2002-04-04 Tor Lillqvist <tml@iki.fi>
* gdk/gdkrgb.c: Fix some bugs, and introduce a minor feature.
(gdk_rgb_convert_4_pack): New function, for 16-color (4 bits per
pixel) static visuals (fixes#858).
(gdk_rgb_convert_gray4_pack, gdk_rgb_convert_gray4_d_pack): Fix
same bugs in both functions: Odd start coordinate (partial byte)
was not handled correctly. Also a partial final byte was not
handled correctly.
(gdk_rgb_do_colormaps): Use G_N_ELEMENTS.
(gdk_rgb_create_info): For pseudocolor visuals, use the 2x2x2
colorcube only for depths 3 and 4. For static color, use it for
depths 3..7 like before. (Depth 5..7 pseudocolor probably never
occurs on X11. It doesn't normally occur on Win32 either, but
there is experimental code in gdkvisual-win32.c to let the user
restrict the size of palette used.)
(gdk_rgb_init): Set gdk_rgb_verbose if the GDK_DEBUG_GDKRGB flag
is set.
(gdk_rgb_select_conv): Use gdk_rgb_convert_8 also for depths 5, 6
and 7 (see above). Use gdk_rgb_convert_4_pack for 4 bits per pixel
static color.
2002-03-29 Hans Breuer <hans@breuer.org>
Fixed dashed line issues (#74441) to an IMO reasonable extend.
That is: use PS_USERSTYLE on WinNT (the next GDI limit appears
to be with lines width > 50); Render horizontal and vertical
dashed lines on Win9x 'by hand'. Dotted selection rectangles
and Dia look nice ...
* gdk/win32/gdkprivate-win32.h : add pen_dashes pointer and
num_pen_dashes to _GdkGCWin32
* gdk/win32/gdkgc-win32.c : initialize pen_dashes and remove
the guesses from dashes to windoze line styles.
(predraw_set_forground) : always ExtCreatePen (PS_SOLID) on
Win9x, which does not support PS_USERSTYLE.
* gdk/win32/gdkdrawable-win32.c : new functions render_line_
<horizontal|vertical>. Use them if not running on NT in
gdk_win32_draw_<rectangle|segments|lines> ()
* gdk/win32/gdkkeys-win32.c (gdk_keyval_name) : return NULL
for keyval == 0 to avoid to have zeros in all menu entries
without accelerator.
2002-03-22 Havoc Pennington <hp@pobox.com>
* gdk/x11/gdkfont-x11.c (gdk_fontset_load): add more explanatory
text to the error message about missing charsets, and use
g_printerr() not g_warning() since this is typically not a
programming error (we do not export any API to ask whether
a font set will have missing charsets so apps realistically can't
do anything other than try the gdk_fontset_load())
* gtk/gtktextview.c (gtk_text_view_key_press_event): return FALSE
if the text view isn't editable and the user presses Return,
so default buttons and such can be activated, #74937
* gtk/gtktextbuffer.c (paste_from_buffer): don't insert
if the insertion point is not editable and the paste
is interactive, #74125
* gtk/gtkwindow.c (gtk_window_move_resize): enhance the #if 0
debug spew
* gtk/gtktextbuffer.c (cut_or_copy): only remove the previous
cut/copied data right before replacing it, when we know we are
going to replace it. Fixes#74049
Fri Mar 22 10:56:19 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkaccelmap.c (gtk_accel_map_save_fd): Fix memory
leak. (#74400, Morten Welinder.)
* gtk/gtkaccelmap.c: Properly handle short returns from
write() calls. (Handling EINTR isn't enough... that only
handles the case where you were interrupted before you
wrote a single byte.)
* gdk/linux-fb/gdkmouse-fb.c gdk/linux-fb/gdkkeyboard-fb.c:
Robustify against short returns from write() calls.
2002-03-22 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkgc-win32.c (gdk_win32_gc_set_dashes): Don't clear
all of the pen_style, just the PS_STYLE_MASK. Seems to fix#74441,
but that is just the special case of one-pixel wide one-pixel
on-off dotted lines. To fully implement generic X11 style dashed
lines a major rewrite would be needed.
