
}
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
while (!_gtk_css_parser_is_eof (scanner->parser) &&
!_gtk_css_parser_begins_with (scanner->parser, '}'))
{
name = _gtk_css_parser_read_value (scanner->parser);
if (name == NULL)
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
continue;
}
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
cssprovider: Emit an error when parsing a binding fails
2011-07-20 00:11:51 +00:00
if (gtk_binding_entry_add_signal_from_string (binding_set, name) != G_TOKEN_NONE)
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Failed to parse binding set.");
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
g_free (name);
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, ";", TRUE))
{
if (!_gtk_css_parser_begins_with (scanner->parser, '}') &&
!_gtk_css_parser_is_eof (scanner->parser))
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Expected semicolon");
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
}
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, "}", TRUE))
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"expected '}' after declarations");
if (!_gtk_css_parser_is_eof (scanner->parser))
_gtk_css_parser_resync (scanner->parser, FALSE, 0);
}
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
skip_semicolon:
cssprovider: Get the location right for semicolon warnings
2011-07-20 00:01:34 +00:00
if (_gtk_css_parser_begins_with (scanner->parser, ';'))
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_DEPRECATED,
"Nonstandard semicolon at end of binding set");
_gtk_css_parser_try (scanner->parser, ";", TRUE);
}
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_BINDING_SET);
return TRUE;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static void
parse_at_keyword (GtkCssScanner *scanner)
{
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
if (parse_import (scanner))
return;
if (parse_color_definition (scanner))
return;
if (parse_binding_set (scanner))
return;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
else
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"unknown @ rule");
_gtk_css_parser_resync (scanner->parser, TRUE, 0);
}
}
Add @binding-set rules to the CSS parser these custom rules can be used to define key themes as in the RC files, a series of bind/unbind rules may be added then: @binding foo-bar { bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) }; unbind "<ctl>d" }
2011-01-28 00:52:03 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static GSList *
parse_selector_list (GtkCssScanner *scanner)
{
GSList *selectors = NULL;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_SELECTOR);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
do {
css: Move selector parsing code into a custom function
2012-02-15 14:42:00 +00:00
GtkCssSelector *select = _gtk_css_selector_parse (scanner->parser);
cssprovider: Clean up declaration parser to conform to more tests The tests will follow with the next commits
2011-04-09 23:48:55 +00:00
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
if (select == NULL)
cssprovider: Clean up declaration parser to conform to more tests The tests will follow with the next commits
2011-04-09 23:48:55 +00:00
{
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
g_slist_free_full (selectors, (GDestroyNotify) _gtk_css_selector_free);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
_gtk_css_parser_resync (scanner->parser, FALSE, 0);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_SELECTOR);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
return NULL;
cssprovider: Clean up declaration parser to conform to more tests The tests will follow with the next commits
2011-04-09 23:48:55 +00:00
}
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
selectors = g_slist_prepend (selectors, select);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
while (_gtk_css_parser_try (scanner->parser, ",", TRUE));
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_SELECTOR);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
return selectors;
}
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
static void
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
parse_declaration (GtkCssScanner *scanner,
GtkCssRuleset *ruleset)
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
styleproperty: unconstify GtkStyleProperty is a real GObject now, so treat it like one and don't use const.
2011-12-31 15:08:23 +00:00
GtkStyleProperty *property;
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
char *name;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_DECLARATION);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
name = _gtk_css_parser_try_ident (scanner->parser, TRUE);
if (name == NULL)
goto check_for_semicolon;
cssprovider: Use _gtk_style_property_lookup() This way we get the actual data node and can use it.
2011-05-19 12:04:51 +00:00
property = _gtk_style_property_lookup (name);
if (property == NULL && name[0] != '-')
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
gtk_css_provider_error (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_NAME,
"'%s' is not a valid property name",
name);
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
g_free (name);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DECLARATION);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
return;
}
if (!_gtk_css_parser_try (scanner->parser, ":", TRUE))
{
gtk_css_provider_invalid_token (scanner->provider, scanner, "':'");
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
g_free (name);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DECLARATION);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
return;
}
cssprovider: Use _gtk_style_property_lookup() This way we get the actual data node and can use it.
2011-05-19 12:04:51 +00:00
if (property)
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
GtkCssValue *value;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
cssprovider: Add a new setter for rulesets We now have set-by-pspec (which is used for normal style properties) and set-by-name (which is used by widget properties).
2011-05-18 02:30:51 +00:00
g_free (name);
cssprovider: Store sections with parsed values Also create a separate section for values to associate the values with. Finally, use this section information when delayed-parsing a widget style property.
2011-06-17 03:58:04 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_VALUE);
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
value = _gtk_style_property_parse_value (property,
styleproperty: Remove base argument from _gtk_style_property_parse_value(). The GtkCssParser takes care of that now.
2012-04-18 20:04:44 +00:00
scanner->parser);
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
if (value == NULL)
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_VALUE);
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DECLARATION);
return;
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
if (!_gtk_css_parser_begins_with (scanner->parser, ';') &&
!_gtk_css_parser_begins_with (scanner->parser, '}') &&
!_gtk_css_parser_is_eof (scanner->parser))
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Junk at end of value");
css: Move property parsing into styleproperty file Just shuffles code around for parsing properties.
