Dolphin Emulator Download

Revision: r7529: texcache-rewrite: Merge latest from trunk.
Revision: r7528 Consolidate wad installation to one function in the GUI now if you install the sysmenu from a wad in the gamelist the menubar will be updated with the new sysmenu version.
Revision: r7527 texcache-rewrite: Virtual EFB copy support in DX9 backend.
In DX9, ConvertEFBRectangle is returning bad results. Hopefully it's easy to
fix.
Revision: r7526 correctly decrypt the files when importing a wii save fixes issue 4462 .
Revision: r7525: OGL: Fix changing MSAA mode during emulation.
Fixes issue 4469.
Revision: r7524: texcache-rewrite: Merge latest from trunk.
Revision: r7523 texcache-rewrite: Unsuccessful attempt to fix DX9 EFB copies.
OGL works, even though it's using the same shaders.
Revision: r7522: typo/ put back a log that i removed.
Revision: r7521 Move the HLE hack of setting.txt file redirection to the BS2Emu we already copy the file to memory here, now we also copy the file to 1/2/data/
This should really fix issue 4287.
Thanks to hosigumayuugi for finding the exact cause of the issue
Revision: r7520 texcache-rewrite: Fix extra glitches introduced in r7515.
Revision: r7519 DX9: Pay attention to the pitch of a locked surface.
Revision: r7518: small fix for 32bit dsp HW reads.
Revision: r7517: texcache-rewrite: In DX9, pay attention to Pitch of locked surface.
Unfortunately, this doesn't fix any glitches. It should definitely be done on
trunk, though.
Revision: r7516 texcache-rewrite: Merge latest from trunk
Revision: r7515 texcache-rewrite: Clean up texture encoders in DX9 and OGL. Still glitchy.
I don't know why it's glitchy. It's like the source region is offset and
stretched. Does it work in trunk?
Revision: r7514 Update the country code in SysConf when wii system language is changed. remove annoying assert for invalid loaders on disc titles. Fixes issue 4287 .
Revision: r7513 texcache-rewrite: Merge latest from trunk
Revision: r7512 texcache-rewrite: Cleanup.
Revision: r7511: Trying to fix mixer problem when AIsrc is @48kHz, should fix games like All Star Baseball 2004/XGRA Extreme-G Racing Association/Conduit 2... Enabled 48kHz output for LLE
(before it was just 32kHz). I hope it works, if there are serious problems, it
will be reverted.
Revision: r7510: UpdateAudioDMA() fix/revert -> fixes Rayman 3/Smugglers Run with LLE, + small optimization for lleint/llejit32.
Revision: r7509 texcache-rewrite: DX9 support.
Please test DX9 and give feedback about the performance. For me, it's faster in
most cases, but slower in some situations involving paletted textures. Every
time the palette changes, it has to re-decode the texture to graphics memory.
The DX11 backend has a shader in place to handle palettes on the GPU, but the
other backends still de-palettize on the CPU.
EFB RAM copies are...still somewhat broken. You'll see if you try it.
Still to come: fix EFB RAM copies; virtual EFB copies for OGL and DX9.
Revision: r7508 texcache-rewrite: Merge latest from trunk
Revision: r7507 texcache-rewrite: Clean up texture encoders, start gathering bits of the texcache that should be common to all backends
Revision: r7506: Fix bug with Dolphin crash on stop when recording.
Revision: r7505: Rewrite to parts of the DTM code to remove the need for temporary files. Should
fix files occasionally getting corrupted. DTM export dialog now asks before
overwriting a file.
Saving a state that is not recording over one that was will now delete the old
recording. Fixes trying to record with the wrong savestate.
Fixed some compiler warnings.
Revision: r7504: fixes AFC type 5 decoding
(At least for Super Mario Sunshine, from whose ucode it was derived - hopefully
this won't break other games.)
Fixes Issue 4437 .
Revision: r7503: only create the data directory for a title, not data/nocopy. fixes issue 4453
fs::open does not need to create the data directory, es_diVerify does that.
change applied to CBoot::Boot_WiiWAD as well (es_diVerify isn't used for wads).