Thu Mar 21 10:04:05 2002 Owen Taylor <otaylor@redhat.com>
* gdk/gdkwindow.c (gdk_window_real_get_depth): Remove
some left over debugging code with a G_BREAKPOINT()
in it.
2002-03-20 Alex Larsson <alexl@redhat.com>
Patches from Carlo E. Prelz <fluido@fluido.as>
Now gtkfb at least compiles and mostly works. The keyboard
stuff needs more loving.
* gdk/linux-fb/gdkkeyboard-fb.c (xlate_get_for_keyval,
xlate_get_for_keycode, raw_get_for_keyval, raw_get_for_keycode):
At least set the out parameters for these calls.
* gdk/linux-fb/gdkmain-fb.c:
(gdk_pointer_grab_info_libgtk_only, gdk_keyboard_grab_info_libgtk_only):
Implement new functions.
Wed Mar 20 16:36:08 2002 Owen Taylor <otaylor@redhat.com>
* gtk/*.c: Patch from Erwann Chenede, #73900 fixing
a lot of warnings with Forte CC, mostly implicit
casts between void * and function pointers.
* gdk/gdkevents.c (gdk_event_get_state): GdkEventVisibility's
state field is not a GdkModifierType. (Also #73900)
Wed Mar 20 10:55:56 2002 Owen Taylor <otaylor@redhat.com>
* gdk/x11/gdkevents-x11.c (gdk_event_translate): Fix assignment
of time in scroll events. (Sven Neumann, #75574)
2002-03-20 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkdnd-win32.c (gdk_drag_find_window): Use
gdk_window_get_toplevel(), as it is the top-level windows that are
registered for DND.
* gdk/win32/gdkwin32.h: Don't include gdkinternals.h.
Sat Mar 16 23:54:56 2002 Kristian Rietveld <kris@gtk.org>
* gdk/x11/gdkwindow-x11.c (gdk_window_set_static_bit_gravity): just
return when we have an input_only window (fix by Owen Taylor)
* gtk/gtktreeprivate.h (struct _GtkTreeViewPrivate): add
scroll_sync_timer
* gtk/gtktreeview.c (install_scroll_sync_handler): new function,
(scroll_sync_handler): ditto,
(gtk_tree_view_unrealize): take scroll_sync_timer into account
(gtk_tree_view_row_deleted): install scroll_sync_timer instead of
calling top_row_to_dy/dy_to_top_row directly
-- this greatly speeds up clearing the model (#73199)
* gtk/gtktreemodelsort.c
(gtk_tree_model_sort_convert_path_to_child_path): fix up this function,
for some reason I really screwed it up (fixes#74663)
Fri Mar 15 15:37:01 2002 Owen Taylor <otaylor@redhat.com>
* gdk/x11/gdkdrawable-x11.c (convert_to_format):
Fix byte shift arithmetic for big-endian. (Tracked
down with help from Tuomas Kuosmanen)
* gdk/x11/gdkkeys-x11.c (update_keymaps): Fudge keymap to have lower
upper case variants if there is only a single keysym per keycode and
the key symbol has upper and lower case variants (#74512)
Thu Mar 14 17:09:52 2002 Owen Taylor <otaylor@redhat.com>
* gdk/gdkwindow.c (gdk_window_process_all_updates): Fix rather
improbable reentrancy problem if a window is destroyed
while updates are being processed on another window. (Noticed
by Michael Meeks, #74708)
2002-03-06 Manish Singh <yosh@gimp.org>
* gdk/Makefile.am
* gtk/Makefile.am: rework so that only things that really need to be
done post install are in the hook
2002-03-05 Manish Singh <yosh@gimp.org>
* gdk/Makefile.am
* gtk/Makefile.am: use install-foo-hook instead of install-foo-local,
since foo-local is run before the their respective builtin targets
in automake 1.5
2002-03-06 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkevents-win32.c (gdk_pointer_grab,
gdk_pointer_ungrab): Bypass calls to the grab/ungrab functions in
gdkinput-win32.c, as they don't effectively do anything anyway.
