_gtk_rbtree_reorder() was moving the node's data while reordering. As we
use the node pointer in the a11y code as a hash key, this didn't work.
So this rewrite changes that. As a bonus, it is less code and faster.
Woohoo!
- Make sure the rotated nodes aren't nil
- Use existing functions for complex computations
- Don't use NULL checks for variables guaranteed to not be NULL/nil
This has one major caveat: The new value is const, so read-only memory.
Any attempt to write to it will cause a crash. Note that we are not ever
supposed to write to it, but bugs happen...
The code used to set nil->parent, which could cause segfaults. Don't do
that. We also need to pass the parent explicitly to the fixup code,
because the node during fixup may be the nil node.
So instead of copying the children and height to the new node, we keep
the old node and copy all the old stuff to it.
This is necessary so the accessibility code can use the node as a key in
the hash table or store the node as a reference to the row instead of
GtkTreeRowReference. And because it already does that (oops), this fixes
a bunch of segfaults with a11y enabled.
gtk_rbtree_adjust() will adjust the summed values of a node and all its
parents in the tree. Currently only implemented by splitting out the
function from gtk_rbtree_free().
.. as a replacement for _gtk_rbtree_node_find_parity(). Instead of 1 or
0, the function now returns the index of node in the complete tree
(counting from the root). And this is of course identical to the row
number.
Instead of just storing the least significant bit of the row number,
store the full row number. This will soon be useful for accessibility.
But CSS could like it, too.
Preferrably should be made just into a local variable for libgtk like
_gdk_debug_flags for libgdk. But for now used by
gtk/tests/textbuffer.c and modules/printbackends/cups/gtkprintbackendcups.c.
2006-01-18 Matthias Clasen <mclasen@redhat.com>
* gtk/gtkrbtree.c:
* gtk/gtktreemodel.c: Fix several g_new() calls which
were using extra indirections or pointless casts.
(#327423, Morten Welinder)
Tue Nov 1 16:18:24 2005 Tim Janik <timj@imendio.com>
* gtk/gtkrbtree.[hc]: get rid of GAllocator usage, allocate and free
nodes via the g_slice_*() API.
_gtk_rbtree_new(): fixed wrong node allocation via g_new().
2005-07-08 Kristian Rietveld <kris@gtk.org>
Fixes#165034, reported by Jorn Baayen.
* gtk/gtkrbtree.[ch] (_gtk_rbtree_set_fixed_height): add a
mark_valid parameter (so we can mark all nodes as valid when we
are setting them to the fixed height).
* gtk/gtktreeview.c (initialized_fixed_height_mode): mark all
nodes as valid when setting the fixed height,
(do_validate_rows): update call to _set_fixed_height,
(gtk_tree_view_row_inserted): mark node as valid when fixed
height mode is enabled and a height has been set,
(gtk_tree_view_build_tree): if fixed height mode is enabled,
mark new nodes as valid in addition to setting the height.
2005-03-20 Matthias Clasen <mclasen@redhat.com>
Make PLT-reduction work with gcc4, and don't include
everything in gdkalias.h:
* gtk/grk.symbols: Group symbols by header and source file.
* gtk/makegtkalias.pl: Protect definitions by the same
preprocessor symbols used to guard the headers. Move
the alias declarations to a separate file which is
produced when calling makegtkalias.pl -def
* gdk/Makefile.am (gtkaliasdef.c): Add a rule to generate
this file.
* gtk/*.c: Include gtkalias.h after the other headers,
include gtkaliasdef.c at the bottom.
* gtk/*.h: Small cleanups.
Mon Aug 9 12:48:04 2004 Matthias Clasen <maclas@gmx.de>
Add hidden aliases for exported symbols which are
used internally in order to get rid of many PLT
entries. (#145519, Arjan van de Ven)
* gtk/Makefile.am: Add rules to generate gtk.def and
from gtk.symbols, and make make check check the abi
with abicheck.sh.
(gtk_private_h_sources): Add gtkinternals.h
(gtk_built_private_headers): Add gtkalias.h
(gtk_extra_sources): Add gtk.symbols
(EXTRA_DIST): Add makegtkalias.pl and abicheck.sh
* gtk/gtk.symbols: New file. Definition of the GTK+ ABI.
The file can be processed by cpp to filter out certain
subsets of symbols.