Revision: r7502 fixes issue 4388.
Revision: r7501: Fix OS X build (hopefully).
Fix a small UI inconsistency.
Fix a warning about shadowed variables.
Revision: r7500 texcache-rewrite: Merge latest from trunk
Revision: r7499 texcache-rewrite: Simplify PSTextureEncoder. I would do this on trunk, but there's a conflict I wasn't expecting.
Revision: r7498 Fix "Disable Lighting" in D3D9 and D3D11.
Fixes issue 3185 .
Revision: r7497: Use the recommended video settings as the default configuration:
- Fast Mipmaps enabled
- Safe Texture Cache enabled at "fast"
- showing shader errors by default
- using Native as internal resolution
- using EFB to Texture by default
- ignoring EFB format changes by default (turns out r7436 made this in Super
Mario Sunshine unnecessary anyway )
Any game which depends on other settings has that specified in the game ini, so
I think this should be fine.
Revision: r7496 Fix EFB format change emulation in D3D11 if MSAA is enabled. Fixes issue 4346.
Revision: r7495 Add a partial fix for 4xSSAA in D3D9 (issue 4090).
Revision: r7494 texcache-rewrite: Split depalettizing shaders into separate file and simplify
Revision: r7493 texcache-rewrite: Fix video config dialog.
Revision: r7492 texcache-rewrite: Merge latest from trunk. I'm going to have to fix the video
config dialog again.
Revision: r7491 texcache-rewrite: Fix an oversight. Mistakes like this mean it's time to clean up the code.
Revision: r7490Fix IR pointer of emulated Wiimote if sensor bar position is set to bottom.
Revision: r7489 - Do not parse numerous video options from the game inis, like:
Show FPS, Input Display, Statistics, Projection Statistics, EFB Copy Regions,
Dump Textures, Dump EFB target, Dump Frames, Free Look, Use FFV1, Show Shader
Errors, Texture Format Overlay, WireFrame, Hotkeys and Adapter
- Also removed various game ini config values which were still parsed from the
Video section.
Revision: r7488 Video configuration dialog stuff:
- Introduced a description text field (!!)
- Fixed ALL setting descriptions
- Various other string changes
- Removed the projection statistics setting
Revision: r7487 Quick fix for Dolphin crashing when using 3D and clicking anywhere on screen.
Basically just re-adding something that I removed a while ago... it wasn't
needed anymore, apparently it is again.
Note: Newest nVidia drivers break Dolphins 3D (270.xx), not sure why yet, gonna
ask around. It really confused me though when trying to fix this issue, spent at
least an hour trying to figure out what the hell was going on. I eventually
realised a build I knew worked in 3D was having issues...


Revision: r7486: New video configuration dialog design:
- added a "Hacks" tab for everything which speeds up emulation
- regrouped stuff in a more sensible way
- Renamed some options to be less confusing.
- by disabling ALL checkboxes, you'll get the most accurate / recommended
configuration now
- Merged the ATC options to a single slider with 4 values (with the rightmost
being "ATC off")
Revision: r7485: Add Catalan language thanks to (and at the urgent request of) Jordi Coma. Also fix a few translation strings in the code.
Reminder: Never mark the empty string for translation.
Revision: r7484: Remove configuration profile support.
I.e. revert most of the video configuration dialog changes since r7483.
Revision: r7483: dsphle - dkjb wii (fixes issue 3415).
Revision: r7482: this should fix ssbb/u (2nd try).
Revision: r7481 WriteARAM() revert, this should fix ssbb/u
Revision: r7480 texcache-rewrite: Merge latest from trunk
Revision: r7479 texcache-rewrite: Use buffer objects to hold palettes. Some devices can't handle 1D textures at 16384 texels in size.
Revision: r7478: Core/Common: Allow LEA(32,...) with M() OpArgs exceeding the 2G range. Fixes JitIL in debug builds.
Revision: r7477: minor correction to the tools menu, no longer uses uninitialized data when system menu is not installed.
Revision: r7476: small WriteARAM() fix i forgot before.
