[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Gnash-commit] [SCM] Gnash annotated tag, 0.8.10, updated. 0.8.10
|
From: |
Gabriele Giacone |
|
Subject: |
[Gnash-commit] [SCM] Gnash annotated tag, 0.8.10, updated. 0.8.10 |
|
Date: |
Fri, 01 Nov 2013 23:33:10 +0000 |
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Gnash".
The annotated tag, 0.8.10 has been updated
to 76e5b1c5ae5938650aae78fdf332432bb07c93d7 (tag)
from 2e609a60e6d1a45eaf4a4e2803b1507070602c49 (which is now obsolete)
tagging 2e609a60e6d1a45eaf4a4e2803b1507070602c49 (commit)
replaces release_0_8_9_final
tagged by Gabriele Giacone
on Sat Nov 2 00:30:26 2013 +0100
- Log -----------------------------------------------------------------
Official Gnash 0.8.10 release
Andrew Guertin (23):
List appropriate headers in libgnashvaapi_la_SOURCES
Move declarations of what's getting installed to the top
Move gnashrc and gnashpluginrc things to the bottom
Move all setting of libgnashbase_la_SOURCES together
Whitespace, useless comment, duplicated code
Add jemalloc to _SOURCES the same way as everything else
Condense libgnashbase_la_LIBADD
Move (most) conditional headers into _SOURCES
Sort _SOURCES
List headers in _SOURCES
Remove dead code
Add headers to _SOURCES in libmedia
Remove reference to unused gnash_gst_version.h
Haiku uses adipe.h, so list it
Whitespace
Group setting of each variable in libmedia Makefile.am
Only add vaapi CPPFLAGS if HAVE_VAAPI in libmedia/Makefile.am
Remove unnecessary gst include flag
-lgstbase-10 is already included in GSTREAMER_APP_LIBS
Remove no-op
Whitespace, no-op
Move headers into _SOURCES in libsound/Makefile.am
Reordering
Bastiaan Jacques (67):
Add a few more changes and a bunch of caveats.
Export the correct symbols.
Another attempt to prevent the plugin from blocking Firefox, seeing as
Fix typo. Bug #32849.
Remove unused includes, add forward declarations, use <ioswfd> where
appropriate, move some includes to implementation files.
Remove defunct macro.
Drop defunct macros.
Revert "Drop defunct macros."
Drop defunct ALP macro and tidy up in configure.ac.
Don't pretent XV is still included.
Remove stray endif.
Fix typo.
Remove the --enable-allstatic option, which does nothing.
Break the outer loop so we don't continue looking after we've found the
GTK headers.
Fix copy/paste errors.
Remove library flags presumably added to workaround typos from
Wrap AC_LANG_SOURCE in square brackets to stop recent automake
complaining.
Remove extension class loading logic that's not used.
Allow --disable-extensions to disable loading extensions to avoid
Forward-declare as_object to make cygnal happy.
Update to upstream Mozilla jemalloc (87f5bb93303c); no local changes.
Add macros to allow jemalloc to build correctly using the Gnash build
system.
Port over mallinfo(), use jemalloc stats API.
Don't define MALLOC_STATS unless specifically enabled.
Comment fix and whitespace cleanup.
Only compile the stats interface if it's enabled.
Provide pointer alignment requirements to jemalloc.
Add missing header file to fix distcheck.
mallinfo() shouln't be static so that it'll override libc's mallinfo().
Use the correct mallinfo struct member to indicate all system memory mapped.
Update for namespace changes.
Fix off-by-one that causes an invalid opcode to read past the end of the
handlers vector. Fixes bug #32950.
Add a comment regarding jemalloc's valgrind support.
Fix Darwin build.
Remove 'typedef' from struct and enum declarations to make g++ -flto -g
happy. See GCC bug #48207.
We don't use zip anymore, so don't bother looking for it.
Drop XPI packaging code which is no longer usable.
--with-cpu isn't used for anything, and you have to specify --host and
--sysroot to get a cross build anyway.
Provide a fallback implementation of NPN_ReleaseVariantValue for old
NPAPI implementations.
Remove IRIX hack that does nothing, and in any event the same test is
repeated later.
Don't automatically link libz2.
Link to liblber only if curl fails to link without it. Bug #25499.
Don't be noisy.
Only link libXi for FLTK2. Don't link libX11 unless needed.
Drop archflag.m4, because it is unused and obsolete: GCC will
Remove macro that used to be part of gettext.
Remove OpenBSD-specific build hacks, which are no longer needed and
Remove remaining macros from the previous gettext installation.
Remove remaining uses of PLUGIN_LINK_UNDEFINED, now defunct.
Remove unused test.
Don't link gstreamer or ffmpeg libraries to targets that don't use it
directly.
Don't link zlib where it's not used directly.
Don't remove core dumps from the build tree, and let automake deal with
obj files.
Remove the mudflap targets in subdirectories, because the mudflap target
in the toplevel Makefile is quite adequate, and running 'make mudflap' in
subdirectories only causes linking problems.
Link gprocessor and rtmpget only to the libraries they use directly.
Link only direct dependencies to findwebcams and findmicrophones.
The macros from libext.m4 are unused.
Remove another macro remaining from the old Gettext installation.
Return an error to the browser if there's a problem launching the Gnash
process, so we don't block in write().
Don't try scripting the plugin without NPAPI support, but still try it
without getvalueforurl. Bug #32965.
Fix null pointer dereference. Bug #32965.
Check for a C++ compiler before checking for a C compiler, so that
missing either one will abort configure. Fixes bug #31023. Bug See
http://lists.gnu.org/archive/html/bug-autoconf/2010-05/msg00001.html
Don't look for system headers we don't use.
Remove checks that define unused macros.
Prune nonexistent directories from libslist and incllist to speed up
configure.
Fix typo.
Prevent _registeredClasses from being prematurely deleted. Bug #34690.
Properly fix bug #34699: Stop including various "Device" headers
Benjamin Wolsey (318):
Use a ptr_vector
Remove commented-out code.
Update completion list.
Reduce scope of iterator.
Drop riscos and xpcom.
Implement clipboard setting.
Drop old configuration code.
Drop unused FD callback code.
Implement screenResolution for Qt4.
Drop remaining reference to HAVE_DMALLOC.
Fix mangled copyright notice.
Implement mouse wheel handling for Qt4.
Copyright typo.
Update news.
Implement screenDPI() in Qt4.
Don't require kde headers for the kde4 (really: Qt4) GUI.
Rename kde4-gnash to qt4-gnash.
Fix kde3 build.
Fix qtopia3 Makefile.
Fix klash_part.
Fix manpages for qt4-gnash.
Fix preformatted manpage.
Fix npapi plugin for qt4-gnash.
Fix rpm for qt4-gnash.
Update TODO.
Implement yesno for qt4.
Indentation.
Line length.
Use functions of Movies directly.
Drop proxy functions and cleanup.
Don't handle AS initialization depending on version;
Don't pass a movie_definition to movie_root's ctor
The displayed version is the one for the root swf, not the VM.
FFmpeg trunk uses libavcodec/version.h to store the version.
Test that a SWF5 loaded into a SWF4 still has global functions,
Move thread id code to the only place it should be used;
Link pthread libs only in libbase.
Don't use pthread cflags except in libbase.
Don't link to everything imaginable.
Drop obsolete linker flag.
Don't include everything imaginable.
There are no OGG_CFLAGS.
Drop unneeded CFLAGS.
Don't link to libglib unless using gtk. Don't link
Don't link to libs unnecessarily.
Don't use CFLAGS unnecessarily.
Don't link to or include things without reason.
Change default server for rtmpy test.
Link OpenGL libs in libbase again.
Add OpenGL cflags back in too.