2011-05-25 23:25:00 +00:00
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
cssprovider: Store sections with parsed values Also create a separate section for values to associate the values with. Finally, use this section information when delayed-parsing a widget style property.
2011-06-17 03:58:04 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_VALUE);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DECLARATION);
css: Move property parsing into styleproperty file Just shuffles code around for parsing properties.
2011-05-25 23:25:00 +00:00
return;
GtkCssProvider: Implement GtkStyleProvider->get_style_property Style properties in the resource files are something like: -GtkComboBox-appears-as-string -GtkEntry-icon-prelight Following the usual naming spec for extensions to CSS.
2010-06-18 15:04:03 +00:00
}
styleproperty: Make _gtk_style_property_parse_value() return a CssValue Also split out initial/inherit handling into a custom GtkCssValue class.
2012-03-26 15:24:02 +00:00
if (GTK_IS_CSS_SHORTHAND_PROPERTY (property))
{
GtkCssShorthandProperty *shorthand = GTK_CSS_SHORTHAND_PROPERTY (property);
guint i;
for (i = 0; i < _gtk_css_shorthand_property_get_n_subproperties (shorthand); i++)
{
GtkCssStyleProperty *child = _gtk_css_shorthand_property_get_subproperty (shorthand, i);
GtkCssValue *sub = _gtk_css_array_value_get_nth (value, i);
gtk_css_ruleset_add (ruleset, child, _gtk_css_value_ref (sub), scanner->section);
}
_gtk_css_value_unref (value);
}
else if (GTK_IS_CSS_STYLE_PROPERTY (property))
{
gtk_css_ruleset_add (ruleset, GTK_CSS_STYLE_PROPERTY (property), value, scanner->section);
}
else
{
g_assert_not_reached ();
_gtk_css_value_unref (value);
}
cssprovider: Store sections with parsed values Also create a separate section for values to associate the values with. Finally, use this section information when delayed-parsing a widget style property.
2011-06-17 03:58:04 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_VALUE);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
else if (name[0] == '-')
{
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
char *value_str;
GtkThemingEngine: Add gtk_theming_engine_register_property(). This function may be used for custom property registration from theming engines. The property names will have the -${engine-type-name}-${prop-name} format, the parser has been modified to allow properties with '-' as the first char.
2010-08-04 15:51:19 +00:00
cssprovider: Store sections with parsed values Also create a separate section for values to associate the values with. Finally, use this section information when delayed-parsing a widget style property.
2011-06-17 03:58:04 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_VALUE);
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
value_str = _gtk_css_parser_read_value (scanner->parser);
if (value_str)
{
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
WidgetPropertyValue *val;
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
val = widget_property_value_new (name, scanner->section);
cssprovider: Store widget style value as string There's no need to turn it into a GtkCssValue, it's always a string.
2012-03-30 00:45:01 +00:00
val->value = value_str;
GtkThemingEngine: Add gtk_theming_engine_register_property(). This function may be used for custom property registration from theming engines. The property names will have the -${engine-type-name}-${prop-name} format, the parser has been modified to allow properties with '-' as the first char.
2010-08-04 15:51:19 +00:00
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
gtk_css_ruleset_add_style (ruleset, name, val);
}
else
{
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
cssprovider: Store sections with parsed values Also create a separate section for values to associate the values with. Finally, use this section information when delayed-parsing a widget style property.
2011-06-17 03:58:04 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_VALUE);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DECLARATION);
css: Make property parsing functions take a css parser Instead of reading a string and then passing that in, let the parse functions use the full power of the parser.
2011-05-18 16:32:22 +00:00
return;
}
cssprovider: Store sections with parsed values Also create a separate section for values to associate the values with. Finally, use this section information when delayed-parsing a widget style property.
2011-06-17 03:58:04 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_VALUE);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
else
g_free (name);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
check_for_semicolon:
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DECLARATION);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, ";", TRUE))
{
if (!_gtk_css_parser_begins_with (scanner->parser, '}') &&
!_gtk_css_parser_is_eof (scanner->parser))
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"Expected semicolon");
_gtk_css_parser_resync (scanner->parser, TRUE, '}');
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
static void
parse_declarations (GtkCssScanner *scanner,
GtkCssRuleset *ruleset)
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
{
while (!_gtk_css_parser_is_eof (scanner->parser) &&
!_gtk_css_parser_begins_with (scanner->parser, '}'))
{
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
parse_declaration (scanner, ruleset);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
static void
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_ruleset (GtkCssScanner *scanner)
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GSList *selectors;
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
GtkCssRuleset ruleset = { 0, };
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_RULESET);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
selectors = parse_selector_list (scanner);
if (selectors == NULL)
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
{
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_RULESET);
return;
}
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, "{", TRUE))
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"expected '{' after selectors");
_gtk_css_parser_resync (scanner->parser, FALSE, 0);
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
g_slist_free_full (selectors, (GDestroyNotify) _gtk_css_selector_free);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_RULESET);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
return;
}
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
parse_declarations (scanner, &ruleset);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (!_gtk_css_parser_try (scanner->parser, "}", TRUE))
{
gtk_css_provider_error_literal (scanner->provider,
scanner,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_SYNTAX,
"expected '}' after declarations");
if (!_gtk_css_parser_is_eof (scanner->parser))
{
_gtk_css_parser_resync (scanner->parser, FALSE, 0);
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
g_slist_free_full (selectors, (GDestroyNotify) _gtk_css_selector_free);
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
gtk_css_ruleset_clear (&ruleset);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_RULESET);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
}
}
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
css_provider_commit (scanner->provider, selectors, &ruleset);
gtk_css_ruleset_clear (&ruleset);
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_RULESET);
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
}
cssprovider: Keep track of passed in error differently Instead of having an error member in the CSS provider's private struct, connect a signal handler when an error is passed in. This has two advantages: 1) It makes the code clearer as we don't have to keep track of an error member anywhere. 2) It causes a non-emission of the g_warning() when an error was passed in, because it only triggers when no signal handlers are connected. So we get identical behavior to GTK 3.0 where warnings where only emitted when no error was passed in.