Revision: r7475: - Time extended in which the tooltip remains visible. (Windows PTFs)
- Better fix/location (wxWidgets source) for the MenuPopup fix. (Windows PTFs).
Revision: r7474 DSPLLE - "Super mario all star 25th anniversary" audio fix
Revision: r7473: - 'Sending' some changes of my previous commit to Windows build only.
Revision: r7472 - fixed a bug on GameList's MenuPopups (Windows Seven/Vista platforms, issue 4403)
- fixed a problem/crash which occurs with some gcc versions (in general on Linux
systems, issue 4400 )
- other minor GUI changes
Revision: r7471 llejit - asrnr fix and reverted Do_ARAM_DMA() change from my previous commit
Revision: r7470: Fix a few strings for translation. Fix a minor icon issue in the game properties dialog. Use a little hack to deal with spaces in file names. Need to find a better way to deal with this.
Revision: r7469: small fix for previous commit (prevents startup crash, if GC game was started after WII game).
Revision: r7468: DSPLLE - Donkey Kong Jungle Beat (WII) / Pikmin 2 (WII) audio fix
Revision: r7467 texcache-rewrite: Merge latest from trunk.
I don't know why IR_X86.cpp and IOdarwin.mm keep appearing in these commits. I'm
not touching them.
Revision: r7466 texcache-rewrite: Enable scaled depth EFB copies. Fixes depth-of-field effect in Wind Waker.
Revision: r7465: Core/PowerPC/JitCommon: Put the quantized paired/single load/store tables into
the code memory.
This makes sure they are below 2G on 64bit systems,
allowing us to continue using the simpler 32bit accesses.
Fixes issue 4357 and issue 4397 .
Revision: r7464 texcache-rewrite: Simplify virtual EFB copy encoders.
Revision: r7463 texcache-rewrite: Fix EFB-to-RAM copies in OpenGL.
Revision: r7462 texcache-rewrite: Merge latest from trunk.
Revision: r7461: DX11: I think I've figured out where the game enables texture offsets for lines and points.
Revision: r7460 DX11: Resolve multisampled EFB before encoding.
Revision: r7459: llejit - fix for flag generation code + dspspy jit workaround.
Revision: r7458: llejit - shift instructions should be correct now.
Revision: r7457: llejit/addaxl - Fix issue 4257 (SSX On Tour crashing with llejit/64bit).
Revision: r7456: llejit/clrp - gbakey working now.
Revision: r7455 texcache-rewrite: OpenGL support about 3/5 of the way done. Virtual EFB copies are not implemented yet. RAM EFB copies are "implemented," but they don't work and I don't know why.
Revision: r7454: texcache-rewrite: Skeletal support for DX9 and OpenGL. They no longer crash on start-up, but they are still utterly broken.
Revision: r7453 texcache-rewrite: Merge latest from trunk
Revision: r7452 texcache-rewrite: EFB Copy config settings work now.
Revision: r7451 GeckoCode float operations: Fixed Issue 4387. (hopefully)
Revision: r7450: Core/DSP/Jit: Ensure aligned stack when pushing registers for ABI calls This fixes issue 4378 : DSP LLE Jit not working on OSX
Revision: r7449: Fix sysconf generation. Fixes issue 4348.
Revision: r7448 texcache-rewrite: Merge latest from trunk.
Revision: r7447 texcache-rewrite: Much better speed when anti-aliasing is on. It seems that binding a multisampled texture to a regular texture slot will kill performance. We need to Resolve the EFB first. This ought to be done on trunk, too...
Revision: r7446 Fix an error introduced in r7083. In the pixel shader manager the farZ and zRange of the z bias were flipped. Switched to using the viewport struct rather than a raw float array to hopefully avoid such confusion in the future. Fixes issue 4060 .
Revision: r7445 texcache-rewrite: Remove obsolete options from video config GUI. NOTE: EFB copy settings are currently ignored.
Revision: r7444: Branch off the texture cache rewrites. I'm putting this on a branch because
a) It's very incomplete.
b) The trunk is feature-frozen right now, or at least that's my impression.

