After doing some performance analysis, it was found that the GTK+ mediaLib code
triggers unnecessary lazy loading of dependent libraries. The current code
uses RTLD_DEFAULT, RTLD_PROBE, RTLD_NEXT, and RTLD_SELF. However, RTLD_PROBE
is all that is necessary, and avoids triggering the lazy loading. So this
commit fixes the code to just use RTLD_PROBE. (Bug 580678)
2007-05-16 Brian Cameron <brian.cameron@sun.com>
* acconfig.h, configure.in, gdk/Makefile.am, gdkprivate.h, gdkrgb.c,
gdkdraw.c, gdk-pixbuf/gdk-pixbuf-scale.c,
gdk-pixbuf/pixops/Makefile.am, gdk-pixbuf/pixops/pixops.[ch],
gdk-pixbuf/pixops/timescale.c: Add Sun mediaLib support so that
hardware acceleration via mediaLib is enabled if mediaLib is
detected via configure. Enhancement request #344813. I was given
permission to commit in the bug report by Matthias Clasen.
* gdk/medialib.[ch]: New files added for mediaLib support.
* docs/reference/gdk-pixbuf/tmpl/scaling.sgml,
docs/reference/gdk/tmpl/rgb.sgml, gdk/gdkdraw.c: Add docs for
mediaLib support.
svn path=/trunk/; revision=17855
2005-01-02 Matthias Clasen <mclasen@redhat.com>
* pixops/pixops.c (bilinear_box_make_weights): Really fix the
math to make the vectors sum up to 1, making GDK_INTERP_HYPER
work again. (#162703, Dennis Nezic)
Sun Aug 22 03:20:56 2004 Matthias Clasen <maclas@gmx.de>
* pixops/pixops.c (pixops_scale_nearest): Fix a mixup in the handling
of dest_channels. (#111922, Christophe Fergeau)
2004-08-11 Matthias Clasen <mclasen@redhat.com>
* pixops/pixops-internal.h:
* pixops/pixops.c:
* pixops/*.S: underscore prefix the mmx implementations, in order
to not export them.
* gdk-pixbuf.symbols: Remove the mmx pixops implementations.
* abicheck.sh: No need to define INCLUDE_MMX_SYMBOLS any more.
Sat Jul 3 00:41:44 2004 Matthias Clasen <maclas@gmx.de>
* pixops/pixops.c (bilinear_box_make_weights): Correct the
math to calculate bilinear weights. (#112412, Brian Cameron)
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>
2003-07-18 Matthias Clasen <maclas@gmx.de>
* pixops/pixops.c (correct_total): Split correction into multiple pieces if no single weight is large
enough to apply the unsplit correction. (#117431, problem reported by Tomas Ögren)
Tue Apr 1 15:33:51 2003 Owen Taylor <otaylor@redhat.com>
* pixops/pixops.c (make_weights): Exploit the fact that all
our filters are separable to simplify the calculation of
the weight tables. (Based on a patch from Brian Cameron.)
* pixops/Makefile.am pixbuf-transform-math.ltx:
Add a short article describing how the math in pixops.c
works.
Thu Jun 13 18:11:11 2002 Owen Taylor <otaylor@redhat.com>
* pixops/pixops.c: Fix problem where when no alpha was
involved, we scaled rather than composited, but then
went ahead and composited anyways. (#76958, Patch
from Matthias Clasen). Needs to be merged to GNOME
1.4 gdk-pixbuf
* pixops/pixops.c (correct_total): New function to make sure
that the weights sum up to the correct total without introducing
negative weights.
(tile_make_weights, bilinear_make_fast_weights,
bilinear_make_weights): Use correct_total.
Fri Feb 8 23:11:15 2002 Owen Taylor <otaylor@redhat.com>
* pixops/pixops.c: Force all weight arrays to sum exactly
to 65535. (Fixes#70971, reported by Federico Mena Quintero)
* Makefile.am (libgdk_pixbuf_1_3_la_DEPENDENCIES): Add
libpixops.la.