(build_key_event_state): Set MOD2 bit if key is in AltGr group.
(gdk_event_translate): Call _gdk_windowing_window_get_offsets()
each time the offsets are used, on the window they refer to,
instead of once in the beginning. The window in question might
change due to event propagation.
(gdk_event_translate): Set key event group to 0 for non-AltGr
keys. Do set the SHIFT bit in the key event state also for
ISO_Left_Tab. Now backtabbing finally works again.
* gdk/win32/gdkunput-win32.c: Minor debugging output change.
* gdk/win32/gdkkeys-win32.c (update_keymap): Set VK_TAB mapping to
Gdk_Tab and GDK_ISO_Left_Tab, like on X11.
(gdk_keymap_translate_keyboard_state): Add similar code as in the
non-XKB case on X11 to generate a more correct
consumed_modifiers. Add debugging output.
(gdk_keyval_name): Use the U+xxxx format for UCS characters
encoded as keyvals. Never return NULL, but hex number
representation if keyval not in table.
* gdk/win32/gdkwindow-win32.c (gdk_window_focus): Call
SetFocus(). Doesn't seem to have any harmful effect, and probably
is close to what this function is supposed to do. But it didn't
fix GtkCombo as I had hoped.
(gdk_window_set_type_hint): Don't intern the _NET_WM_* atoms that
weren't used and wouldn't have any meaning on Win32 anyway.
Mon Mar 4 12:43:48 2002 Owen Taylor <otaylor@redhat.com>
Fixes for non XKB operation. (#73103)
* gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state):
Fix problem where GDK_Tab was converted to GDK_ISO_Left_Tab
always, not just if Shift was pressed.
* gdk/x11/gdkkeys-x11.c gdk/x11/gdkprivate-x11.c
gdk/x11/gdkevents-x11.c: Move determination of group
to _gdk_x11_get_group_for_state(), fix to handle non-XKB.
* gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state): Only
add in SHIFT_MASK | LOCK_MASK to consumed_modifiers if the shifted
and unshifted keysyms are different. Only add in group_switch_mask
if the group-switched and non-group-switched keysyms are diferent.
* gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state):
When calling XKeycodeToKeysym, use group * keysyms_per_keycode / 2
not group * keysyms_per_keycode. (Tor Lillquist)
* gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state):
Remove the hack to fudge the return value for GDK_Tab + Shift
since that didn't work -- it didn't affect the reverse mapping.
Instead fudge the non-XKB keymap instead.
Fri Mar 1 00:46:49 2002 Owen Taylor <otaylor@redhat.com>
* configure.in: Add a --enable-explicit-deps option to control
whether dependency libraries are written into .la files and .pc
files. Default is auto: if static libraries are off and shlib
dependencies are found, don't write deps.
* gdk/Makefile.am gtk/Makefile.am sanitize-la.sh: Shell script
used to strip dependencies out of .la files.
* Makefile.am: Add a slightly modified distcheck rule that passes
--enable-gtk-doc to the configure inside.
(So that 'make dist' succeeds inside.)
* configure.in: If pango was compiled with
--disable-explicit-deps, then repeat the checks for X and
freetype ourselves so we don't depend on linking to libraries
that pango doesn't reveal in the link line. Add some more paranoia
for whether what we detect ourselves matches what Pango backends
we found.
2002-03-03 Tor Lillqvist <tml@iki.fi>
* gdk/gdk.def
* gdk/win32/gdkevents-win32.c: Add and (trivially) implement
gdk_pointer_grab_info_libgtk_only() and
gdk_keyboard_grab_info_libgtk_only().
2002-03-02 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkevents-win32.c (build_keypress_event): Reorder
code, keyval should get a value in all cases now.
(build_keyrelease_event): Set GDK_CONTROL_MASK here, too, if it
was a control character produced with Alt+keypad digits.
(gdk_event_translate): Minor cleanups to keyboard input
handling. Still need more major work.