Don't get too excited. This code works with the DX11 backend ONLY. The other
backends are broken and will crash on start-up. Eventually the other backends
should get updated too.

I'm honestly not sure if this work is on the "right track." Whatever gets merged
to trunk will be much different than this. Still, here are some of the
improvements this branch has compared to trunk:

- Enormous performance improvement for all games...on MY computer. I'd like some
feedback from other users.
- The "Accurate Texture Cache" option is obsolete. It now behaves like "Safe"
mode, but much faster.
- Render-to-texture effects (aka EFB copies) are virtualized, meaning a high-
resolution "look-aside" texture can be used in place of the actual RAM copy if
the game made no changes to the texture. This is the same as the "EFB Scaled
Copy" feature we had before, BUT...
- The "look-aside" texture can be interpreted with a palette and retain its high
resolution! Thus, things like the Twilight Princess map and Metroid Prime's
thermal and X-ray visors now work in high resolution.
Revision: r7443 Create new branch texcache-rewrite for major changes I have to the texture cache system.
Revision: r7442: Core/DSP/Jit: Fix accumulator limiting in several extended ops Corrects some audio in Pikmin when using DSPLLE Jit.
Revision: r7441 Edited wiki page Windows_Build through web user interface. MS fixed the Win7SDK + VS SP1 issue.
Revision: r7440: Fix some strings for translation in the video config dialog.
Revision: r7439: - Mac build fix.
Revision: r7438: Linux build fix.
Also move the primary translation pot file to the external repository to
facilitate updating translations when strings in the code are changed. Now the
only time the primary dolphin repository needs to be changed for translations is
when a new language is added.
Revision: r7437 * One Window 3-State changes (VideoSettings GUI):
- Added final support to string type (planned implementation reaches 100%)
- Added support to post processing shaders (OpenGL backend)
- Code Maintenance:
- Reorganized the class hierarchy
- General polishing... deleted some redundant code
Revision: r7436: VideoCommon: Update color writing mask if pixel format changes.
Possibly fixes issue 4340.
Revision: r7435: Revert r7086.
Preferring the slow code until it gets optimized properly...
Revision: r7434: Fix SDL 1.3 support.
Revision: r7433: Use mutexes in LogManager to make it threadsafe, as suggested by beistin's patch. Change some LogManager function names to be consistent with Dolphin's naming conventions.
Revision: r7432 roll back r7431.
Revision: r7431 Add a "force no texture filter" option, useful for some 2D games (like Mega Man 10) and people who like their pixelated graphics.
Revision: r7430: videocommon depends on core for Fifo Recorder.
Revision: r7429 Fix the segmentation faults on emulation start on linux. The HW::Init() call needs to be before the video backend initialization. This is the way it was before soren's change in revision 7188. Now the emulator initialization sequence is pretty much back to the way it was before.
Revision: r7428: Linux build fix.
Revision: r7427 Create sys conf if it doesn't exists instead of asking questions.
Revision: r7426: Fix issue 4267. (but not really).
Revision: r7425 Fixed issue 4141.(key press beeping) Initializing the video backend in the GUI thread seemed to be the cause for whatever reason. Hopefully other platforms don't hate this change.
Revision: r7424 Set up some more emulated wiimote default keys. Enabled the nunchuk extension by default, with some key mappings. Fixes Issue 4266.(mostly).
Revision: r7423: DX11: Fix issue 4345.
Revision: r7422 - Eliminated a memcpy in DX11's vertex shader disk cache loading. Maybe games will boot faster now, who knows?
Revision: r7421: I think clean up DX11 code a bit. Used "smart" pointers to avoid need to manually AddRef/Release in many places. Eliminated the D3DBlob class (replaced with just D3D10CreateBlob). Eliminated some Init/Shutdown functions (moved stuff into ctors/dtors). This should not affect behavior at all, only code maintainability.
Revision: r7420 - Linux build fix
Revision: r7419: - Removed RTTI support
Revision: r7418 DX11: Correct line width and point size by taking viewport rect into account.
Twilight Princess map looks almost perfect except for the drop shadows around
the borders: the lines are too skinny, but the corner points seem to be correct.
