Package: mupen64plus / 1.5+dfsg1-13
Metadata
Package | Version | Patches format |
---|---|---|
mupen64plus | 1.5+dfsg1-13 | 3.0 (quilt) |
Patch series
view the series filePatch | File delta | Description |
---|---|---|
fix desktop file.patch | (download) |
mupen64plus.desktop.in |
5 2 + 3 - 0 ! |
fix desktop file |
destdir.patch | (download) |
Makefile |
2 1 + 1 - 0 ! |
use destdir during installation Bug: http://code.google.com/p/mupen64plus/issues/detail?id=231 |
plugin searchpath.patch | (download) |
main/main.c |
3 1 + 2 - 0 ! |
use /usr/lib/mupen64plus as plugin searchpath |
dejavu font.patch | (download) |
opengl/osd.cpp |
9 3 + 6 - 0 ! |
use font from ttf-dejavu-core |
as needed.patch | (download) |
blight_input/Makefile |
2 1 + 1 - 0 ! |
allow compilation using -wl,--as-needed |
system libpng.patch | (download) |
rice_video/Makefile |
16 1 + 15 - 0 ! |
use libpng from system instead of buildin one |
system zlib.patch | (download) |
Makefile |
1 0 + 1 - 0 ! |
use zlib from system instead of buildin one |
system libbz2.patch | (download) |
Makefile |
11 2 + 9 - 0 ! |
use libbz2 from system instead of buildin one |
remove gln64.patch | (download) |
Makefile |
8 1 + 7 - 0 ! |
remove non-dfsg clean gln64 from build |
remove nondfsg icons.patch | (download) |
glide64/messagebox_gtk2.c |
26 0 + 26 - 0 ! |
remove non-dfsg clean icons from build Bug: http://code.google.com/p/mupen64plus/issues/detail?id=239 |
glide64 noasm.patch | (download) |
glide64/rdp.cpp |
10 8 + 2 - 0 ! |
don't include x86-asm in non-x86 plattforms |
rice texturepack crash.patch | (download) |
rice_video/TextureFilters.cpp |
2 1 + 1 - 0 ! |
fix segfault in rice_video using texturepack |
rice screenflickering.patch | (download) |
config/RiceVideoLinux.ini |
40 39 + 1 - 0 ! |
fix screenflickering in some games with rice_video |
static binutils libs.patch | (download) |
Makefile |
5 3 + 2 - 0 ! |
link binutils libraries with unstable abi as static library |
fix 7z subfolder.patch | (download) |
main/rom.c |
2 2 + 0 - 0 ! |
don't segfault when caching 7z archive with subfolders Bug: http://code.google.com/p/mupen64plus/issues/detail?id=256 |
xdg basedir.patch | (download) |
Makefile |
3 2 + 1 - 0 ! |
use xdg base directory compliant directories |
ftbfs dynarec.patch | (download) |
Makefile |
32 17 + 15 - 0 ! |
don't try to compile dynarec with not set dynarec architecture If we cannot find a valid architecture for dynarec we shouldn't try to build the needed object files which are architecture dependend. |
ftbfs kfreebsd.patch | (download) |
glide64/Ini.cpp |
2 1 + 1 - 0 ! |
don't stop build of kfreebsd due to unknown os Bug: http://code.google.com/p/mupen64plus/issues/detail?id=257 Bug-Debian: http://bugs.debian.org/544550 |
ftbfs debugger.patch | (download) |
debugger/memory.c |
22 22 + 0 - 0 ! |
define dynarec debugger functions for non-x86 systems The debugger checks for valid memory for dynarec. In case of architectures without dynarec support it should skip this check. The needed macro is not defined, but can be replaced with an empty check. Also functions to decode recompiled code have to be implemented as dummy functions. |
debian archs.patch | (download) |
pre.mk |
146 130 + 16 - 0 ! |
define new cpu architectures mupen64plus must exactly know for which architectures it is build and which word size and endianness it has. Most architectures supported by debian weren't listed here. That means that the build would probably fail or generate unusable binaries. . To better support unknown architectures NO_ASM must be enabled by default and known good architectures can then override this value. The user can still force to change that behavior by NO_ASM=0 or NO_ASM=1 on the commandline. . |
ftbfs glibc210.patch | (download) |
glide64/Combine.cpp |
3 2 + 1 - 0 ! |