2011-04-11 00:32:26 +00:00
static void
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_statement (GtkCssScanner *scanner)
cssprovider: Keep track of passed in error differently Instead of having an error member in the CSS provider's private struct, connect a signal handler when an error is passed in. This has two advantages: 1) It makes the code clearer as we don't have to keep track of an error member anywhere. 2) It causes a non-emission of the g_warning() when an error was passed in, because it only triggers when no signal handlers are connected. So we get identical behavior to GTK 3.0 where warnings where only emitted when no error was passed in.
2011-04-11 00:32:26 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (_gtk_css_parser_begins_with (scanner->parser, '@'))
parse_at_keyword (scanner);
else
parse_ruleset (scanner);
cssprovider: Keep track of passed in error differently Instead of having an error member in the CSS provider's private struct, connect a signal handler when an error is passed in. This has two advantages: 1) It makes the code clearer as we don't have to keep track of an error member anywhere. 2) It causes a non-emission of the g_warning() when an error was passed in, because it only triggers when no signal handlers are connected. So we get identical behavior to GTK 3.0 where warnings where only emitted when no error was passed in.
2011-04-11 00:32:26 +00:00
}
cssprovider: Remove now useless variables
2011-04-08 00:57:33 +00:00
cssprovider: More error handling into load_internal() Now the parsing functions starting at parse_stylesheet() don't have to care about errors anymore.
2011-04-11 17:57:42 +00:00
static void
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
parse_stylesheet (GtkCssScanner *scanner)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_push_section (scanner, GTK_CSS_SECTION_DOCUMENT);
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
_gtk_css_parser_skip_whitespace (scanner->parser);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
while (!_gtk_css_parser_is_eof (scanner->parser))
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
if (_gtk_css_parser_try (scanner->parser, "<!--", TRUE) ||
_gtk_css_parser_try (scanner->parser, "-->", TRUE))
continue;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_statement (scanner);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
API: css: Add GtkCssSection This struct keeps track of an area of text in a CSS file and uses it when specifying information. Also, the cssprovider keeps track of sections when parsing a file.
2011-06-17 00:02:24 +00:00
gtk_css_scanner_pop_section (scanner, GTK_CSS_SECTION_DOCUMENT);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
static int
gtk_css_provider_compare_rule (gconstpointer a_,
gconstpointer b_)
{
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
const GtkCssRuleset *a = (const GtkCssRuleset *) a_;
const GtkCssRuleset *b = (const GtkCssRuleset *) b_;
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
int compare;
compare = _gtk_css_selector_compare (a->selector, b->selector);
if (compare != 0)
return compare;
return 0;
}
static void
gtk_css_provider_postprocess (GtkCssProvider *css_provider)
{
GtkCssProviderPrivate *priv = css_provider->priv;
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
g_array_sort (priv->rulesets, gtk_css_provider_compare_rule);
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
}
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
static gboolean
gtk_css_provider_load_internal (GtkCssProvider *css_provider,
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *parent,
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
GFile *file,
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
const char *text,
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
GError **error)
{
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
GtkCssScanner *scanner;
cssprovider: More error handling into load_internal() Now the parsing functions starting at parse_stylesheet() don't have to care about errors anymore.
2011-04-11 17:57:42 +00:00
gulong error_handler;
gtkcssprovider: plug several leaks In finalize(), clear all rulesets. In parse_declaration(), Free the GValue under unhandled error situations. In gtk_css_provider_load_internal(), Do not leak the file contents.
2011-05-30 17:26:28 +00:00
char *free_data = NULL;
cssprovider: More error handling into load_internal() Now the parsing functions starting at parse_stylesheet() don't have to care about errors anymore.