Use C++-style function signature.
Fix to openBSD's testsuite.
Revert "Don't use pthread cflags except in libbase,
Get movie_root in the most direct way.
Don't declare or include as_environment where not needed.
Don't use an exception for exclusively local error handling.
Drop unused member.
Don't generate random bools when we want random bytes.
Initialize member.
Include order.
Initialize member in copy ctor.
Fix some cppcheck warnings.
Fix more cppcheck warnings.
Pass by reference (cppcheck warning).
Initialize members in ctor.
Fix operator typo in assert.
Initialize members in ctor.
Tests for copyPixels.
Make updateObjects() public.
Implement copyPixels()
Test BitmapData.copyChannels().
Update news.
Implement BitmapData.copyChannel(); tests pass.
Implement BitmapData.noise.
Cleanup.
Fix copyPixels() for overlapping ranges
Fix copy paste error
Try constructing a boost::zip_iterator
Use std::transform in copyChannel().
Add another test for copyChannel().
Update news.
Add test for overlapping copyChannel() ranges.
Make opcode output operators less verbose
Make the noise RNG more generic
Separate Video DisplayObject and AS implementation.
Put intrusive functions in ref_counted.h
Fixes to XML properties.
Append node before setting attributes.
Fix other idmap tests.
Cleanup and expect swfdec passes.
Update NEWS.
Improvements to XML and XMLNode
Handle _focusrect property correctly.
Add our own tests for _focusrect.
Test null _focusrect for MovieClips too.
Improve floodFill speed considerably.
Make number tests pass.
Test and correct case-insensitivity.
Split long line.
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
Update README with recent information.
Installation instructions are in INSTALL.
Drop obsolete comment.
Print XMLNodes with an empty name.
Implement XML.contentType correctly.
Cleanup XML_as.
Const correct.
Fix setTextFormat warning.
Make sure alpha is 0xff, not 0, by default.
Use Gnash coding style and drop unused members.
Don't check that a handler property is a function.
Fix warning.
Drop silly comment.
Fix __resolve lookup with prop flags.
Do not continue to advance during yesno().
Handle getter-setter and object __resolve.
Fix object-resolve for version 6 as well.
Clean up FileReference code.
Use AsBroadcaster.
Cleanups of fontlib.
Fix libva conditional.
Minor cleanups.
Show some delete problems.
Fix setMilliseconds format string.
Clean up StreamSoundBlockTag.
Handle StreamSoundHead tags in their own file.
Use legal identifier.
Log only once.
Cleanup includes
Remove obsolete comment.
Don't store as_functions in the sprite definition.
Use the colour we create.
Don't construct unused iterators.
Use the ObjectURI we locate.
Fix some new compiler warnings.
Remove unused stop_count.
Fix includes and new compiler warnings.
Fix new unused variable warnings.
Add documentation for recent changes.
Minor cleanups.
Header cleanups.
Remove silly copy-paste error.
Don't send unhandled key events to Buttons.
Fix button key listeners.
Remove unneeded member, minor renames.
Documentation, drop debugging, cleanup.
Don't use notifyEvent().
Expect pass.
Test and implement Button.getDepth().
Test more and log_unimpl.
Correct tests results.
Use sample count from mp3 data when available.
Add some notes.
Clean includes.
Add sound utilities.
Use SoundUtils.h
Use SoundUtils.h
Remove done TODOs.
Pass SoundInfos by value.
Const correct and document interface.
Pass sound data from callers, not sound handle.
Add separate storage for streaming sounds.
Add separate functions for stopping sounds.
More splitting.
Update documentation.
Include order.
Reduce duplication.
Fix pedantic error.
Make macros standard compliant.
Fix more pedantic errors.
Use C comments in C.
More pedantic errors.
Add a StreamingSound input stream and data class.
Update comments to match implementation.
Rearrange classes, restrict access.
Remove comment.
Introduce a common base class for sounds.
Use LiveSound base class and clean up.
Fix loop count.
Fix loop count and bug from previous commit.
Minor cleanups.
Don't add sound twice to playing instances.
Make things private.
Make things private.
Implement seeksamples.
Don't parse raw sound.
Don't waste space.
Drop unnecessary headers.
Don't assign false again.
Add an mp3 click track with ID3 data.
Start ID3 implementation.
Use a separate function for id3.
Update test totals.
Put ifdef back in because Lenny needs it.
Fix metadata build for Lenny.
Rationalize event handling.
Add interface to check for a streaming sound.
Merge branch 'main'
Move QueuedEvent to where it's needed.
Add click.mp3 to dist.
Use <ltdl.h>
Fix some cppcheck warnings.
Drop debugging and add note.
Drop old irritating debugging message.
Don't add sounds if there's no decoder available!
Don't multiply by 100 times too much.
Make notifyEvent nonvirtual.
Fixes to dragging.
Drop some unneeded things.
Minor cleanup.
Trim some long lines.
Split long line.
Don't waste space.
Use boost::optional and make DragState less mutable.
Add used headers.
Make MovieClip.prototype.startDrag less picky.
Test that MovieClip.startDrag works for TextFields.
Add test for ActionStopDrag opcode.
Only MovieClips have an unload handler.
Only MovieClips have constructors.
Add comments.
Queued constructors are only for MovieClips
Add an empty function.
Don't include headers unnecessarily.
Reduce rgba interface.
Add default case to keep gcc happy.
Fix swfdec test.
Expect passes from strk's fix.
Empty text has no height either.
Fix off-by-one error for DisplayObject removal.
Implement perlin noise (partially).
Optimize algorithm.
Log unimpl and fix unused variable warnings.
Add documentation.
Test what happens when there are no octaves.
Minor changes.
Do some basic colour range testing.
Implement octaves.
Implement fractal noise.
Correct non-fractal algorithm.
Implement channels.
Test channels.
More small corrections based on manual tests.
Apply alpha to greyscale if requested.
Tabs to spaces.
Handle octave offsets.
Don't use so many member functions.
Use an enum for channels.
Include the correct header for exp2.
Fix more symbol errors.
Headers and cleanup.
Update documentation and const correct.
Clean up test.
Testsuite cleanups.
Various header and namespace corrections.
More namespace and header cleanups.
Always use our own exp2 function.
Update NEWS.
Documentation.
More documentation and cleanup.
Remove another member function.
Remove unused variable.
Don't duplicate bind functionality.
Use existing boost functions.
Simplify mathematics.
Noise doesn't need to be a member.
Correct documentation and move function.
Rough but effective synchronization mechanism.
Don't access invalid sounds.
Only update last frame time when we advanced.
Tidy up a bit.
Move sound tests into a separate directory.
Add empty buffers as synchonization.
Add test for multiple sound streams.
Drop streamSamplesFetched() again.
Remove generated swf files.
Clean up.
Drop incorrect preprocessor conditionals.
We use streaming sound for synchronization now.
Add documentation.
Allow abandoning synchonization.
Document failure return.
Add comment.
Frame skipping is now handled in movie_root.
Add a trigger expression.
Add sound stream field.
Log unimplemented for a permitted case.
Minor tidying up.
Use an ostream operator<<.
Rename Point to Vector to avoid confusion.
Update documentation.
Restore old Button behaviour.
Use a typedef.
Add note about regression.
Fix RNG for old boost versions.
Remove semaphore too.
Make debugging output conditional.
Code style.
Add output operators for fill styles.
Output formatting.
Fix assertion.
Don't assign to an unused variable.
Initialize _glWidget
Use empty() not size().
Use a boost array and initialize _width and _height.
Don't assign to pluginsdir.
Remove unused assigments.
Fix cppcheck warnings.
Use empty().
Const correct.
Fix cppcheck warnings (preincrement).
Const correct.
Fix return type of operator=.