fix ftbfs using g++-4.4 and libc 2.10 due to missing const keywords G++ 4.4 checks correctly if a pointer is converted from const to non-const without an explicit cast. As glibc 2.10 provides a more strict declaration of strstr and mupen64plus doesn't use const in that places, it will fail hard. |
version string.patch | (download) |
pre.mk |
4 2 + 2 - 0 ! |
remove misleading -development in version string |
default optimisations.patch | (download) |
pre.mk |
11 3 + 8 - 0 ! |
don't override optimisations set by build environment |
gtk open filter.patch | (download) |
main/gui_gtk/main_gtk.c |
2 1 + 1 - 0 ! |
show *.n64 and *.v64 in filtered gtk open dialog When somebody tries to open files with the extension n64 or v64 the filter says that they are supported but they will not be shown. If the user switches to to "all files" mode they will be shown. GTK has a special gtk_file_filter_add_pattern to add the actual pattern and gtk_file_filter_set_name to give them a name which will presented to the user. To show [vn]64 files the filter for z64 has to be altered and not only the name of the filter. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=264 Bug-Debian: http://bugs.debian.org/546046 |
noexecstack.patch | (download) |
Makefile |
7 0 + 7 - 0 ! |
don't enable executable stack by default A non executable stack is a security feature on modern platforms to protect against an attack made possble by a flaws that allows an attacker to fill stack memory with executable code. . As mupen64plus needs executable pages store its recompiled code the code blocks and wrapper blocks must be allocated page aligned to change the protection bits to executable of pages which belong to that memory block. |
fix_readpng.patch | (download) |
rice_video/liblinux/bmp.c |
6 3 + 3 - 0 ! |
fix crash when loading png due to reordering optimizer The optimizer can reorder instructions like the initialisation of variables if it looks safe to him. It isn't said that all variables are initialised right before the call of the first setjmp and thus we cannot rely on there value to determine if they are point to actual data. |
jttl_fix_romclosed.patch | (download) |
jttl_audio/main.c |
8 3 + 5 - 0 ! |
stop audio processing thread before freeing buffers SDL_PauseAudio is called to stop the audio, but that doesn't stop the audio processing right away. This means if it still is in SDL_MixAudio to change the volume and we start right away to free all buffers, it will segfault. We can call SDL_CloseAudio before freeing all buffers to be sure that my_audio_callback is stopped and doesn't process any data from the buffers anymore when we clean everything up in RomClosed. |
rice_nodebug.patch | (download) |
rice_video/Makefile |
6 0 + 6 - 0 ! |
don't enable excessive debug functionality in rice_video rice_video has extensive capabilities to debug itself. This feature is also enabled when enabling mupen64plus debugger, but isn't used by it. As it creates a massive slowdown after a while this feature is better disabled in Debian. |
interpreter_x86_fldcw.patch | (download) |
r4300/macros.h |
2 1 + 1 - 0 ! |
correctly set floating point control word In interpreter mode small assembler snippeds are used to set the floating point control word to emulate the MIPS coprocessor behaviour. The makefile variable isn't propagated to the c preprocessor and thus we need to check otherwise if target architecture is x86 or compatible. |
correct_fpr32_mapping.patch | (download) |
main/savestates.c |
17 16 + 1 - 0 ! |
correct mapping of floating point registers in mips i mode |
load_aidacrate.patch | (download) |
main/savestates.c |
1 1 + 0 - 0 ! |
set aidacrate during savestate load Without allowing the audio plugin to set the aiDacrate the program will start to behave strangely because the audio plugin will work with a wrong frequency. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=169 |
load_vistatus.patch | (download) |
main/savestates.c |
2 2 + 0 - 0 ! |