2011-04-11 17:57:42 +00:00
if (error)
error_handler = g_signal_connect (css_provider,
"parsing-error",
G_CALLBACK (gtk_css_provider_propagate_error),
error);
else
error_handler = 0; /* silence gcc */
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
if (text == NULL)
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
{
GError *load_error = NULL;
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
if (g_file_load_contents (file, NULL,
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
&free_data, NULL,
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
NULL, &load_error))
{
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
text = free_data;
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
}
else
{
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
GtkCssSection *section;
cssprovider: Allow storing the parent stylesheet
2011-04-12 00:06:08 +00:00
if (parent)
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
section = gtk_css_section_ref (parent->section);
cssprovider: Allow storing the parent stylesheet
2011-04-12 00:06:08 +00:00
else
API: cssprovider: Change parsing-error signal Instead of path, line and position, emit the section the error happened in. This has a lot more information to consume.
2011-06-17 05:39:11 +00:00
section = _gtk_css_section_new_for_file (GTK_CSS_SECTION_DOCUMENT, file);
gtk_css_provider_error (css_provider,
parent,
GTK_CSS_PROVIDER_ERROR,
GTK_CSS_PROVIDER_ERROR_IMPORT,
"Failed to import: %s",
load_error->message);
gtk_css_section_unref (section);
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
}
}
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
if (text)
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
{
cssprovider: Keep only the curent section while parsing We don't need a list of sections, the section can give us the parent when we need it (d'oh).
2011-06-17 02:52:43 +00:00
scanner = gtk_css_scanner_new (css_provider,
parent,
parent ? parent->section : NULL,
file,
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
text);
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
css: Rewrite the parser Instead of relying on GScanner and its idea of syntax, code up a parser that obeys the CSS spec. This also has the great side effect of reporting correct line numbers and positions. Also included is a reorganization of the returned error values. Instead of error values describing what type of syntax error was returned, the code just returns SYNTAX_ERROR. Other messages exist for when actual values don't work or when errors shouldn't be fatal due to backwards compatibility.
2011-04-14 02:47:18 +00:00
parse_stylesheet (scanner);
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
gtk_css_scanner_destroy (scanner);
css: Rewrite selectors Selectors now go into their own C file. The new selectors are modeled a lot closer to the CSS spec. In particular the specificity computation matches CSS 2.1 exactly. For details about the why, see also: http://mail.gnome.org/archives/gtk-devel-list/2011-May/msg00061.html https://bugzilla.gnome.org/show_bug.cgi?id=649798
2011-05-14 11:27:31 +00:00
if (parent == NULL)
gtk_css_provider_postprocess (css_provider);
cssprovider: Make load_internal() load the contents This way, we achieve two things: 1) We can unify file loading to one location 2) We can emit the error from file loading using the parsing-error signal. This is very useful for @import handling in particular.
2011-04-11 18:08:57 +00:00
}
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
gtkcssprovider: plug several leaks In finalize(), clear all rulesets. In parse_declaration(), Free the GValue under unhandled error situations. In gtk_css_provider_load_internal(), Do not leak the file contents.
2011-05-30 17:26:28 +00:00
g_free (free_data);
cssprovider: More error handling into load_internal() Now the parsing functions starting at parse_stylesheet() don't have to care about errors anymore.
2011-04-11 17:57:42 +00:00
if (error)
{
g_signal_handler_disconnect (css_provider, error_handler);
gtkcssprovider: plug several leaks In finalize(), clear all rulesets. In parse_declaration(), Free the GValue under unhandled error situations. In gtk_css_provider_load_internal(), Do not leak the file contents.
2011-05-30 17:26:28 +00:00
cssprovider: More error handling into load_internal() Now the parsing functions starting at parse_stylesheet() don't have to care about errors anymore.
2011-04-11 17:57:42 +00:00
if (*error)
{
/* We clear all contents from the provider for backwards compat reasons */
gtk_css_provider_reset (css_provider);
return FALSE;
}
}
return TRUE;
cssprovider: Add a load_internal function that all loads chain to
2011-04-11 17:55:34 +00:00
}
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* gtk_css_provider_load_from_data:
* @css_provider: a #GtkCssProvider
Add annotations to gtk_css_provider_load_from_data https://bugzilla.gnome.org/show_bug.cgi?id=646859
2011-06-01 00:37:43 +00:00
* @data: (array length=length) (element-type guint8): CSS data loaded in memory
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
* @length: the length of @data in bytes, or -1 for NUL terminated strings. If
* @length is not -1, the code will assume it is not NUL terminated and will
* potentially do a copy.
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
* @error: (out) (allow-none): return location for a #GError, or %NULL
*
* Loads @data into @css_provider, making it clear any previously loaded
* information.
*
* Returns: %TRUE if the data could be loaded.
**/
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
gboolean
Fix prototype of gtk_css_style_provider_load_from_data GError arguments must be **.
2010-11-18 01:24:18 +00:00
gtk_css_provider_load_from_data (GtkCssProvider *css_provider,
const gchar *data,
gssize length,
GError **error)
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
{
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
char *free_data;
gboolean ret;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
g_return_val_if_fail (GTK_IS_CSS_PROVIDER (css_provider), FALSE);
g_return_val_if_fail (data != NULL, FALSE);
if (length < 0)
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
{
length = strlen (data);
free_data = NULL;
}
else
{
free_data = g_strndup (data, length);
data = free_data;
}
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
gtk_css_provider_reset (css_provider);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
ret = gtk_css_provider_load_internal (css_provider, NULL, NULL, data, error);
g_free (free_data);
styleprovider: Make all implementors emit the new signal ... and listen to it in the cases where they use other style providers.