Revision: r7417 * fix crash for x64 and debug build
Revision: r7416 VideoConfig GUI:
+ Added "Reset To Default" functionality for per-game profiles.
- Solved a window centering problem that occured during switching between
Default Profile and a Per-game Profile.
+ Added final & complete support to reflect changes on the fly and real-time in
all cases (even when you edit Default profile).
Revision: r7415: Fix scons build.
Revision: r7414: Add graphics FIFO recorder and player for debugging the graphics system.
Revision: r7413: Core/DSP: Split extended opcode 0xc0/mask 0xc0 to account for 0xc3/mask 0xc3variant.
In assembly, these are 'ld $ax0.d,$ax1.r,@$arS with their n,m and nm variants,
which have been special cased for S==3. The regular 'ld can be decomposed
into lrri $ax0.d,@$arS and lrri $ax1.r,@$ar3, while the S==3 case decomposes
to lrri $axR.h,@$arD and lrri $axR.l,@$ar3. The latter variant will be
disassembled to 'ldax $axR,@$arD after this change. The assembler recognizes
both the new 'ldax variant and the old 'ld with @$ar3 but the disassembler
only outputs 'ldax. Besides the readability, this allows for more correct
register use analysis(when it's done).
Revision: r7412: When using the "Start Renderer in Fullscreen" option, really start in fullscreen. In other words this now switches to fullscreen before the renderer
is initiated instead of after. This is a partial fix for issue 4316.
Also, if the render window size changes while frame dumping, scale the resulting
video to prevent clipping on linux. This is a complete fix for issue 4316 on
linux. I don't know how to implement this on windows though.
Revision: r7411: Have DX9 plugin also dump frames on every VI as opposed to every new frame (follow-up to r7131 and issue #4064 ).
Revision: r7410: Unfortunately, the changed fullscreen order in the fix for 3D Vision was causing some problems. I'm sure there's another way to do it but for now, 3D Vision users can always compile new builds with the fix.
Revision: r7409 DSP:
* LLE dspinit hax (demo disks, zelda collectors edition, ocarina of time working (again) with lle)
* LLE interpreter "DSPonThread" working again
Revision: r7408: vs2010: make sure minimal rebuild is turned off
Revision: r7407 only paint GameListItems if any exist
Revision: r7406: Only create a GameListItem cache if the image is valid.
Revision: r7405: VideoConfig revision:
+ increased VideoConfig integrity
+ fixed a clamorous fault, before completely forgotten, about handling
parameters attached to RadioButtons controls
Revision: r7404: - Attempt to fix a little problem displaying games within Profile Selection in Video Properties
Revision: r7403 Eliminate ini file loading in the paint, sort, and tooltip events.
Revision: r7402 Linux build fix.
Revision: r7401 Fix a few issues with the size of the logger pane. Also some other general clean up issues.
Revision: r7400 * Added proper string conversion support to VideoSettings (now you should see list titles correctly)
Revision: r7399 dsplle - prevent crashing, if there is unknown opcode (interpreter, only important for tests and certain homebrew ucodes), more cleaning, removal of obsolete TODO-s
Revision: r7398: - Attempt to fix a little problem displaying games within Profile Selection in Video Properties
Revision: r7397: Sorry, apparently that last commit was pointless. Removing most of it. Best have
as little 3D Vision mess in the code as possible.
But hey, at least somewhere along the line someone fixed hotkeys for us 3D
Vision users.
Revision: r7396: More 3D Vision related stuff. This is the last one I promise.
Hotkeys now work when running 3D Vision. You've just got to select the render
window first (as in click on the fullscreen image), don't ask me why, it's
something to do with RendererHasFocus in Frame.cpp. Annoyingly, escape crashes
the emulator and the screenshot button doesn't do anything, though I guess any
other hotkey should be working. I wouldn't usually commit this but I remember
someone really wanting to be able to connect their Wiimote while in 3D, so I
figured I might aswell.
Revision: r7395: Linux build fix, and silence some compiler warnings.