Fri Mar 1 18:39:44 2002 Owen Taylor <otaylor@redhat.com>
* gdk/x11/{gdkevents-x11.c,gdkmain-x11.c,gdkprivate-x11.h,
gdkwindow-x11.c}: Robustify tracking of pointer grab window.
* gdk/x11/gdkmain-x11.c: Keep track of current keyboard
grab window.
* gdk/x11/gdkmain-x11.c (gdk_pointer_grab_info_libgtk_only,
gdk_keyboard_grab_info_libgtk_only): Private libgtk => libgtk
API for finding out current grab information.
* gtk/gtkmain.c (rewrite_event_for_grabs): Rewrite events
so that the effective behavior of owner_events = TRUE is changed
to "deliver events to same window group normally" instead
of "deliver events to same application normally. #69934
* gtk/gtkrange.c: Use an explicit gtk_grab_add() so that
it works within the GtkList combo, where there is a
owner_events = FALSE gdk_pointer_grab() already in effect.
(#65006, reported by Damon Chaplin)
2002-03-01 Tor Lillqvist <tml@iki.fi>
* gdk/gdkkeys.c (gdk_keyval_convert_case)
* gdk/x11/gdkkeys-x11.c (gdk_keyval_convert_case): If the keyval
to be converted is one of those that consist of a Unicode
character plus our special flag bit, use the
g_unichar_to{lower,upper} functions (Bug#73112).
* gdk/gdkkeys.c (gdk_keyval_convert_case): Drop the #ifdefs for
the GDK keysym symbols. As we include gdkkeysyms.h, we know they
are there. And if GDK keysyms some day are added/removed, this
function needs changes, too, anyway.
Fri Mar 1 17:00:28 2002 Owen Taylor <otaylor@redhat.com>
* gdk/gdkpolyreg-generic.c: Fix some reported (but not significant)
UMR's from purify. (#73160, Morten Welinder)
Thu Feb 28 19:55:01 2002 Owen Taylor <otaylor@redhat.com>
* gdk/gdkkeys.c: include gdkkeysysms.h if
compiling gdk_keyval_convert_case.
* gtk/gtkmenubar.c (gtk_menu_bar_cycle_focus): Implement
<Control>Tab <Control><Shift>Tab to cycle between
all menu bars in a toplevel once one is up.
* tests/testgtk.c: Add a second menubar, this
example is already full of crack anyways.
* gtk/gtkmenushell.c (gtk_menu_shell_key_press): Padd
unhandled events up to the parent menu shell.
* gtk/gtkmenuitem.c (gtk_menu_item_select_timeout): Only
pop up the menu if the parent menu shell is still active.
* gtk/gtkcontainer.[ch] (_gtk_container_focus_sort): Add a
old_focus argument, export privately.
2002-03-01 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkwindow-win32.h (struct _GdkWindowImplWin32): Remove
the event_mask, it is now in GdkWindowObject.
* gdk/win32/gdkwindow-win32.c: Change accordingly. Set the
GDK_STRUCTURE_MASK in gdk_window_set_events(), as it is always set
in gdk_window_new(), too. (Bug#72921)
* gdk/win32/gdkevents-win32.c: Change accordingly here, too.
(vk_from_char): New function, calculates the virtual keycode
corresponding to the char in a WM_CHAR message.
(build_keypress_event, build_keyrelease_event): Use it.
(build_keypress_event): Call ImmReleaseContext() after using the
input context. This might plug a memory or resource leak.
(build_key_event_state): Remove #if 0 code.
(gdk_event_translate): Actually, it would be preferrable to always
handle just the WM_KEYDOWN and WM_KEYUP messages, not WM_CHAR at
all, and thus drop the contorted logic with ignore_wm_char etc.
* gdk/win32/gdkkeys-win32.c: (gdk_keymap_get_entries_for_keyval):
Debugging output.
(gdk_keymap_translate_keyboard_state): Return correct value. (But
_gtk_key_hash_lookup() doesn't check the return value...)
Thu Feb 28 14:49:08 2002 Owen Taylor <otaylor@redhat.com>
* gdk/x11/gdkwindow-x11.c (gdk_window_foreign_new): Initialize
private->event_mask from attributes we retrieved for
the window.