Output as int.
Drop unneeded configure tests.
Fix null pointer dereference.
Remove useless PATH_MAX declaration.
Const.
Brad Smith (3):
Remove unnecessary OpenBSD libpng detection hack.
Fix newer FFmpeg detection with pkg-config.
Properly use mkstemps() wherever it is available
Chi-Thanh Christopher Nguyen (1):
Fix ffmpeg version detection on Gentoo.
Gabriele Giacone (43):
Disable swfdec testsuite in dist.
Fix rpm packaging.
Add ~codename to deb package version.
Remove .la from gnash-dev package
Add mysql-devel to rpm builddep
Make rpm versioning follow fedora guidelines
Fix files inclusion in rpm
Fix enable-renderer switch.
Don't link against MySQL libs (leftover from wallpaper fix for extensions/
Add MediaHandler rc option to usermanual.
Revert "Change default server for rtmpy test."
Lowercase renderer.
Add multiarch paths. Fixes natty build.
Packaging supports python >2.6. Fixes natty packaging.
Fix cygnal manpage.
Add multiarch paths.
Suppress last sid warning. Make cppcheck happier.
Update testcases page generator.
Obsolete usleep to gnashSleep.
Remove unused var.
Inefficient emptiness checks.
Memory leak. Replace putenv with setenv.
Redundant null check. Fix indent.
Fix build on sid with libav 0.7.1.
Typo
Remove broken message (see 89aa145).
Remove cygnal/ from SUBDIRS. Fix distcheck. See 40b80d3
Fix qt4 include paths - see bug #34452.
Fix qt4 include paths again.
Update gitignore.
Fix and update translations.
Fix update-po. See bug #34654.
Another qt4 w/o opengl fix.
qt4 fix for lenny build.
If revno.h already exists and we're not in git, don't regenerate. If
included
Make cookie file not world-readable. Fixes http://bugs.debian.org/649384
More const for NPP_GetMIMEDescription.
Add boost headers to swfmill tests.
CVE-2011-4328 fix. mkstemps and boost::iostreams. See bug #34903
Fix for old boost and libc6 libs.
Fix boost::iostreams headers detection on openbsd.
Add multiarch paths on dpkg systems only.
Add debian translations to desktop files.
Hicham HAOUARI (7):
Update code dependencies documentation
Add libltdl development package to the list of code dependencies
Add BOOST_CFLAGS to plugin/npapi/Makefile.am
Comment hardcoded docdir value in doc/C/Makefile.am for now
Update NEXT_RELEASE in packaging/snapshot.am
Various rpm packaging fixes
Remove libklashpart.la from debian packaging
Johannes Obermayr (1):
Choose right kde4 libdir on openSUSE based systems
Jonas 'Sortie' Termansen (11):
Require headers to be installed before allowing linking. This prevents a
bug where the library is found, but no headers are found, but the compile still
continues, and not ./configure telling the user to install the library's
development files.
Require the development headers to be installed to use qt, and not just
use qt if the library is installed. Installing qt in a manner that these files
would detect, while not having the dev files around, will let configure allow
building the qt gui, but the build will break.
Recommend installing the libspeexdsp library when configuring. Note: on
debian-ish platforms, this is distributed separately from libspeex, while on
RPM-platforms they are distributed together.
Configure generates a script that installs everything it suggests.
Fixed broken out-of-dir regression in configure.ac.
Made base-attempter.sh more user-friendly.
Fixed building GTK without AGG, but rather OpenGL.
Don't memcpy data if you need to endian-swap it anyways.
Added base-attempter.sh to EXTRA_DIST in Makefile.am
Documented tu_file.h and makeFileChannel now also accepts a path.
Removed parse parameter from AudioDecoder interface.
Markus Gothe (3):
added DSOEXPORT's
added DSOEXPORT's
added DSOEXPORT
Paul Menzel (1):
README: replace bzr with git
Rob Savoye (770):
new files from branch, seperate the rendeer from the fb gui. Add OpenVG
support.
add double buffering and off screen rendering options from branch.
new files from branch, seperate the rendeer from the fb gui. Add OpenVG
support.
merge master into older OpenVG bzr branch, which was imported into git.
copy files from old bzr branch to get around merge conflicts.
migrate more missing files from old bzr branch.
hack so all the new files actually compile, but probably don't work yet.
rename and make OpenGLES1.1 specific.
rename Renderer_gles.cpp to Renderer_gles1.cpp
fix type, use ovg instead of ogv
hack on old OVG patch till it compiles, but of course won't work yet
split up value for gles1 and gles2
add initial stubs
separate out gles1 from gles2, disable Mali tools for now, use Mesa
instead
separate out gles1 from gles2, disable Mali tools for now, use Mesa
instead
separate out gles1 from gles2
separate out gles1 from gles2
add initial stubs for OpenGLES2.x support
renamed to separate gles1 from gles2
build OpenGLES2 support
add OpenVG render constant to gnashconfig
minor style cleanup
Choose a renderer
use a boost:scoped_ptr instead of a real pointer
more seperation of OpenGLES1 from OpenGLES2
only list the default libraries once
use RENDERER_OPENVG instead of HAVE_VG_OPENVG_H
minor style cleanup of conditionals
delete unused files
added gles1 option to -r
always pass the file description for the framebuffer
move method code from header to source
move method code from header to source
always pass the file description for the framebuffer
be less verbose
be gles1 specific
cleanup
use a scoped pointer
minor cleanup
handle error if fbmem memory doesn't exist.
cleanup comment
minor cleanup
use underbar for private data name
reformat
big merge from master
put everything behind a namespace
reindent
generic EGL Support for embedded Open stuff
add many more tests
add many more tests
work under X11 using GDK
add conditional for EGL
add EGLDevice, and the test_egl test case
add testing method to append RenderImages to the vector
rename draw_poly() to drawPoly() to be consistent.
rename draw_poly() to drawPoly() to be consistent.
memset() the memory for the test buffer after allocating.
Move class definition to header file. Fix indentting
add renderer and openvg namespaces
move EGL_DEPTH_SIZE to the stdnard attributes, so it works for OPenVG,
OpenGLES1, and OPenGLES2
move inside renderer::opengl namespaces
move test_egl to check_PROGRAMS, it's basically done
change order of the tests
use drawPoly() now instead of draw_poly()
Don't build EGL test case unless configured for EGL support
rename EGLDevice::init to initDevice to be more unique
use drawPoly() now instead of draw_poly()
add the EGL libraries for the test cases
Inherit Renderer_ovg from EGLDEvice, and initDevice() when constructing
the renderer
flesh out initial stubs for OPenGLES that match the Render API
split off GTK window creation and callbacks code to it's own file
create_Renderer_ovg has been renamed to create_handler
add GTK support file
use new to allocate memory for buffer, remove unused code
initial support to render in GTK window
disable GTK support for now
init EGL separate from the device
initialize EGL. now that it's a separate step
use boost to time operations
optionally print the elapsed time when destructing.
don't print the renderer test total time
draw_poly is now drawPoly
add accesors for EGL config data. Use cerr instead of log_debug for
debugging so it always prints. Use Gdk instead of Xlib for display stuff
test the new accesors
rename m_display_width * height to _display_*.
hack in simple OpenVG drawing test, for debug only
use EGL_RENDER_BUFFER to get the right values
add tests for all the new accessors
add new accessors for EGL config data
Add support for Pbuffers and native pixmaps.
add methods for using Pbuffers
renable GTK support
add pbuffer and pixmap support
big merge from master
fix paths for new subdirectories
move new files from branch into subdirs to match master.
fix spelling of cairo
rearrange noinst_HEADERS
add path to include Renderer_ogl.h
add namespace and path for openvg
Don't create OpenVG renderer for now
include the fb.am file again
restore using fb.am
remove rendere includes
stubs for adding DirectFB support.
use size_t instead of int
add DirectFB support
initialDirectFB support and test case
use size_t instead of EGLint
build the DirectFB test case
print out the various config bits
make initDevice() work. Add more silly debug function for more data
structuress.