2012-04-08 15:35:24 +00:00
_gtk_style_provider_private_changed (GTK_STYLE_PROVIDER_PRIVATE (css_provider));
cssprovider: Handle non-NULL-terminated data Do a dumb copy if that happens.
2011-06-17 03:15:48 +00:00
return ret;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* gtk_css_provider_load_from_file:
* @css_provider: a #GtkCssProvider
* @file: #GFile pointing to a file to load
* @error: (out) (allow-none): return location for a #GError, or %NULL
*
* Loads the data contained in @file into @css_provider, making it
* clear any previously loaded information.
*
* Returns: %TRUE if the data could be loaded.
**/
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
gboolean
gtk_css_provider_load_from_file (GtkCssProvider *css_provider,
GFile *file,
GError **error)
{
styleprovider: Make all implementors emit the new signal ... and listen to it in the cases where they use other style providers.
2012-04-08 15:35:24 +00:00
gboolean success;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
g_return_val_if_fail (GTK_IS_CSS_PROVIDER (css_provider), FALSE);
g_return_val_if_fail (G_IS_FILE (file), FALSE);
cssprovider: Change declaration parsing for error reporting Previously, we only checked for errors after parsing the full declaration. Now we detect errors with the property before even attempting to parse its value. The benefit here is that the error reporting reports the correct line and position numbers.
2011-04-07 22:08:10 +00:00
gtk_css_provider_reset (css_provider);
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
styleprovider: Make all implementors emit the new signal ... and listen to it in the cases where they use other style providers.
2012-04-08 15:35:24 +00:00
success = gtk_css_provider_load_internal (css_provider, NULL, file, NULL, error);
_gtk_style_provider_private_changed (GTK_STYLE_PROVIDER_PRIVATE (css_provider));
return success;
Added GtkCssProvider. GtkCssProvider is a GtkStyleProvider implementation which is able to read CSS-like input to style widgets. At the moment, only the basic parser features are available.
2010-04-11 21:03:25 +00:00
}
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* gtk_css_provider_load_from_path:
* @css_provider: a #GtkCssProvider
* @path: the path of a filename to load, in the GLib filename encoding
* @error: (out) (allow-none): return location for a #GError, or %NULL
*
* Loads the data contained in @path into @css_provider, making it clear
* any previously loaded information.
*
* Returns: %TRUE if the data could be loaded.
**/
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
gboolean
gtk_css_provider_load_from_path (GtkCssProvider *css_provider,
const gchar *path,
GError **error)
{
cssprovider: Reimplement gtk_css_provider_load_from_path() Call gtk_css_provider_load_from_file() instead of the internal function. This has two advantages: 1) It simplifies the code a lot 2) It gets rid of GMappedFile usage. GMappedFile does not work everywhere, so this is finally portable.
2011-04-11 23:22:16 +00:00
GFile *file;
gboolean result;
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
g_return_val_if_fail (GTK_IS_CSS_PROVIDER (css_provider), FALSE);
g_return_val_if_fail (path != NULL, FALSE);
cssprovider: Reimplement gtk_css_provider_load_from_path() Call gtk_css_provider_load_from_file() instead of the internal function. This has two advantages: 1) It simplifies the code a lot 2) It gets rid of GMappedFile usage. GMappedFile does not work everywhere, so this is finally portable.
2011-04-11 23:22:16 +00:00
file = g_file_new_for_path (path);
result = gtk_css_provider_load_from_file (css_provider, file, error);
g_object_unref (file);
return result;
GtkCssProvider: Add support for @import rules Now other CSS files can be referenced from the currently parsed file: @import url (other-file.css); @import url (/some/file.css);
2010-10-08 16:36:46 +00:00
}
Add Gtk resource file with the default CSS file
2012-01-12 18:14:33 +00:00
static gboolean
_gtk_css_provider_load_from_resource (GtkCssProvider *css_provider,
const gchar *resource_path)
{
GFile *file;
char *uri, *escaped;
gboolean result;
g_return_val_if_fail (GTK_IS_CSS_PROVIDER (css_provider), FALSE);
g_return_val_if_fail (resource_path != NULL, FALSE);
escaped = g_uri_escape_string (resource_path,
G_URI_RESERVED_CHARS_ALLOWED_IN_PATH, FALSE);
uri = g_strconcat ("resource://", escaped, NULL);
g_free (escaped);
file = g_file_new_for_uri (uri);
g_free (uri);
result = gtk_css_provider_load_from_file (css_provider, file, NULL);
g_object_unref (file);
return result;
}
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* gtk_css_provider_get_default:
*
* Returns the provider containing the style settings used as a
* fallback for all widgets.
*
* Returns: (transfer none): The provider used for fallback styling.
* This memory is owned by GTK+, and you must not free it.
**/
GtkCssProvider: Add gtk_css_provider_get_default(). This method returns a GtkCssProvider with the default style information loaded.