Fri Jan 11 18:05:07 2002 Owen Taylor <otaylor@redhat.com>
* pixops/pixops.c: Fix integer overflow for the values
passed to the pixel functions. (#55148)
Fri Jan 11 17:53:58 2002 Owen Taylor <otaylor@redhat.com>
* pixops/pixops.c (pixops_scale_nearest): Remove meaningless
FIXME.
* pixops/pixops.c (pixops_process): Comment the computation
of run_end_index and fix some edge cases.
* pixops/{scale_line_22_33_mmx.S, composite_line_22_4a4_mmx.S:
composite_line_color_22_4a4_mmx.S}: Handle the case where
dest <= dest_end properly. (#52001)
Thu Jan 3 19:31:58 2002 Owen Taylor <otaylor@redhat.com>
* pixops/pixops.c (scale_line_22_33): Fix and uncomment
non-MMX version.
* pixops/pixops.c (pixops_composite_nearest): Remove a
division.
* pixops/pixops.c (pixops_composite): Add some docs
about the parameters.
* pixops/README: Add notes about the correct algorithms
for alpha compositing and how to implement them quickly.
2001-08-19 Sven Neumann <sven@gimp.org>
* pixops/pixops.c: applied some simple but effective optimizations
to pixops_composite_nearest() and pixops_composite_color_nearest().
* gdk-pixbuf-csource.c: Add missing <stdlib.h> include.
* io-png.c: (setup_png_transformations): Add ifdef so we compile
without warnings with G_DISABLE_CHECKS on.
* io-pnm.c: (gdk_pixbuf__pnm_image_load_increment): Add a missing
const.
* io-wbmp.c: (getin), (get_mbi): Add a missing const.
* io-xbm.c: (gdk_pixbuf__xbm_image_load_real): Get rid of some
unused locals and add an initial value to quiet the compiler's
unintialized variable warning.
* pixops/pixops.c: Put an ifdef around some dead code.
2001-05-04 Havoc Pennington <hp@redhat.com>
* configure.in: fix some shell typos
* gtk/gtkcolorsel.c (gtk_color_selection_destroy): warning fix
* gtk/gtkimage.c: handle animations
* gtk/gtkcheckbutton.c (gtk_check_button_size_request): request
border_width * 2, not just border_width
* gtk/gtkscale.c: add "format_value" signal to allow people
to override the way values are drawn.
(gtk_scale_get_value_size): fix width/height mistake,
and compute size from actual displayed text, not
from made-up text.
* gtk/gtktexttag.c (gtk_text_tag_class_init): fix return type in
signal registration
* tests/testtext.c: Add "Remove all tags" menu item for testing
* gtk/gtktextbuffer.c (gtk_text_buffer_remove_all_tags): implement
* demos/gtk-demo/main.c (main): add hack so we can find modules
without installing gtk
* demos/gtk-demo/textview.c (insert_text): demo font scaling
* gtk/gtkcellrenderertext.c: Add "scale" property (font scaling
factor)
(gtk_cell_renderer_text_set_property): remove some bogus
g_object_notify
* gtk/gtktexttag.c: add "scale" property which is a font scaling
factor
* gtk/gtktextlayout.c (add_text_attrs): add font scale attribute
to layout
* gtk/gtktextiter.c (gtk_text_iter_is_start): rename from
gtk_text_iter_is_first
2001-05-04 Havoc Pennington <hp@redhat.com>
* pixops/pixops.c (pixops_process): merge fix from stable: Patch
from hoshem@mel.comcen.com.au to fix nonzero X offsets. Fixes
bug #50371.
* gdk-pixbuf/pixops/pixops.c (pixops_composite_nearest): merge
from stable: Patch from OKADA Mitsuru <m-okada@fjb.co.jp> to fix
confusion of using "src" instead of "p".
(pixops_composite_color_nearest): Use a more accurate (and
correct, to begin with) compositing method. This cures checks
showing through on images with no alpha.
* gdk-pixbuf.c (gdk_pixbuf_fill): fix bug that left some trailing
bytes unfilled.
* gdk-pixbuf-io.h: fix UpdatedNotifyFunc to use signed ints
* gdk-pixbuf-loader.h (struct _GdkPixbufLoaderClass): Change
area_updated signal to use signed ints. Removed animation-related
signals.