Add more test to match the egl tests
add spport for accessing more of the low level data about Displays,
Screen, etc...
start movingcode from libvaapi to here
change the name of the directfb conditional to BUILD_DIRECTF_DEVICE
add printDisplayLayerConfig()
start moving vaapi suppor from libvaapi
remove extra files
build vaapi device support
X11 support for Device class
convert old hwaccel option to enable-device option, now handles vaapi,
egl, directfb
implement printing X11 errors, add test case
base class for devices
initialize data in copnstructor. Add support and tests for pixel format
Add support for pixel format
add base class header for Devices
rename x11 to has_x11 to be more clear
specify a default value for AC_DEFINE()
use proper ifdefs for X11 includes. USe Root window as default window
make native window type typedef a long instead of a void *
derive from base Gnashdevice class now. Add pixel format support. Make
more consistent with other Devices API
initDevice now takes a argc, argv to be consistent.
use the proper HAVE_X11*
minor reindenting
add comments about EGL initialization process.
initDevice() now takes argc, argv
work as a derived class of GnashDevice now.
add some comments
exit if the deviuce can't be initialized
add constant name to debug config print methods
always attempt to find egl and directfb by default
minor editing of print statements
remove useless comment
don't store width, height, or bpp, it's better to look them up from the
surface
new glue file so the FB follows the same conventions as the other GUIs
drop gui_fb.h, add fb_glue.h. Don't try to build GLES support yet.
use the gui namespace to access createFBGui
wrap the prototype for createFBGuie in the guio namespace
add emacs comment block
don't use virtual for a derived class
move AGG specific stuff to fb_glue_agg.cpp
add all methods reqired by the base class
add comments
add cteareCacheBitmap back in, plus test case
add the device headers to CPPFL:AGS
add X11 device constant
add the device headers to CPPFL:AGS
add device test, drop using GTK
we don't want any render specific code in here.
return the device type of the instantiated class
return the device type of the instantiated class
add new required method for getType()
add support for querying display devices
don't try to use a device that isn't initialized
use C++ style comments
add comment blocks, add gui namespace
open an X11 window and start an event_loop. These are only used for
testing, and use Xlib instead of GTK for potrability
add more constructors to reduce the API steps
create an X11 window with an EGL visual
handle X11 visuals, as the Mesa EGL runs on X11
add more comments
don't call EGLDevice.attachWindow() for now.
start documentation on how rendering works.
add more explanation of devices.
add doxygen style comments
see if we can render VG in our test window
build devices as libgnashdevice to get ready to move this to it's own
subdirectory.
remove gdk_exit(), it doesn't belong here.
add libgnashdevice to the list
add BOOST_LIBS to the link list
GnashTexture moved fom libbase to libreander, as it's for OpenGL. Cleanup
link lists.
add comment that on X11, you can only use double buffering
improve the test for getRedSize()
move OPenGL specific file to librender
add libdevice
link in libgnashdevice
remove al the device code, it's now in it's own libdevice directory
moved to libdevice
add some libcore paths to CPFLAGS, remove all librender references
don't use the red pixel size test yet
moved from librender
elinate compiler warning
don't configure optional directories unless enabled
add libdevices to CPP paths
add libdevices to CPP paths
eglDevices.h is in libdevice now
rearrange files to make devices code in it's own dir
libvaapi is now under libdevice/vaapi
move libvaapi under libdevices
start getting vaapi into the standad API
eglDevices.h is in libdevices/egl now
test case for vaapi device
work towards surviving distcheck
add headers to _SOURCES to make distcheck happy
recreate the old libvaapi Makefile as a fragment
don't FAIL all the teets if initDevice() doesn't work, make them untested
instead.
don't try to access any devicde specific stuff if it's not initialized
add private variable for the glue layer.
cleanup formatting to be more readable
initial files for raw framebuffer device
add raw framebuffer device
change from renderer to device namespace now that this code is in it's
own directory
Revert "change from renderer to device namespace now that this code is in
it's own directory"
add a raw framebuffer to the device enum
fix namespace mess
implement support for a raw framebuffer
add a raw framebuffer to the output device list
add test case for the glue API
add accessors for the memory and it's size
minor indenting fix
add namespcae for GnashDevice
minor reformatting change
output device code move to the glue layer
fix typo
rearrange heavily where code lives, as it's being separated into the
rawFB device layer and the glue layer from fb.cpp
initialize the device before the renderer
fix typo is test, yes needs to be xyes
remove direct device access. Add some trace starement for debugging
move colormap code to RawFBDevice
move colormap code to RawFBDevice
move colormap code to RawFBDevice
how did this get in there ?
be less verbose
add support for a manual line drawing test to make sure initialization
actually worked.
fix single buffer test to be correct, DirectFB is double buffered by
default, unlike everything else
add support for a manual line drawing test to make sure initialization
actually worked.
hack on the attributes to get it working with a 16bit depth on real HW
with no X11
move openvg to it's own test, it now handles multiple OpenVG versions.
add OpenVG debug display at end. COmment out libltdl till we fix it
right. Limit the set of default devices to what the babbage board supports
rearrange headers for different OpenVG versions
migrate libpng patch from master
diable libltdl for not causes problems on ltib
fix typo in libgnashrender_la_CPPFLAGS
currently when cross compiling for ltib with an external toolchain, more
libraries need to be added at link time
reenable testr test case again
fix OpenVG headers
ifdef out a few tests that don't deem to exists on ltib
minor tweaks
don't try to link librender to the test.
for now always build test cases.
apply patch from master to disable libltdl
add memory access accesors
add eglinfo so we have a version for our target
move most of the target code to makefile fragments to be more readable.
fix test
add more include paths as some of the input events need them for deeply
included headers bu other headers
InputDevice.h is now in libdevice/events
fix OPENVG_CFLAGS for the two version, 1.0.1 with libMesa, 1.1.2 with
ltib.
tweak x11 device config
drop the usage of the Gui class, instead just queue up event data to be
retrieved by the Gui later.
new test cases
use a std::queue for all input events
Use 0x7 instead of 0x1 for the mouse button mask, so it works with 3
button mice.
add simple test case for input events.
moved to libdevice/events/
input events are now in libdevice
scanForDevices() doesn't need the Gui argument anymore. ::run() now polls
the input devices in the main loop, and notify's the Gui of events.
use new method for coordinate range checking, add some more documentation
also look for a FAKE_FRAMEBUFFER in the environment, and use that for the
device to open.
fix X11 device configury
add TS_LIBS so we can use a touchscreen
mostly just reformatting
add the gui namespace
fix spelling of --enable-input option. Add --enable-devtests, which
enables developer only tests.
make some tests only build with --enable-devtests
update with the one from master
base class for other classes that need to handle deisplay devices.