2010-08-02 14:16:03 +00:00
GtkCssProvider *
gtk_css_provider_get_default (void)
{
static GtkCssProvider *provider;
if (G_UNLIKELY (!provider))
{
provider = gtk_css_provider_new ();
}
return provider;
}
Privately export css_provider_get_theme_dir
2011-01-26 16:17:34 +00:00
gchar *
_gtk_css_provider_get_theme_dir (void)
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
{
const gchar *var;
gchar *path;
var = g_getenv ("GTK_DATA_PREFIX");
if (var)
path = g_build_filename (var, "share", "themes", NULL);
else
gtk: clean up the private horror - add gtkmodulesprivate.h and move stuff there from gtkprivate.h - add gtkprivate.c and move stuff there from gtkmain.c - add gtkwin32.c and move stuff there from gtkmain.c - don't redefine GTK_DATADIR and friends in gtkprivate.h - have _gtk_get_datadir() and friends on all platforms - remove the horrid hacks where gtkprivate.h can't be included, or must be included later due to redefinition of the compile-time directories
2011-10-22 06:48:13 +00:00
path = g_build_filename (_gtk_get_data_prefix (), "share", "themes", NULL);
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
return path;
}
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
/**
* gtk_css_provider_get_named:
* @name: A theme name
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
* @variant: (allow-none): variant to load, for example, "dark", or
* %NULL for the default
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
*
Fix gtk_css_provider_get_named() docs.
2010-11-13 21:57:49 +00:00
* Loads a theme from the usual theme paths
Add theming docs Both API and file format is documented, there's still missing a migration doc though.
2010-10-19 08:59:46 +00:00
*
Fix gtk_css_provider_get_named() docs.
2010-11-13 21:57:49 +00:00
* Returns: (transfer none): a #GtkCssProvider with the theme loaded.
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
* This memory is owned by GTK+, and you must not free it.
*/
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
GtkCssProvider *
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
gtk_css_provider_get_named (const gchar *name,
const gchar *variant)
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
{
static GHashTable *themes = NULL;
GtkCssProvider *provider;
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
gchar *key;
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
if (variant == NULL)
key = (gchar *)name;
else
key = g_strconcat (name, "-", variant, NULL);
Remove all default css and make Raleigh a pure fallback theme It was problematic to maintain Raleigh going forward, as any changes in it affected all themes. Also, its more robust if each theme is a full standalone css rather than relying on an inherited css base. So, this changes Raleigh to a standalone theme that we can tweak without accidentally breaking other themes, and makes the default theme empty. In fact, we don't even add the default provider anymore as its always empty.
2012-02-09 22:16:03 +00:00
if (G_UNLIKELY (!themes))
themes = g_hash_table_new (g_str_hash, g_str_equal);
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
provider = g_hash_table_lookup (themes, key);
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
Remove all default css and make Raleigh a pure fallback theme It was problematic to maintain Raleigh going forward, as any changes in it affected all themes. Also, its more robust if each theme is a full standalone css rather than relying on an inherited css base. So, this changes Raleigh to a standalone theme that we can tweak without accidentally breaking other themes, and makes the default theme empty. In fact, we don't even add the default provider anymore as its always empty.
2012-02-09 22:16:03 +00:00
if (!provider)
{
gchar *resource_path = NULL;
if (variant)
resource_path = g_strdup_printf ("/org/gtk/libgtk/%s-%s.css", name, variant);
else
resource_path = g_strdup_printf ("/org/gtk/libgtk/%s.css", name);
if (g_resources_get_info (resource_path, 0, NULL, NULL, NULL))
{
provider = gtk_css_provider_new ();
if (!_gtk_css_provider_load_from_resource (provider, resource_path))
{
g_object_unref (provider);
provider = NULL;
}
}
g_free (resource_path);
}
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
if (!provider)
{
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
gchar *subpath, *path = NULL;
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
if (variant)
Hardcode 3.0 version for style files and paths. It's about time to start using the right paths.
2010-11-13 21:33:34 +00:00
subpath = g_strdup_printf ("gtk-3.0" G_DIR_SEPARATOR_S "gtk-%s.css", variant);
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
else
Hardcode 3.0 version for style files and paths. It's about time to start using the right paths.