Revision: r7394: fix a memleak and quiet some warnings
Revision: r7393: Fix gcc build.
Revision: r7392: Fixed some memory leaks. Only one was mine ;P
Revision: r7391: * GUI Video-Settings changes:
- Removed recent ghanges about separate per-game Video Settings
- Added "One Window 3-State Support" (very close to full implementation)
Revision: r7390: OS X also needs to have the display resolution changed before
setting wxFULLSCREEN_ALL.
Revision: r7389: Linux needs to have the wxWidgets window go to fullscreen after the display resolution is changed.
Revision: r7388: *sigh*
I knew it would be something tiny. 3D Vision is fixed now, I'll make this 3D
Vision specific if it causes any problems, but I've tested it and it seems fine.
For anyone interested, because fullscreen wasn't set for wx first, the menus and
stuff were being included in the size of the frame (which is being used for the
backbuffer size), so it was getting 1920x1062 or something weird like that (for
1080p). In the render function, it then checks to see if the frame resolution is
equal to any supported resolution and if it isn't, bumps you down to 640x480,
apparently an unsupported 3D Vision mode... though now I think about it, I'm not
sure why it wasn't messing with regular fullscreen mode.
This would have been so much easier with a working mouse...
I'm not even joking, my replacement mouse just arrived at the door...
Revision: r7387: Linux build fix.
Revision: r7386: Moved per-game graphics configuration to the game-list right click menu. It will be too difficult to make the "profiles" drop-down thing work with 3-state vs 2-state checkboxes. The per-game settings now have "undetermined" states, except for the radio buttons(I'll fix that later). It's super hacky right now. Video config (probably all config stuff) could be redone.
Revision: r7385: It now works in the debug builds, but not in release builds for some reason. I'm not used to big projects like this so it's probably something obvious that I'm missing. Exclusive full screen mode seems to need the fullscreen resolution for the backbuffer when the DX device is created, so I used adapters[adapter].resolutions[f].xres and yres to get that, but in release
builds, that gives me the windowed resolution (in debug it give me the
fullscreen res, like I believe it's supposed to), which it reallly doesn't like.
It's not a difficult fix probably, but I've been staring at my monitor in
confusion for too long now and need to get to bed. If anyone wants to test, you
can still enable the 3D Vision option, it just wouldn't be 3D (obviously).
Revision: r7384: Core/DSP/Jit: Fix instruction size retrieval in Branch emitters.
Reading the instruction from memory at the given address and using that
as index in the opTable leads to more correct results than using the
address as index into opTable.
Also assert when trying to execute a bad instruction instead of crashing
without warning.
Revision: r7383: Update build instructions to reflect r7380 and r7381.
Revision: r7382: Update build instructions to reflect r7380 and r7381.
Revision: r7381: Simplify the SCons build:
With wxWidgets 2.9.2 now in Externals, SCons was the only external dependency other than Xcode and building just SCons from Macports caused a fair mount of extra stuff to be built unnecessarily, so just include a copy of scons-local. Always build wxWidgets from Externals which allows us to get rid of the home-grown autoconf machinery in SconsTests. The main use case for args.cache was the wxconfig option,
which is no longer relevant after the above changes and
caching options has a lot of potential for confusion,
especially since other build systems don't tend to do this,
so we no longer do it now.
Revision: r7380: Import r67258 of the wxWidgets trunk, which I expect will before
long become wxWidgets 2.9.2, which in turn is expected to be the
last 2.9 release before the 3.0 stable release.
Since the full wxWidgets distribution is rather large, I have
imported only the parts that we use, on a subdirectory basis:
art
include/wx/*.*
include/wx/aui
include/wx/cocoa
include/wx/generic
include/wx/gtk
include/wx/meta
include/wx/msw
include/wx/osx
include/wx/persist
include/wx/private
include/wx/protocol
include/wx/unix
src/aui
src/common
src/generic
src/gtk
src/msw
src/osx
src/unix
Revision: r7379: Hide DX11 on WinXP (fixes Issue 3719 )
Revision: r7378: Removed the OpenMP option from the DolphinWX project. It was causing a slowdown when the EFB to texture option was activated. This means that the OpenMP texture decoder option will not be hidden for non-OpenMP builds.