* io-gif.c, io-gif-animation.h, io-gif-animation.c: Massive
rewrite action
* gdk-pixbuf-animation.c: Add GdkPixbufAnimationIter to abstract
all the pesky details. Remove old frame-based API. Make
GdkPixbufAnimation an abstract base class, derived by the loaders.
2000-10-06 Havoc Pennington <hp@redhat.com>
* Makefile.am (SRC_SUBDIRS): contrib subdir
* gdk/gdkpixbuf.h: Move GdkPixbufAlphaMode to gdk-pixbuf library,
so it can be used in Xlib version
* demos/testpixbuf.c (update_timeout): error checking from 1.0
tree
* gtk/gdk-pixbuf-loader.c (gdk_pixbuf_loader_get_pixbuf): Sync
change from 1.0 tree that returns first frame of animation
if the loaded file is an animation.
* contrib: add new directory to hold stuff that comes with GTK+
but isn't really part of GTK+ (for now, gdk-pixbuf-xlib)
* configure.in: add contrib/*
2000-10-06 Havoc Pennington <hp@redhat.com>
* gdk-pixbuf.h: add GdkPixbufAlphaMode
2000-10-06 Havoc Pennington <hp@redhat.com>
This entry is a summary of the merged-in changes from 1.0.
Relevant original ChangeLog entries are spliced in after
this entry; the files they refer to are from the 1.0
gdk-pixbuf sources.
* pixops/pixops.c (pixops_composite_nearest): sync a small fix
from 1.0
* io-xpm.c (xpm_seek_string): add fscanf error check from 1.0
Add progressive loader from 1.0
* io-tiff.c (gdk_pixbuf__tiff_image_begin_load): mem leak fixes
from 1.0 tree
* io-pnm.c: new version from 1.0 tree
* io-jpeg.c (gdk_pixbuf__jpeg_image_load): sync from 1.0, use
malloc not g_malloc
* io-gif.c (lzw_read_byte): sync from 1.0, change a g_error to
g_warning
(gif_get_next_step): return 0 here, sync from 1.0
* gdk-pixbuf-util.c: sync email address change for Cody
Russell
2000-09-11 Jeffrey Stedfast <fejj@helixcode.com>
* gdk-pixbuf/io-pnm.c: Pretty much totally rewrote again because
last nights code was still "broken". Should now properly handle
all error conditions gracefully.
2000-09-10 Jeffrey Stedfast <fejj@helixcode.com>
* gdk-pixbuf/io-pnm.c: Rewrote.
2000-09-09 Federico Mena Quintero <federico@helixcode.com>
* gdk-pixbuf/pixops/pixops.c (pixops_composite_nearest): Compute
the correct dest offset.
2000-08-25 Federico Mena Quintero <federico@helixcode.com>
* gdk-pixbuf/io-xpm.c: #include <unistd.h>
2000-08-05 Larry Ewing <lewing@helixcode.com>
* gdk-pixbuf/io-tiff.c: stop leaking context->tempname.
* gdk-pixbuf/io-xpm.c: same as above.
2000-07-26 Michael Meeks <michael@helixcode.com>
* gdk-pixbuf/io-jpeg.c (gdk_pixbuf__jpeg_image_load): make
g_malloc a malloc.
2000-07-21 Larry Ewing <lewing@helixcode.com>
* gdk-pixbuf/io-xpm.c: add a fake progressive loader so that
xpm at least supports the progressive interface like the one in
io-tiff.c. This should be reimplemented as an actual progressive
loader.
2000-07-19 Jonathan Blandford <jrb@redhat.com>
* demo/pixbuf-demo.c (update_timeout): changed scaling level to
make it look better.
* gdk-pixbuf/testpixbuf.c (update_timeout): Patch from michael
meeks to handle errors better.
Wed Jun 21 16:38:13 2000 Owen Taylor <otaylor@redhat.com>
* gdk-pixbuf/* docs/reference/gdk-pixbuf/*: Welcome aboard,
gdk-pixbuf.
* gtk/gdk-pixbuf-loader.c (gdk_pixbuf_loader_class_init):
fixups for GObject.