2010-11-13 21:33:34 +00:00
subpath = g_strdup ("gtk-3.0" G_DIR_SEPARATOR_S "gtk.css");
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
Try to load css themes from the XDG config directory first Will fall back to previous locations. https://bugzilla.gnome.org/show_bug.cgi?id=646631
2012-04-27 16:32:18 +00:00
/* First look in the user's config directory
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
*/
Try to load css themes from the XDG config directory first Will fall back to previous locations. https://bugzilla.gnome.org/show_bug.cgi?id=646631
2012-04-27 16:32:18 +00:00
path = g_build_filename (g_get_user_data_dir (), "themes", name, subpath, NULL);
if (!g_file_test (path, G_FILE_TEST_EXISTS))
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
{
Try to load css themes from the XDG config directory first Will fall back to previous locations. https://bugzilla.gnome.org/show_bug.cgi?id=646631
2012-04-27 16:32:18 +00:00
g_free (path);
path = NULL;
}
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
Try to load css themes from the XDG config directory first Will fall back to previous locations. https://bugzilla.gnome.org/show_bug.cgi?id=646631
2012-04-27 16:32:18 +00:00
/* Next look in the user's home directory
*/
if (!path)
{
const gchar *home_dir;
home_dir = g_get_home_dir ();
if (home_dir)
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
{
Try to load css themes from the XDG config directory first Will fall back to previous locations. https://bugzilla.gnome.org/show_bug.cgi?id=646631
2012-04-27 16:32:18 +00:00
path = g_build_filename (home_dir, ".themes", name, subpath, NULL);
if (!g_file_test (path, G_FILE_TEST_EXISTS))
{
g_free (path);
path = NULL;
}
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
}
}
if (!path)
{
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
gchar *theme_dir;
theme_dir = _gtk_css_provider_get_theme_dir ();
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
path = g_build_filename (theme_dir, name, subpath, NULL);
g_free (theme_dir);
if (!g_file_test (path, G_FILE_TEST_EXISTS))
{
g_free (path);
path = NULL;
}
}
gtk_css_provider_get_named: don't leak subpath (#636777)
2010-12-08 14:13:21 +00:00
g_free (subpath);
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
if (path)
{
Free the resource file string.
2012-03-04 16:51:07 +00:00
char *dir, *resource_file;
GResource *resource;
Support a gtk.gresource file in themes This lets themes install a .gresource for efficient loading of resources. This resource file will (if it exists) be automatically loaded (mmaped) when the theme is used and unloaded when not used.
2012-02-03 14:50:49 +00:00
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
provider = gtk_css_provider_new ();
Free the resource file string.
2012-03-04 16:51:07 +00:00
dir = g_path_get_dirname (path);
resource_file = g_build_filename (dir, "gtk.gresource", NULL);
resource = g_resource_load (resource_file, NULL);
g_free (resource_file);
if (resource != NULL)
g_resources_register (resource);
Support a gtk.gresource file in themes This lets themes install a .gresource for efficient loading of resources. This resource file will (if it exists) be automatically loaded (mmaped) when the theme is used and unloaded when not used.
2012-02-03 14:50:49 +00:00
cssprovider: When loading themes, don't pass in an error We want to parse existing themes as well as possible instead of failing. And the g_warning() is preserved.
2011-05-18 16:37:23 +00:00
if (!gtk_css_provider_load_from_path (provider, path, NULL))
{
Fix crash with theme resource file The freeing the GResource in gtk_css_provider_reset caused problems because it was called from gtk_css_provider_load_from_path() inside gtk_css_provider_get_named(). We fix this by delaying the setting of priv->resource until we successfully loaded the file.
2012-02-03 15:37:54 +00:00
if (resource != NULL)
{
g_resources_unregister (resource);
g_resource_unref (resource);
}
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
g_object_unref (provider);
provider = NULL;
}
else
Fix crash with theme resource file The freeing the GResource in gtk_css_provider_reset caused problems because it was called from gtk_css_provider_load_from_path() inside gtk_css_provider_get_named(). We fix this by delaying the setting of priv->resource until we successfully loaded the file.
2012-02-03 15:37:54 +00:00
{
/* Only set this after load success, as load_from_path will clear it */
provider->priv->resource = resource;
g_hash_table_insert (themes, g_strdup (key), provider);
}
Fix a memory leak in GtkCSSProvider Found by Felix Riemann https://bugzilla.gnome.org/show_bug.cgi?id=641558
2011-02-05 02:13:56 +00:00
g_free (path);
Plug a small memory leak
2012-02-11 18:05:32 +00:00
g_free (dir);
GtkCssProvider: Add gtk_css_provider_get_named() This functions loads a CSS file from a installed theme.
2010-09-22 21:39:15 +00:00
}
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
}
GtkCssProvider: Take into account variant When loading a theme, make sure we take into account the variant so we don't use the plain version when the theme changes. Also make sure to fallback to the plain theme when loading a variant fails. https://bugzilla.gnome.org/show_bug.cgi?id=640983
2011-01-31 00:30:00 +00:00
if (key != name)
g_free (key);
GtkCssProvider: Add gtk_css_provider_load_from_path()
2010-08-20 09:50:39 +00:00
return provider;
}
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
static int
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
compare_properties (gconstpointer a, gconstpointer b, gpointer style)
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
{
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
const guint *ua = a;
const guint *ub = b;
PropertyValue *styles = style;
return strcmp (_gtk_style_property_get_name (GTK_STYLE_PROPERTY (styles[*ua].property)),
_gtk_style_property_get_name (GTK_STYLE_PROPERTY (styles[*ub].property)));
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
}
static int
compare_names (gconstpointer a, gconstpointer b)
{
const WidgetPropertyValue *aa = a;
const WidgetPropertyValue *bb = b;
return strcmp (aa->name, bb->name);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
}
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
static void
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
gtk_css_ruleset_print (const GtkCssRuleset *ruleset,
GString *str)
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
{
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
GList *values, *walk;
WidgetPropertyValue *widget_value;
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
guint i;
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
_gtk_css_selector_print (ruleset->selector, str);
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
g_string_append (str, " {\n");
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
if (ruleset->styles)
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
{
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
guint *sorted = g_new (guint, ruleset->n_styles);
for (i = 0; i < ruleset->n_styles; i++)
sorted[i] = i;
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
/* so the output is identical for identical selector styles */
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
g_qsort_with_data (sorted, ruleset->n_styles, sizeof (guint), compare_properties, ruleset->styles);
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
for (i = 0; i < ruleset->n_styles; i++)
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
{
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
PropertyValue *prop = &ruleset->styles[sorted[i]];
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
g_string_append (str, " ");
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
g_string_append (str, _gtk_style_property_get_name (GTK_STYLE_PROPERTY (prop->property)));
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
g_string_append (str, ": ");
cssstyleproperty: Get rid of unused API Both _gtk_css_style_property_print_value() and _gtk_css_style_property_compute_value() aren't necessary anymore and are replaced by _gtk_css_value_print() and _gtk_css_value_comptue() respectively.