Add OpenVG support. Move device handling code to DeviceGlue for use by
GUIs
OpenVG glue for GTK. This requires libMesa
return the window, not the root display
look for Freescale Babbage board device names for input events
add new DeviceGlue.h file
add gtk/gtk_glue_ovg.* files for OpenVG with GTK support
use Ovg instead of egl in file names
Use OpenVG instead of AGG
add newline at end of file
remove unused device header files, this is now handled by the glue layer
add the two new header files
fix --enable-device= option
fix all test cases to match reality now that we have real hardware that
supports EGL in a framebuffrer
tweak EGL config attributes to work on a real 16bit display instead of a
fake one.
remove trailing slash on filename
add accesors for the devices so it's easier toi use in a derived class.
tweak fake framebuffer settings to match our real 16bit LCD. Move mmap()
from initDevice() to attachWindow(), so it's consistent with EGL
remove debug crap
free the XVisual after using it.
add EGL support to run OpenVG on the desktop
add OpenVG renderer support to the framebuffer GUI
add OpenVG renderer support to the framebuffer GUI
use a 32bpp on the desktop, 16bpp on embedded
less versbose
fix indenting
add comments
work with on;y OpenVG, no AGG
fix device handling for the framebuffer
getFBHandle is now just getHandle
add getHandle()
temporaily ifdef out calibration untill we can get the stage size
add swapBuffers()
add the renderer
initialize the renderer
add doxygen style comments
use current drawShape() API, not the old API.
add helper classes for boost::variant. Implement all the types in
apply_fill_style
add libsound.la to eliminate linking issues under ltib.
derived methods aren't virtual
add raw framebuffer support, fix tests to match rality of real OpenVG on
a real 16bit touchscreen
correct note, bitmap_info is now CachedBitmap
clear the device list before appending to it so we don't get duplicates
return false for bindClient() and swapBuffer() if not derived
getDrawableWindow() has been replaced by getHandle()
handle multiple attributes fot the EGL configuration so they can be
changed dynamically.
implement GetType() to return the Fill style associated with the style.
derive from CachedBitmap
move helper functions for Fill Styles to it's own file
rename to follow the Gnash style
derived classes don't need the virtual keyword
rename to follow the Gnash style
move code from header to be easier to debug and more compact
move code from header to be easier to debug and more compact
move code to header to be easier to debug and more compact
add new files
swapPbuffers() was renamed to swapBuffers()
rename Renderer_ovg.* to be OpenVGRenderer to be more in the Gnash style
minor editing tweaks
use nil not t for emacs
openvg/Renderer_ovg.h is now openvg/OpenVGRenderer.h
openvg/Renderer_ovg.h is now openvg/OpenVGRenderer.h
openvg/Renderer_ovg.h is now openvg/OpenVGRenderer.h
initialize the surface, use hardware gradients instad of crappy software
ones ripped off from OpenGL. Handle embedded bitmaps
add support for hardware gradients
openvg/Renderer_ovg.h is now openvg/OpenVGRenderer.h
add range checking to avoid false positives
more all the attribute tables to their own file
more all the attribute tables to their own file
more all the attribute tables to their own file. Try multiple EGL
configurations.
if attachWindow() fails, nothing else will work properly.
tweak EGL config so it works on both 32 and 17bpp displays, with both
OpenVG 1.0 and OpenVG 1.1
clean up formatting
don't swap if single buffered
call prepDrawingArea() for OPenVG too
add accessor for variant to return GnahImage
get image handling of bitmaps working
update comment
Add comments.
be less verbose
add accessor for the display depth from the device
be less verbose
be less verbose
use a back buffer instead of single buffering
be less verbose
clear the display before rendering the next frame
reset input devices to no, or we can't select just one
add evtest from freedesktop.org
Select a subset of possible input devices, in our case, just the Babbage
touchscreen.
only minor reformatting changes
libdevice/events/InputDevice.cpp
add more event types
properly support Linux input events types that use absolute position
reporting. Initize the touchscreen input_absinfo.
openvg.h, not open.h
nil not NIL for emacs block
clear the entire display, not just the size of the swf being renderered.
improve 32/16 bit handling for bitmaps
simple test case for gradients, not designed for automated testing.
implement working linear gradients via OpenVG hardware acceleration.
use more ramp stops
dump the value of the Alpha Mask Size
don't try to mask if it's disabled. Fx radial gradients to use the color.
move some code from the source to the header, so the static functions for
scanning devices work without Linux Input events.
Renderer_ovg.h is now OpenVGRenderer.h
don't dump
remove commented out lines
fix bitmap loading
merge apply() into createPatternBitmap()
add rcfile, so the renderer is configurable at runtime. Use tslib with
Babbage tablet instead of linux input events.
cleanup position handling, as we can use ts_calibrate instead of our own
move scanDevices() from the header back to the src
don't convert mouse locations when using tslib, as tslib does it for us.
only minor formatting changes
cleanup image creation
add doxygen comments
improve formatting
fix comment
get the ramp data from gnash to do hardware accelerated gradient correctly
add accesor to get the vector of Gradient Records.
minor reformatting only
add private data for the aspect ratio, so we can adjust bitmaps to be
correct.
enable debug message, remove blank lines
reformatting
remove extra FBGlue constructor/destructor
Use CXForm for all bitmaps.
disable the defaults when enable-device is specified
initialize AGG renderer and set the size
add EGL_CFLAGS for all egl code
return something
destroy the surrace if it exists
add dejagu.h to the dist
merge from release_0_8_9 branch to get all copyright changes.
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
big merge from release branch to get mostly just copyright date changes
merge from master for rpm changes
huge merge to catch up with master, still a few changed API bugs to fix
fix function definition
adjust namespace definitions
merge from master
check for gles1 and gles2 errors separetely
reenable Texture support for OpenGL, look in libdevice for headers
remove stubs
remove gles2 as it will be better to recreate in the far future
remove gles2 as it will be better to recreate in the far future
pass in defaut argument
rename create_Renderer_ogl to create_handler
make const correct to work with the latest Gnash internal rendering API
drop gles2 header
cleanup warnings about unused variables
drop original version that is unused
make whether the drawing area needs to be intitizlized when realized
dynamic, rather than a compile time option.
add error message fr bad window handle
drop references to gles2
fix config ifdef for DirectFB support
merge with master
add utilities back into DIST_SUBDIRS
convert dump/Makefile.am to dump/dump.am
create_Renderer_ogl is now renderer::opengl::create_handler
disable developer only tests by default
drawPoly() is still draw_poly() in this branch
silence warning
PTHREAD_LIBS should always follow BOOST_LIBS
add LTDL_LIBS to the link
add utilities back in as a STD_DIR
commit file #23105 for patch #7523
Merge branch 'master' into hwaccel
merge from master
merge from master
remove smart_ptr.h include
big merge from master
add LIBVA_LIBS to LDADD if VAAPI is enabled
add LIBVA_LIBS and LIBVA_X11_LIBS to LDADD if VAAPI is enabled to the
Cairo glue too
add default for gtk_gnash_LDADD, renders add only those libs
rtmpget needs PTHREAD_LIBS on natty
remove fluendo mp3 plugin
don't prefix va with lib when finging libva support
support AGG again in additional to OpenVG. Use the command line args to
change renderers.
update copyright date
cleanup to eliminate compiler warnings
merge from master
VaapiImageFormat.h is in vaapi now
merge from master
don't initialize OpenVG like OpenGL
throw an exception if we can't create an OpenVG context, as nothing will
work anyway
throw an exception if we can't make the context current. Add more comments
use HAVE_VA_VA_H instead of RENDERER_AGG for including the vaapi glue
header
vaapi headers need vaapi/ prefixed to the name.
Gnash*Texture files moved to librender
revert libva changes till a better solutuon is found
fix typo in help string
use found_libva_incl instead of found_va_incl for the VAAPI conditional
don't always add the libva GLX library
cleanup handling of libva libraries so they work for all guis.
throw and exception instead of returning an error if the window handle is
bogus.
fix test for pkgroot
define boolean as int when using mingw32
only define boolean for jpeg 7
pass argc & argv to init. Add comments
merge from master
merge from master
Merge branch 'master' into hwaccel
move #endif to not drop the closing brace
initialize the4 renderer before preping the drawing area.
less verbose
less verbose
Merge branch 'master' into hwaccel
attach the window to the device
Add support for the GnashDevice display
use createRenderHandler() instead of create_Renderer_agg so it works with
the new API
don't build OpenVG support if no dev files found.
remove commented out code blocks
handle the user error of openvg being specified, but not having any dev
files.