set video width and status during savestate load Not setting status and width for video plugins could result in corrupted output in some programs. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=169 |
fix_r0_override.patch | (download) |
r4300/pure_interp.c |
62 62 + 0 - 0 ! |
don't allow to override r0 register r0 on MIPS is hardwired to zero. It is allowed to write to this register, but the result of any operation with the target register r0 is discarded. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=278 |
resume_on_start.patch | (download) |
main/gui_gtk/main_gtk.c |
5 4 + 1 - 0 ! |
resume when pressing on start when in pause mode When the program is paused and user presses start it will try to restart the program without actual resetting it correctly. As result it will just not behave as expected. When it is paused it should continue instead. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=282 |
osd pause crash.patch | (download) |
main/main.c |
22 13 + 9 - 0 ! |
fix osd crash after pause-stop-start-pause of emulator The osd message handler for pause messages are managed outside the message queue of the osd and will only be updated during pauseContinueEmulation, but the message is bound to an OpenGL context and thus must be also be deleted when the emulator is stopped and the video plugin shuts down. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=283 |
system liblzma.patch | (download) |
Makefile |
9 3 + 6 - 0 ! |
use liblzma from system instead of buildin one Bug: http://code.google.com/p/mupen64plus/issues/detail?id=255 Bug-Debian: http://bugs.debian.org/543552 |
rice_fog.patch | (download) |
rice_video/RenderBase.cpp |
2 1 + 1 - 0 ! |
reduce flickering when enabling fog in rice When enabling fog in rice video plugin some polygons near the camera starts to flicker, turn to white or disappear completely. This happens due to negative fog coords send to OpenGL. This is invalid and should be clamped to 0.0 in that case. Bug: http://code.google.com/p/mupen64plus/issues/detail?id=30 |
ftbfs gvariant type conflicts.patch | (download) |
rice_video/typedefs.h |
10 6 + 4 - 0 ! |
fix ftbfs due to glib gvariant type definitions |
dont install unneeded.patch | (download) |
install.sh |
7 2 + 5 - 0 ! |
don't install files which aren't used by us |
rsp_ucode2_reset.patch | (download) |
rsp_hle/hle.h |
4 2 + 2 - 0 ! |
reset status of specific ucode2 hacks after starting again |
rsp_hle_bigendian.patch | (download) |
rsp_hle/hle.h |
2 2 + 0 - 0 ! |
fix wrong high level emulation of rsp on big endian systems |
rice crash vendorstring.patch | (download) |
rice_video/GraphicsContext.h |
2 1 + 1 - 0 ! |
don't crash on long opengl vendor string |
rice ati symbols.patch | (download) |
rice_video/Makefile |
1 1 + 0 - 0 ! |
allow to load rice on systems with missing opengl extensions |
link_gdk.patch | (download) |
pre.mk |
4 2 + 2 - 0 ! |
link directly against gdk We use functions of gdk-2.0 directly and thus we should link against it directly instead of indirectly through gtk+2.0 |
path_max.patch | (download) |
glide64/winlnxdefs.h |
4 4 + 0 - 0 ! |
add dummy path_max on systems without it Posix only defines that PATH_MAX may is defined through limits.h. GNU/Hurd has removed this definition without violating the standards. . Mupen64Plus still relies on it and we work around that problem by defining PATH_MAX as 4096 in all places it is needed. This is not 100% correct as their are ways to generate longer paths, but mupen64plus must be heavily refactored in some places to fix the actual problem. Bug-Debian: http://bugs.debian.org/588066 |
optional_signinfo.patch | (download) |
main/main.c |
2 1 + 1 - 0 ! |
only use siginfo subsystem if system offers it Bug-Debian: http://bugs.debian.org/588066 |
glide64_hurd.patch | (download) |
glide64/wrapper/glidesys.h |
2 1 + 1 - 0 ! |
define gnu/hurd as unix system for glide64 Bug-Debian: http://bugs.debian.org/588066 |