* gdk/Makefile.am gdk/gdkpixbuf-render.[ch] gdk/gdkpixbuf.[ch]: Bits of
gdk-pixbuf with GDK dependencies moved into GDK.
* gtk/Makefile.am gtk/gdk-pixbuf-loader.[ch]: Temporarily
move gdk-pixbuf-loader here until GObject has signals.
* demos/: New directory of demos. Move demos from
gdk-pixbuf here.
* demos/pixbuf-init.c: Small bit of code to check for
loaders in ../gdk-pixbuf/.libs/gdk-pixbuf, and if found,
set GDK_PIXBUF_MODULEDIR appropriately.
* gdk/gdkcompat.h: Remove GDK_DRAWABLE_PIXMAP compat
define which no longer makes sense.
* gdk-pixbuf/pixops/pixops.c: (pixops_composite_nearest),
(composite_pixel), (composite_line):
* gdk-pixbuf/test-gdk-pixbuf.c: (simple_composite_test_one_type):
Now that Owen explained the algorithm, I think I implemented it
right for the cases where the destination pixbuf is not opaque.
The test does seem to confirm it.
* gdk-pixbuf/.cvsignore:
* gdk-pixbuf/Makefile.am:
* gdk-pixbuf/test-gdk-pixbuf.c (store_pixel), (fill_with_pixel),
(load_pixel), (simple_composite_test_one),
(simple_composite_test_one_type), (simple_composite_test), (main):
Added some tests of composite to highlight problems in the old
implementation. These tests run without any user interaction.
Just do "make check".
* gdk-pixbuf/pixops/pixops.c (pixops_composite_nearest),
(composite_pixel), (composite_line): Fix composite to do a textbook
"A over B" composite. This was clearly the intent, and it was easy
to fix the code to do it. (Note to those that read my earlier tries
at a patch; this version fixes typos that were present in all
my patch attempts. I checked the final version by adding more tests.)
2000-04-11 Federico Mena Quintero <federico@helixcode.com>
Most of this patch is based on a patch by Havoc Pennington
(hp@redhat.com) to make GdkPixbuf's structures opaque and to
remove the libart dependency.
* gdk-pixbuf/gdk-pixbuf.h: Removed the public structures.
(GdkColorspace): New enum that for now only contains
GDK_COLORSPACE_RGB.
(GdkPixbufDestroyNotify): New type for the pixbuf's pixels destroy
notification function.
(GdkInterpType): New num with interpolation types.
* *.[ch]: Replace the libart stuff with our own stuff.
* pixops/*.[ch]: Likewise.
* gdk-pixbuf/gdk-pixbuf-private.h: New file with the private
declarations of the GdkPixbuf structures.
* gdk-pixbuf/gdk-pixbuf.c (gdk_pixbuf_new_from_art_pixbuf):
Removed function.
(gdk_pixbuf_get_format): Constify.
(gdk_pixbuf_get_n_channels): Constify.
(gdk_pixbuf_get_has_alpha): Constify.
(gdk_pixbuf_get_bits_per_sample): Constify.
(gdk_pixbuf_get_pixels): Constify.
(gdk_pixbuf_get_width): Constify.
(gdk_pixbuf_get_height): Constify.
(gdk_pixbuf_get_rowstride): Constify.
* gdk-pixbuf/gdk-pixbuf.c (gdk_pixbuf_copy): New function to copy
a pixbuf.
* gdk-pixbuf/gdk-pixbuf-data.c (gdk_pixbuf_new_from_data): Added a
bits_per_sample argument; currently only 8 bits per sample are
supported.
* gdk-pixbuf/gdk-pixbuf-animation.c (gdk_pixbuf_frame_get_pixbuf):
New accessor.
(gdk_pixbuf_frame_get_x_offset): New accessor.
(gdk_pixbuf_frame_get_y_offset): New accessor.
(gdk_pixbuf_frame_get_delay_time): New accessor.
(gdk_pixbuf_frame_get_action): New accessor.