Use OPENVG_* instead of OVG_* for debug message
eliminate warnings, even though this is just a debugging app for
developers.
eliminate compiler warnings
add README explaining the DirectFB code needs to be fully implemented.
fixed unused compiler warnings and update the copyright date.
update the copyright date.
update the copyright date.
update the copyright date.
adjust the order of variable initialization in the constructors. Update
the copyright date, and get rid of a few compiler warnings.
update the copyright date.
we always need OPENVG_STATIC_LIBRARY defined in CXXFLAGS
use a different config constant for OpenVG headers.
create_Renderer_ogl() is now openvg::create_handler()
don't print Matrix debug info
be less verbose
oneric moves libOpenVG into it's own directory
add comment about oneric and openvg library locations.
fix comment typos
merge from master
eliminate unused variable warning
eliminate unused variable warning
create_Renderer_ogl is now renderer::opengl::create_handler
merge from master
Merge branch 'master' into hwaccel
Merge branch 'master' into hwaccel
eliminate warnings, even though this is mostly just stubs for future de
elopment.
drop a few blank lines
add 2011 to copyright date.
eliminate warning about unused variable
Merge branch 'master' into hwaccel
remove style changes, they're now in a new branch instead
don't build the fb gui nor EGL by default, only if openvg is soecified.
add option to disable debug logging, which is a performance hit
disable logging if configured that way
it helps if the option is spelled right.
Revert "it helps if the option is spelled right."
Revert "disable logging if configured that way"
add comments, tweak dimensions for debugging.
remove extra blank lines
tweak privates & constructor. Implement linear gradients
use StyleHandler for gradients
fix bitmap handling to work with StyleHandler. Remove all accessor style
classes
support bitmaps, improve gradient support
fix comments
fix handling of solid fills, use StyleHandler, remove unused
apply_fill_style
less verbose
less verbose
merge from style branch
Merge branch 'hwaccel'
use dump.am instead of dump as a SUBDIR
using += in bourne shell is not portable. Fixes natty-ARM builds
the pythonmod also needs a path to libdevice
optional utilities need cygnal to be built first.
drop the pythonmod for packages, as pyGTK has been depreciated.
minor style changes in brace placement
change t to nil for emacs tab-mode
add new option to force a mouse pointer so touchscreen SWF files work on
the desktop with a mouse.
Use new option to optionally force the mouse pointer to be visible.
minor style changes in brace placement
build all devices too for distcheck
build the raw device if the FB GUI and AGG are to be built
add test for resolution field for input events
include errno to keep Lenny happy
include errno to keep Lenny happy, and optionally use inpout event field
for resolution
Ubuntu patch for recent ffmpeg versions merged in by hand, as it was
based on an older release.
fixes for other ffmpeg versions and distributions
more fixes for other ffmpeg versions and distributions
include avcodec.h if it exists
change order of ifdef
use LIBAVFORMAT_VERSION_MAJOR instead of LIBAVUTIL_VERSION_INT
optionally include avcodec.h
minor version tweaks to get Lenny to build with the patch.
revert merge mess
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
Don't do anything if we have no timers, just return so we don't waste cpu
cycles
increase the default delay to 50ms
only look for an invoke packet if _controlfd is actually set
use a regualr iterator instead of std::for_each, so we can do loop
through all the liveChars only once
set the delay back to 10ms
don't use AC_MSG_ERROR, as it exits after message.
forceably remove ubuntu patch as the revert seems to have not worked
correctly.
Merge branch 'master' into noise
don't build the OpenGL renderer by default, as it suffers from both
performance and rendering quality issues.
Don't configure in Linux input events on any BSD distro. For bug #34373.
fix typo, only build inout events on Linux
add a few new things
gtkglext is unneeded, as the OpenGL support is no longer built by default.
Always build libdevice/Makefile so distcheck is happy. Add Make
conditional for LINUX. Fix inpout_event list text
Use conditional for LINUX so input devices are only built on Linux
systems. Put cygnal libdevice in DIST_SUBDIRS or they won't get included in the
source tarball.
Only have one AC_OUTPUT, as otherwise the optional directories cause
distcheck to fail.
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
minor reformatting
for distcheck, set cxxflags to -O0 to speed up builds. Only use SUBDIRS,
as automake groks the correct optional directories. Otherwise directories
appear twice, causing distcheck to fail.
Don't disable optimization for distcheck
clarify that software double buffering is for AGG only.
enable double buffering by default for AGG with a framebuffer
fix typos in comment
add virtual accessor to get the offscreen buffer memory.
add accessors for framebuffer and ofscreen buffer
clear the offscreen buffer in addition to the framebuffer
tweak indenting, minor change
update from latest builds
tweaks for the OLPC XO 1.75, which doesn't distribute kde or qt
minor reformatting changes only
merge with master
Merge branch 'master' into noise
Merge branch 'master' into noise
set the delay timeout to the frame rate
Use the default of 10ms for low framerate files
Use __null instead of NULL for pointers
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
include Point2d.h so EGL devices still builds after other changes that
broke compilation
remove debug accessor for drawbounds, as it's unused
add comment about the interval
add comment about the threshold value
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
Don't create the shared object directory if they're read only, bug #34837
add xulrunner-dev or firefox-dev for Build Deps
const is needed for some of the NPAPI functions for Oneiric with
firefox-dev
const is needed for some of the NPAPI functions for Oneiric with
firefox-dev
BZ2_LIBS needs to be in the link when using --as-needed
Fix #34708, but breaks #32956. YouTube video player is now the correct
size
Merge branch 'master' into noise
Buffer is in cygnal namespace now, not amf
Buffer is in cygnal namespace now, not amf
Don't do anything if we have no timers.
Cleanup compilation warnings when using -pedantic for #34911
Cleanup compilation warnings when using -pedantic
cleanup various compilation warnings and errors so make check works with
-pedantic
don't manually draw a line for now
add test for the stride
Add swapBuffers() for double buffering. USe a standard pointer for
_fbmem, so boost doesn't try to free it. Fix test case
add comments about swapBuffers() beyhaviour
add getStride() to base class
make isSingleBuffered() really work.
use getStride() instead of calculating it ourselves
use isSingleBuffered() instead of ifdef
remove commented out lines
fix copyright date
remove leftover double buffering code that is now replaced by
swapBuffers()
add getStride() as it's now in the base class
cleanup AGG double buffering to work with libdevice
use the protected GnashDevice from the base class, instead of our own.
merge from master
use isSingleBuffered() instead of getOffscreenBuffer()
move swapBuffers() from header to source file
remove unused code that's now in swapBuffers(). Remove unused local copy
of the framebuffer data, it's now in RawFBDevice
be less verbose
mention improved double buffering support
reenable InvalidatedRegions for AGG in a framebuffer
add resize_view()
add resize_view()
reenable fullscreen support
add accesor to get the movie's FPS
use the movie frame rate to adjust the delay between frames. Also make
the default interval larger, too small a value eats the cpu
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
merge from noise branch so adjusting the time interval works the same way
for the framebuffer
add support for EV_MSC events when used for keyboard input
Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash
add initial support for uinout, ie.. User Mode mouse control, so we can
move the cursor.
new file for the Uinput device
add .gdb_history
add events/UinputDevice.cpp to build
add a UInput device to the base class.
use new uinput device to control the cursor
don't enable the PS/2 mouse by default, we use input events for the mouse
instead.
cleanup optional PS/2 mouse handlingm use InputDevice::convertAbsCoords()
instead of MouseDevice::convertCoordinates()
fix input device mouse support.