2012-07-16 12:28:58 +00:00
_gtk_css_value_print (prop->value, str);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
g_string_append (str, ";\n");
}
cssprovider: Redo styles as an array Saves even more memory and avoids an extra copy during assigning.
2012-03-01 05:40:15 +00:00
g_free (sorted);
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
}
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
if (ruleset->widget_style)
{
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
values = NULL;
for (widget_value = ruleset->widget_style; widget_value != NULL; widget_value = widget_value->next)
values = g_list_prepend (values, widget_value);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
/* so the output is identical for identical selector styles */
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
values = g_list_sort (values, compare_names);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
for (walk = values; walk; walk = walk->next)
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
{
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
widget_value = walk->data;
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
g_string_append (str, " ");
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
g_string_append (str, widget_value->name);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
g_string_append (str, ": ");
cssprovider: Store widget style value as string There's no need to turn it into a GtkCssValue, it's always a string.
2012-03-30 00:45:01 +00:00
g_string_append (str, widget_value->value);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
g_string_append (str, ";\n");
}
Don't use hashtables for storing css rule properties This saves a lot of memory
2012-02-15 16:06:30 +00:00
g_list_free (values);
cssprovider: Keep two hash tables per ruleset One for the style properties, one for the widget style properties. This way we can make one hash table by pspec which means we don't have to repeat the pspec lookup.
2011-05-18 02:36:43 +00:00
}
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
g_string_append (str, "}\n");
}
static void
gtk_css_provider_print_colors (GHashTable *colors,
GString *str)
{
GList *keys, *walk;
char *s;
keys = g_hash_table_get_keys (colors);
/* so the output is identical for identical styles */
keys = g_list_sort (keys, (GCompareFunc) strcmp);
for (walk = keys; walk; walk = walk->next)
{
const char *name = walk->data;
GtkSymbolicColor *symbolic = g_hash_table_lookup (colors, (gpointer) name);
g_string_append (str, "@define-color ");
g_string_append (str, name);
g_string_append (str, " ");
s = gtk_symbolic_color_to_string (symbolic);
g_string_append (str, s);
g_free (s);
g_string_append (str, ";\n");
}
g_list_free (keys);
}
/**
* gtk_css_provider_to_string:
* @provider: the provider to write to a string
*
* Convertes the @provider into a string representation in CSS
* format.
*
* Using gtk_css_provider_load_from_data() with the return value
* from this function on a new provider created with
* gtk_css_provider_new() will basicallu create a duplicate of
* this @provider.
*
* Returns: a new string representing the @provider.
Add a Since tag gtk_css_provider_to_string is a post-3.0 addition. Fixes bug 659516
2011-12-30 05:38:06 +00:00
*
* Since: 3.2
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
**/
char *
gtk_css_provider_to_string (GtkCssProvider *provider)
{
GtkCssProviderPrivate *priv;
GString *str;
guint i;
g_return_val_if_fail (GTK_IS_CSS_PROVIDER (provider), NULL);
priv = provider->priv;
str = g_string_new ("");
gtk_css_provider_print_colors (priv->symbolic_colors, str);
cssprovider: Rename "SelectorStyleInfo" to "Ruleset" No code changes, just clarification of the name
2011-05-17 20:04:40 +00:00
for (i = 0; i < priv->rulesets->len; i++)
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
{
cssprovider: Add an empty line when printing ... between the @define-color statements and the rules.
2012-04-05 20:28:21 +00:00
if (str->len != 0)
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
g_string_append (str, "\n");
cssprovider: Refactor handling of rulesets Keep rulesets as an on-stack/heap structure instead of allocating all instances separately. Also, pass a ruleset to the ruleset parser, so we can make the ruleset parser do lots of fancy things that might be useful for performance.
2011-05-18 02:20:05 +00:00
gtk_css_ruleset_print (&g_array_index (priv->rulesets, GtkCssRuleset, i), str);
API: gtk: Add gtk_css_provider_to_string() This is intented to convert the contents of a CSS provider back to a string. It is not complete yet but good enough for starting a testsuite. :)
2011-04-06 20:03:35 +00:00
}
return g_string_free (str, FALSE);
}
Reference in New Issue Copy Permalink