2002-02-28 Tor Lillqvist <tml@iki.fi>
* gdk/win32/Makefile.am: Move bdfcursor.c from
libgdk_win32_la_SOURCES to EXTRA_DIST.
* gdk/win32/gdkgc-win32.c (gdk_win32_colormap_color_pack): In the
bitmap case, use just the LSB of the pixel value. Does this match
X11 semantics? Or should it be pixel!=0?
Wed Feb 27 12:34:21 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkiconfactory.c: Register all stock icons
"unscaled" so that we don't have icons only available
at some sizes. (#68896)
Wed Feb 27 11:59:05 2002 Owen Taylor <otaylor@redhat.com>
Some ASCII vs. locale fixes from Darin Adler (#55837)
* gtk/gtkcombo.c (gtk_combo_find): Implement real
case-insensitive comparison using g_utf8_casefold().
* gtk/gtkaccellabel.c (gtk_accel_label_refetch): Fix
assignment of gunichar to char by replacing g_unichar_totitle
with g_ascii_toupper ... the contents would only be ascii
here anyways.
* gdk/x11/gdkim-x11.c (_gdk_x11_initialize_locale)
* gtk/gtkfontsel.c (gtk_font_selection_show_available_fonts):
Replace a calls to g_strcasecmp() with g_ascii_strcasecmp().
Tue Feb 26 21:35:31 2002 Owen Taylor <otaylor@redhat.com>
* gdk/gdkkeyuni.c: Shouldn't map KP_Enter to '\n',
or KP_Tab to \t' since we don't map the corresponding
main keyboard characters in that way. (#61561,
Padraig O'Briain)
Tue Feb 26 19:16:28 2002 Owen Taylor <otaylor@redhat.com>
* gdk/win32/gdkkeys-win32.c (gdk_keyval_name):
* gdk/x11/gdkkeys-x11.c (gdk_keyval_name):
Make sure Page_Up, Page_Down get those names, not Prior,
Next.
2002-02-26 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkkeys-win32.c: Implement the functions that until
now just were non-functional stubs. For "hardware keycodes", we
use Windows virtual keycodes. Not scancodes, although that at
first might seem more low-level and a better match to X11
keycodes.
The Windows API is really mixed up and confused with respect to
scancodes and virtual keycodes. (Surprised?) Some scancodes are
generated by two keys on the keyboard (!), and although the
keyboard messages do have a flag to indicate which key the user
pressed, other API that take a scan code as input don't let you
specify which actual key you mean.
(update_keymap): Function to build a X11-like representation of
the keyboard. Each key has four keysyms: two levels (nonshifted
and shifted) and two groups (normal and with AltGr).
(gdk_keymap_get_direction): Use the codepage corresponding to the
thread's input locale, not the system codepage.
* gdk/win32/gdkglobals-win32.c
* gdk/win32/gdkmain-win32.c
* gdk/win32/gdkprivate-win32.h
* gdk/win32/gdkwindow-win32.h
* gdk/win32/gdkwindow-win32.c: Remove the input_locale and
charset_info fields from GdkWindowImplWin32. Input locale is
per-thread in Windows, and as GDK on Windows really only works
when the GDI interaction all happens in just one thread anyway,
this state can be global. Use globals _gdk_input_locale and
_gdk_input_codepage instead. Set these based on the thread's input
locale (keyboard layout, or which IME is active).
* gdk/win32/gdkevents-win32.c: Set the group and hardware_keycode
fields in GDK key events. On input locale change messages, set
the global state variables, and inform update_keymap() that it
has to rebuild the keymap.
Sun Feb 24 23:23:28 2002 Owen Taylor <otaylor@redhat.com>
* tests/testselection.c (main): Convert over to using
an GtkInvisible - selections don't work properly
with now NO_WINDOW GtkToggleButton. (Example really
should be using GtkClipboard.)