use nil instead of t for emacs indent-tabs-mode
Merge branch 'dbbuf'
initialize netfd to eliminate compiler warning
only dump data when -d is specified
eliminate compiler warnings
eliminate compiler warnings
initialize private data
eliminate compiler warnings
build libdevice before libbase so the vaapi support gets built first
use correct path to vaapi headers and library
add GnashVaapiTexture.cpp back into the build.
cygnal needs libgnashrender if vaap is built.
only build EGL tests when specified
drop dependency on libgnashbase.la
use the right conditional
add GNASH_LIBS to AM_LDFLAGS
only build developer tests when specified.
add lingnashrender to the link
don't make EGL dependent on OpenVG
move endif so EGL builds
Don't build OpenVG as part of =all
only build developer tests if configured that way
return an error if no message received
Build all devices and renderers for distcheck
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
updated with new strings
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
updated with new strings
add macro so new strings get found by xgettext, so they can be translated.
updated with new strings
updated with new strings
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
add macro so new strings get found by xgettext, so they can be translated.
updated with new strings
updated with new strings
print messages cleaned up for translations
print messages cleaned up for translations
updated with new strings
cleanup code style
updated with new strings
print messages cleaned up for translations
print messages cleaned up for translations
updated with new strings
add gettext macro to all log_error() messages.
change w/out to without
print messages cleaned up for translations
updated with new strings
Capitalize URL
print messages cleaned up for translations
fix typos
updated with new strings
fix minor typos
fix minor typos
updated with new strings
advance all characters before notifying them of loads.
add new translations, even though they are incomplete.
update translations
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
update copyright date
remove extraneous = from help strings option names
fix typo with gettext macro
regenerated the preformatted versions
build a dump-gnash manpage too, since we now package it.
build fb-gnash and dump-gnash packages too.
add dump-gnash as an aliases to this manpage
add dump-gnash manpage
drop unused rules for optional package buulds, everything is selectable
at runtime now.
add fb-gnash and dump-gnash packages
drop duplicate copies
install the dump-gnash man page
update for new packages
remove extraneous man page
aliasesdoc/C/preformatted/gnash-gtk-launcher.1.in
install manpage aliases as files, not links
don't ignore man pages
cleanup the handling of man pages
remove unused preformatted tjay are now aliases the correct way
install-man1 doesn't need to depend on EXTRAMANPAGES
install-man1 doesn't need to depend on EXTRAMANPAGES
fix typo in filename
enable docbook, don't gzip man pages
enable docbook, install info pages
update copyright date, remove commented out block.
add man page alises file to the dist
keep NEXT_RELEASE at 0.8.10 till after the release
update en_GB translation, regenerate pot file.
correct typos
correct typos
fix typo
tweak copyright message
remove extraneous spaces in string
change version to 0.8.10
change version to 0.8.10
revert time interval/FPS change for now.
Do not edit /usr/lib out of the path, fremove the entire -L option path
instead. For bug #35300
Merge branch 'master' into release_0_8_10
generate GNU style ChangeLog from git-log for release.
Sandro Santilli (183):
Test for bug #32833
Automate runner for test of bug #32833
Fix bug #32833 for AGG renderer (render missing bitmap fills in solid red)
Revert the change about missing bitmaps as it broke another test. Leave
comments about the case right in the code.
Have AGG renderer handle disposed bitmaps and missing bitmaps
differently, fixing #32833 w/out breaking other things.
Fix the missing and disposed bitmap in bitmapfill case for cairo too (bug
#32842)
Distribute missing_bitmap.xml
Set release date...
Prepare for next turn (0.8.10)
Do not advertise existance of non-existent gles renderer (single-line
change))
Fix misleading comments about --hwaccel and --renderer
Do not advertise unavailable omap hwaccel. Fixes ticket #32888
Have GTK gui always trust RunResource about hwaccel and renderer. Have
commandline parser use rcfile setting as a default, or first supported renderer
otherwise. Fixes #32925 in a safer way.
Have QT4 gui handle non-existent renderers by throwing an exception
rather than assuming AGG. Have it trust RunResources to keep rcfile handling
centralized.
As for renderer, read rcfile for the default mediahandler too. Fixes
#32859.
Query movie_root for action execution timeout configuration.
Handle ActionLimit exceptions thrown during movie advancement like the
others, that is virtually asking user if he wants to abort execution.
Implement yesno for GTK gui. Visible effect: you now get a popup when
script execution takes over the time limit (default: 15 secs).
Add note about the improved script limit user interface
Add support for 'scriptsTimeout', 'scriptsRecursionLimit' and
'lockScriptLimits' rc directives, testcase their support and use them from
movie_root.
Document scriptsTimeout, scriptsRecursionLimit and lockScriptLimits
rcfile directives.
Don't use an arbitrary string as a format
Add support for scriptTimeouts in GTK preference window
Don't write integers rcdirectives in hex form (after LCKey, that is)
Complete GTK preferences support for script limits
Reworded NEWS item about script limits
_really_ continue on scriptTimeout if the user chooses not to disable
scripts.
Make sure testsuite run is not affected by user-defined script limits
The Rc test can't really test default values, as there's no way to
construct an RcInit file w/out having it read the user rcfile :/ This commit
sets the default values for script limits
Do not execute any actionscript outside of the main loop. Fixes bug
#32956.
Change the default urlOpenerFormat from "firefox -remote 'openurl(%u)'"
to "xdg-open '%u'"
Add binary test for https://savannah.gnu.org/bugs/index.php?32950
Automatically run test for https://savannah.gnu.org/bugs/index.php?32950
Fix testcase to use opcode 0xFF, not 0xFE (this way it's confirmed to
segfault pre bastiaan fix)
Distribute test_lastopcode_v6.swf, as we're using it during make check.
ignore backup files (most notably gnashconfig.h~)
Fix build: create_Renderer_ogl became renderer::opengl::create_handler
Define output operator for SWF::TagType
Fix build with DEBUG_SET_INVALIDATED defined
safer path to string_table (invalidated ranges debugging)
Print DisplayObject target rather than just name (st_invalidated
debugging)
cleanup/indenting
movie_root doesn't need MovieFactory.h
Make sure to destroy movie_root before clearing the MovieFactory
(library). This prevents MovieLoader (owned by movie_root) adding more items to
the library, thus keeping SWF loaders alive and keeping Curl busy, with the
consequence of aborts or segfaults on exit. Fixes bug #32579
Comments cleanup
Drop unused class members
Check that invalidated ranges don't contain spot which aren't updated.
This is an xfail in a jump-back case.
Do not invalidate MovieClip on timeline rewind unless DisplayList really
changes. Takes CPU use down in many cases (seen it from 100% to 8%)
expect the new success in invalidated bounds detection
Typo in error message
Play all movies given on command line, serializing them. This is a basis
for furhter evaluating leftover singletons. Not all GUIs are handling this the
same way. GTK, in particular, seems to be getting old windows stuck.
Use a separate Player for each movie on command line. Makes GTK gui
survive, and it's cleaner anyway.
Cleanup rules to build kde3-gnash. Fixes bug #31782
Look for qxembed.h (used) rather than kapp.h or kapplication.h (unused)
to tell if we're good to go kde3 with gui.
Print KDE4_CFLAGS and KDE4_LIBS when building klash4 plugin (they will be
used)
Have gprocessor fetch samples from the NullSoundHandler, to allow testing
for sound events. Do not expect the relative failure in Sound.as anymore.
Add test for bug #31314 (Sound.onLoad, google dict). Expected to fail.
Test that Sound.onLoad is called with an argument, true when load
succeeded.
Add callback for Sound.onLoad
Implement onLoad callback for AS Sound object. Fixes #31314.