* gtk/abicheck.sh: New file. Script to check the actually
symbols exported from libgtk-x11.2.0.so against the symbols
found in gtk.symbols.
* gtk/makegtkalias.pl: New file. Perl script to generate the
header containing the alias definitions for internally used
exported symbols from a list of symbols.
* gtk/gtkinternals.h: New file. An uninstalled header listing
symbols which must be exported for some reason and do not appear
in any other header.
* gtk/*.c: Include gtkalias.h
2004-03-05 Federico Mena Quintero <federico@ximian.com>
Fixes#136082 and #135265, patch by Morten Welinder.
* configure.in: Use AC_SYS_LARGEFILE.
* */*.c: #include <config.h>
Mon Mar 31 13:52:13 2003 Jonathan Blandford <jrb@redhat.com>
* gtk/gtkrbtree.c (_gtk_rbtree_reorder): null initialize reorder's
members so Purify won't complain.
Wed Nov 20 19:29:50 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtkrbtree.c (gtk_rbtree_reorder_sort_func): return a value
in the range of [-1, 1] instead [0, 1],
(gtk_rbtree_reorder_invert_func): ditto,
This makes treeview reordering working on platforms as FreeBSD and
solaris, thanks go to Heiner Eichmann for finding this out and
sending in a patch (#98251).
Wed Nov 20 19:27:07 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtktreestore.c (node_free): only free the data list if
node->data is not NULL,
(gtk_tree_store_remove): free the data list from the node if
needed. (#94728, pointed out by Peter Bloomfield).
Wed Nov 20 19:23:13 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtktreeview.c (gtk_tree_view_remove_column): stop editing
and set the edited_column to NULL if edited_column is set.
(#91288, modified patch from Josh Parsons).
Wed Nov 20 19:20:34 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtktreeview.c (gtk_tree_view_size_allocate): do a full
redraw if vadjustment->value is bigger than tree_view->priv->height,
(gtk_tree_view_row_deleted): free the row ref if it isn't valid
anymore. (#83726, reported by Robert Kinsella).
Thu Jun 6 17:27:01 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtkrbtree.c (_gtk_rbtree_reorder): don't leak the GArray
(figured out by Matthias Clasen, fixes#82026).
Thu May 16 14:51:14 2002 Owen Taylor <otaylor@redhat.com>
Fixes for #78258
* gtk/gtkrbtree.c demos/gtk-demo/editable-cells.c:
Fix some 64-bit warning problems. (Patch from Matthias
Clasen.)
* gtk/gtkrbtree.c (_gtk_rbtree_debug_spew_helper): Surround
a bunch of debug code with G_ENABLE_DEBUG.
* configure.in demos/gtk-demo/main.c: Replace
#ifndef G_OS_WIN32 with proper #ifdef HAVE_FLOCKFILE.
Wed Mar 20 22:59:23 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtkrbtree.[ch]: add _gtk_rbtree_set_fixed_height()
* gtk/gtktreeprivate.h: add fixed_height_check field
* gtk/gtktreeview.c (gtk_tree_view_init): initialize
scroll_sync_timer and fixed_height_check
(do_validate_rows): add fixed_height_check. If all validated rows
in the first cycle have the same height, then we set that height for
the entire tree. This is some sort of 'fake' optimization, but helps
a lot for the common case. We keep validating the entire tree in
the background though.
(gtk_tree_view_set_model): reset fixed_height_check
Sun Jan 27 23:23:23 Kristian Rietveld <kris@gtk.org>
* gtk/gtkrbtree.[ch]: add _gtk_rbtree_mark_invalid (my previous
commit already needed this ...)
(/me broke the build for the first time ... :)
Sat Dec 8 10:04:54 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtkrbtree.c (_gtk_rbtree_debug_spew): Add debug spew to lots
of places, so that people with crashy trees can just run their
program with GTK_DEBUG=TREE.
Fri Dec 7 20:06:14 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtktreeview.c (validate_visible_area): get logic right,
#66249
* gtk/gtkrbtree.c (_gtk_rbtree_remove_node): fix bug where
removing a node ended up with a corrupt tree. Really really nasty
bug.
(_gtk_rbtree_debug_spew): new debug helper function
(_fixup_validation): new inline function to clean up code
readability a lot.
Fri Nov 2 16:45:17 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtkrbtree.c (_gtk_rbtree_remove_node): Clean up height code
a bit. I don't think it's completely correct yet, but it's
getting there.