* gdk-pixbuf/gdk-pixbuf-render.c
(gdk_pixbuf_render_pixmap_and_mask): Instead of returning a solid
mask rectangle for pixbufs without an alpha channel, set the
*mask_return to NULL.
* gdk-pixbuf/gdk-pixbuf-util.c (gdk_pixbuf_add_alpha): Constify.
* gdk-pixbuf/gdk-pixbuf-scale.c: Fix includes.
* gdk-pixbuf/gdk-pixbuf-scale.c (gdk_pixbuf_scale): Added some
preconditions. Maybe we should also check for the colorspace,
bits per pixel, and such.
(gdk_pixbuf_composite): Likewise.
(gdk_pixbuf_composite_color): Likewise.
(gdk_pixbuf_scale_simple): Likewise, and fail gracefully if we
cannot allocate the new pixbuf.
(gdk_pixbuf_composite_color_simple): Likewise.
* gdk-pixbuf/gnome-canvas-pixbuf.c (gnome_canvas_pixbuf_render):
Use art_rgb_rgba_affine() or art_rgb_affine() since we no longer
have an ArtPixBuf.
* gdk-pixbuf/io-bmp.c: Fix includes.
* gdk-pixbuf/pixops/pixops.c (pixops_scale_nearest): Fixed cast in
an lvalue.
* TODO: Populated.
* configure.in: Removed checks for libart.
* gdk-pixbuf/Makefile.am: Removed references to libart.
(noinst_HEADERS): Added gdk-pixbuf-private.h.
* gdk-pixbuf/Makefile.am (libgdk_pixbuf_la_LDFLAGS): Incremented
the version number of the libtool library to indicate that this
definitely is not compatible with the old usage. I know you love
me. I know you do.
* configure.in: Bumped version number to 0.7.0.
* README: Updated.
* gdk-pixbuf-config.in (--libs): We no longer require libart.
* DEPENDS.libgdk_pixbuf: We no longer depend on libart.
* gdk-pixbuf.spec.in: Updated, but I don't guarantee anything.
2000-03-27 Owen Taylor <otaylor@redhat.com>
* gdk-pixbuf/pixops/pixops.c: Fix problem with
precendence of >> and & which was keeping composite_color
from working with non-gray images.
Mon Jan 31 12:38:50 2000 Owen Taylor <otaylor@redhat.com>
* gdk-pixbuf/pixops/pixops.c (scale_line): Fix from Peter
Wainwright to fix 4a => 4a scaling.
2000-01-22 Federico Mena Quintero <federico@helixcode.com>
* configure.in: Bumped version number to 0.5.0.
* gdk-pixbuf/pixops/timescale.c (dump_array): Removed unused variable.
(main): Return 0.
* gdk-pixbuf/pixops/pixops.c (pixops_composite_color_nearest): Put
parentheses around + in a shift to remove a compiler warning.
(pixops_process): Likewise.
Patch from Ross Golder <rossigee@bigfoot.com> to create an RPM
specfile:
* configure.in: Added gdk-pixbuf.spec to AC_OUTPUT.
* Makefile.am (EXTRA_DIST): Added gdk-pixbuf.spec.in
2000-01-05 Owen Taylor <otaylor@redhat.com>
* gdk-pixbuf/pixops/pixops.c: Fixed src_channels/dest_channels
confusion. (Again caught by Tomas Ogren's Purify)
2000-01-05 Owen Taylor <otaylor@redhat.com>
* gdk-pixbuf/pixops/pixops.c (pixops_process): Fix computation of end of run indices.
* gdk-pixbuf/pixops/pixops.c: Fix double increments when dest_channels == 4.
(Pointed out by Tomas Ogren)
2000-01-05 Owen Taylor <otaylor@redhat.com>
* gdk-pixbuf/pixops/: Directory full of pixel data scaling
code that will eventually migrate into libart.
* configure.in acconfig.h: Add checks for MMX compiler support
* gdk-pixbuf/gdk-pixbuf.h gdk-pixbuf/gdk-pixbuf-scale.c:
Nice wrapper routines for the code in pixops that operate
on pixbufs instead of raw data.
* gdk-pixbuf/testpixbuf-scale: Test program for scaling
routines.