Add note about Sound.onLoad and google dict
Some steps forward in MediaParserFfmpeg::seek(). Fixes the google dict
case (ticket #33085) and shouldn't break anyting since prior to this commit the
function simply returned false...
Do not stop probeTimer while playing embedded sound. Fixes bug #33091.
Add "duration" and "position" named strings
Use named strings for "duration" and "position"
Add test for Sound.postition and Sound.duration (they fail for loaded mp3)
Add test for Sound.position and Sound.duration when playing embedded
sound (tests fix for bug #33091)
Drop unused variables
indent
Add interface to get nominal frame rate from MovieTester
Fix build scripts for dump gui (currently only builds if FB gui is
selected :/)
Do not try to write to a non-opened stream
Print debug lines for ActionPushData to really print what's pushed, not
what was on the stack _before_ pushing...
Tweak debug lines
Add automated test for embedded event sound .position, aka "tictac"
(https://savannah.gnu.org/bugs/?33091)
Fix leak on sound_handler destructor (only exposed when using
NullSoundHandler, that is test runners...)
Add automated test for bug #33176 (tags after last advertised frame)
Make success condition pattern of last test (tags after last showframe)
stricter (I've seen it give a false success otherwise)
Flush "orphaned" tags (tags found after last expected showframe) right
before executing tags of first frame. Fixes bug #33176 (SWF9/AVM1/April3.swf)
Properly distribute tags_after_last_showframe.xml testcase
Expect success in test for frames after last expected showframe
Fix signed vs. unsigned comparison warning
Fix typos reported in bug #33264
Add note about compatibility fix (#33176)
Do not complain about unexistent media handler if none was specified.
Fixes bug #33339.
Do not call .front() on an empty vector (the media handler list when
--disable-media is given). Fixes bug #33338
Put CPP code (rather than C) in revno.h, to reuse from ActionScript
Pass -I$(top_builddir) to makeswf when building AS tests, to give
check.as a chance to find revno.h
Encode dependency of every SWF file in actionscript.all on check.as and
dejagnu_{init,fini}.as
Print revision info in addition to rcsid
Reduce rcsid to the sole base filename. Actual revision will be
automatically printed by check.as
Another check.as user requiring proper -I to find revno.h ...
Fix MAKESWF flags for online-tests
Re-enable building index for online tests, for some reason removed in
commit 6afad08f5ac4416ac0205cd3c2df8ef703adc531 (which was unrelated...)
Flushing of orphaned tags belong to action (not debug) logging
Implement onSoundComplete for embedded event sounds, fixing bug #23020
Use number of milliseconds since start as the log timestamp.
Check that onSoundComplete is called at least once when all sounds are
finished playing (test for #23020)
USE_VAAPI conditional is not being used, so don't define it... This is
trying to reduce confusion while debugging bug #33393
Missing space in error message
Use std::strerror to tell more about std::fopen failures
Apply contributed patch to documentation to note gnashrc directives are
case-insensitive [bug #33684]
Try UTF-8 character in URL string, trying to reproduce bug #33718
(unsuccessfully)
Don't need c_str() to pass a string to our logger functions
Don't need c_str() to pass a string to our logger functions (again)
include <cstring> for strerror (see
http://gnashdev.org:8010/builders/lenny-linux-x86/builds/525/steps/compile/logs/stdio)
Add test for bug #33521 (constant pool used by functions). Fails.
Fix runner for the new func_dict.xml test, add an afunc_dict.xml test for
testing the constant pool bug with anonymous functions.
Add another constant pool (dictionary) test. Showing more gnash bugs
about it.
Typos
Tabs to spaces
Improve the test for dictionary use in function to show that when coming
back from functions, the previous dictionary is again in effect
Log once about SetTabIndexTag being unimplemented
Add SeaMonkey to the list of browsers known to work (as reported by newl
in IRC). Reword the section a little bit.
Add End actions between DoAction blocks to show that availability of
constant pool to different DoAction blocks is not due to some side-effect of
missing End opcodes
Unexpect ConstantPool successes, or fixing bugs ain't fun
Have base MediaHandler use AudioDecoderSimple for UNCOMPRESSED sound.
Add news about uncompressed sound support (was broken in 0.8.9)
CallFrame: Drop item from stack before invoking actions in called frame.
Add test for constant pool usage on CallFrame action (succeeds)
Add test for constant pool and clip event code interaction (gnash fails)
Make ConstantPool managed by VM
I guess you need ConstantPool class files too...
Don't loop 50 times in tests when a single run is enough
Add a trace-based test for _root.onLoad user-function execution order
Put RegisterClass tests in their own dir
Be more verbose during execution of onLoad events
Be verbose also about the second onLoad event...
Assing a name to the clip-with-events too
Fix abs_builddir
Put INITACTION tests in their own dir
Rename files to avoid mixing CamelCase with underscore_separated...
Review and cleanup InitActionTest2.c
Substitute "green" album cover with focused green-toned image (#33764)
Substitute "lynch" album cover with focused red-toned image (#33764)
Substitute "offspring" album cover with focused blue-toned image (#33764)
Build libmingutils.la in misc-ming.all before descending in subdirs
Ignore generated 'libtool' and a test in misc-ming.all
Add test for lastIndexOf with UTF8 characters (for bug #33718) -- gnash
fails
Fix lastIndexOf with UTF-8 (bug #33718) - expect success in testcase.
Add another test for UTF8 with indexOf, and simplify the fix for
lastIndexOf
Attempt to test #33754 (NewAdd with undefined) but gnash succeeds...
Automate test for bug #33754 (gnash fails)
Fix ActionNewAdd with short stack (bug #33754)
Forgot to add shortstack.trace..
Add test for frame0 actions execution order.
Stop mc3 timeline at end of test, reduce frame rate to 1fps
Move ChangeLog files in their own directory (under doc/)
Fix path to ChangeLog files
Queue onLoad events with a second liveChars scan in movie_root
A fix and a tolerance improvement for Matrix.as test
Fix libming warnings
Add note about Captivate output support
typo (#33841)
Use a random SHM key, hopefully fixing conflicts in chrooted builds
Use nanoseconds for the LC key, simpler, and HEX wasn't supported ...
Add missing <map>, for composition (build on lenny failed otherwise)
Update version 5 of MovieClip.as test
Update version 6 to 8 of MovieClip.as test
Fix build with GNASH_DEBUG_SOUNDS_MANAGEMENT defined
Attempt to fix redefinition warning on hurd
More results validation from wiki page
Drop spurious newline from per-frame debug line
Found another spurious newline
Improve AS error log about registerClass to include supposed export id
Remove unused code
Reduce SharedLib constructors to 1, destructors to 0
Sound.stop() invoked on an unattached Sound stops all event sounds.
Add cross-reference from NEWS item to savannah ticket
Revert to building the FB gui automatically by default
Re-enable distribution of cygnal and libdevice.
Revert "Re-enable distribution of cygnal and libdevice."
Typo: DefineFone2 -> DefineFont2. See bug #34387.
Unify message for Gui::hideMenu and Gui::showMenu, addressing issue #34428
Add Midori to the list of supported browsers (webkit based)
Update translation files [unchecked, to be reverted if found broken]
Fix parsing of lossless 16bit bitmaps
Fix message typo (#34651)
Update translation files
Few more translations in the italian po file
Provide a thumbnailer for SWF files and GNOME desktop config for it.
Distribute gnashthumbnailrc.in (#35075)
Set release date (too late...)
Sebastian Parborg (2):
Fixed made the gtk gui respect the Renderer option
Made gnash choose the first available render as default.
-----------------------------------------------------------------------
hooks/post-receive
--
Gnash
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Gnash-commit] [SCM] Gnash annotated tag, 0.8.10, updated. 0.8.10,
Gabriele Giacone <=