* gdk/x11/gdkselection-x11.c (gdk_selection_property_get):
Fix problems if sizeof(GdkAtom) != sizeof(Atom). (Unlikely)
* gdk/x11/gdkproperty-x11.c (gdk_property_get): Set
actual_length when converting atoms as well.
(Gregory Merchan, #72074)
2002-02-24 Alexander Larsson <alla@lysator.liu.se>
* gdk/x11/gdkwindow-x11.c (gdk_window_set_background):
Always NULL out bg_pixmap, even when it used to be None
or parent relative.
Sat Feb 23 20:59:05 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkstyle.c gtk/gtkrc.[ch]: gtk/gtksettings.c gtk/gtkstyle.c:
Patch from Richard Hestilow to add a gtk-font-name GtkSetting.
* gdk/x11/gdkevents-x11.c: Add Gtk/FontName XSETTING.
2002-02-24 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkevents-win32.c (build_keypress_event,
build_keyrelease_event, gdk_event_translate): Set also group in
the key press/release events, always to zero for now.
Sat Feb 23 11:54:12 2002 Owen Taylor <otaylor@redhat.com>
* gtk/*.h gdk/*.h: Massive padding addition to class structures.
* gtk/gtktextmark.h: Fix a FIXME about G_CONST_RETURN.
Fri Feb 22 18:36:45 2002 Owen Taylor <otaylor@redhat.com>
Fixes for AIX compilation from Miroslaw Dobrzanski-Neumann,
#72182.
* demos/gtk-demo/main.c gtk/gtktreeview.c gtk/gtktoolbar.c
gtk/gtkpaned.c gtk/gtkobject.c gtk/gtkcontainer.c
gtk/gtkcolorsel.c gtk/gtkcellrender.c gdk/x11/xsettings-client.h:
Squash trailing commas on enumerations.
* gdk/gdktypes.h: Drop GDK_RELEASE_MASK to 1 << 30 instead
of 1 << 31 to work around AIX compiler problem. (C standard
seems to imply that compiler is required to use an unsigned
type for the enum value in this case.)
* gdk/gdkpixbuf-drawable.c: Fix lvalue casts.
* gtk/gtkwindow.h: Fix use of enum types for bitfields ...
compilers may choose to use a signed type for bitfields.
Wed Feb 20 14:26:47 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkkeyhash.[ch]: Implement "fuzzy" key binding lookups;
allow matches on key and level but not group. Also, implement
ignoring "consumed modifiers correctly."
* gtk/gtkaccelgroup.c gtk/gtkbindings.c: Convert to using
GtkKeyHash.
* gtk/gtkdebug.h gtk/gtkmain.c: Support GTK_DEBUG=keybindings
* gdk/x11/gdkevents-x11.c (gdk_event_translate): Fill in
the group for key release events as well as key press events.
* gdk/gdkkeys.h gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state):
Rename unused_modifiers to consumed_modifiers, make the docs and
non-Xkb implementation match the Xkb implementation.
* gdk/linux-fb/gdkkeyboard-fb.c gdk/win32/gdkkeys-win32.c: Propagate
doc and parameter name changes.
* gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state):
XkbTranslateKeyCode doesn't handle LockMask, we need to handle
it ourselves.
* gdk/x11/gdkkeys-x11.c (gdk_keymap_translate_keyboard_state): Force
<Shift>Tab to give GDK_ISO_Left_Tab, since we need consistency
to allow dealing with ISO_Left_Tab.
* gtk/gtkwindow.c gtk/gtktextview.c gtk/gtkscrolledwindow.c
gtk/gtkpaned.c gtk/gtkcombo.c gtk/gtknotebook.c:
Remove inappropriate uses of GDK_ISO_Left_Tab. (GDK_ISO_Left_Tab
or <Shift>Tab both are equivalent as a binding specifier.)
* gtk/gtkbutton.c (gtk_button_class_init): Make ::activate
GTK_RUN_ACTION, so you can bind an accelerator to it.
* gtk/gtklabel.c (gtk_label_set_uline_text_internal): Call
gdk_unicode_to_keyval on the mnemonic character.
* tests/testgtk.c: Add a test for the new fuzzy key binding matching.