Revision: r7377: Enabled OpenMP support in the Windows 32bit builds.
Revision: r7376: Fixed a memory leak that occurred while compiling the OpenCL kernels. Reset the frame counter after loading a save.
Revision: r7375: Added a toggle switch for the OpenMP texture decoder in the graphics settings.
Revision: r7374: Game-specific video settings:
* Moved the config validity check to reflect the emulator enforced settings at
the config dialog
* Removed some obsolete validity checks
* Corrected "EnablePixelLighting" typo in the ini files
Revision: r7373: Make sure to not do any XFB emulation at all if it's not enabled. Should fix issue 4294 .
Revision: r7372: Build fix. :3.
Revision: r7371: Missed a dialog before.
Revision: r7370: Core/DSP: Fix loop/loopi for Interpreter and JIT.
A repeat count of zero still executed the instruction once. Fixes Zelda Ucode
audio crackling.
Fixes DSPLLE audio in: Luigi Mansion, Pikmin, Super Mario Galaxy, Zelda CE, Zelda Four Swords, Zelda WW, Zelda TP.
(Thanks LM for testing)
Revision: r7369: Change VERSION back to STATE_VERSION. Update dolphin-emu.pot file for recent changes to strings in code.
Revision: r7368: Linux build fix. Why would you swap with an empty temporary vector instead of clearing a vector? What were you thinking?
Revision: r7367: Changed the frame display to the number of polls. Changed frame counter to VI counter. Changed g_FrameCounter to u32. Patch by bzb95.
Revision: r7366: Eliminated memory leaks of the save state code and put it in a namespace. It is prettier than before, but it could be better (less global usage). Other minor stuff.
Revision: r7365: Finish making dialogs close when escape is pressed. Fixes issue 4265. At least I think I got them all.
Revision: r7364: Make sure that the video software backend doesn't get stuck in an infinite loopon shutdown.
Revision: r7363: Fix the framerate counter of the software graphic backend.
Revision: r7362: Hack Misc.cpp into building on OS X.
Revision: r7361: Fix the video software backend. (closes issue 4269) Some warning fixes and cleanup.
Revision: r7360: More work on making dialogs close when escape is pressed, and some general GUI code clean up.
Revision: r7389: Make "GetLastErrorMsg" thread safe.
Revision: r7358: Fix and issue with the escape button closing the hotkey dialog when trying to set a hotkey to escape.
Fix a few more dialogs to close with escape.
Fix an issue with the pulse audio sound backend when the dsp emulation method is changed, and the frequency setting was not respected.
Revision: r7357: DX11: Preserve w coordinates of lines and points. Fixes glitches when a line is partially behind the camera.
Revision: r7356: Fix the nogui build.
Revision: r7355: Reinstate AfterInit(), which is still needed for MacOpenFile() to work properly and possibly also wxProgressDialog.
Revision: r7354: Build fix for linux, and fix a deadlock when the core is initiated.
Revision: r7353: A bit of cleanup to Core Init/Stop, Frame, and Main. Cleanup XAudio2 to attempt to fix the crash on stop(didn't help :p). For some reason CFrame::DoStop is called twice.(might be the issue)
Revision: r7352: Some more work on making dialogs close when escape is pressed. Also make sure dialog modality is ended properly so that they can be closed on OSX.
Revision: r7351: The escape key should now work to close most of the dialogs in dolphin. Let me know of any that I missed. I am sure there are some. I also notice that anytime a wxNotebook is used a page should be set to get proper focus on the dialog.
Revision: r7350: Make the hotkey dialog close when the escape key is pressed. I commit this as an example of how all dialogs should be set up. The buttons used to close the dialog should use wxID_OK or wxID_CANCEL. Alternatively you can use SetEscapeId or SetAffirmativeId to set the escape id to the id you use.
Revision: r7349: Fix auto window size options when not rendering to main.
Switch gtk keystate_lock back to a recursive mutex. Now that the mutex attributes are being initialized properly this doesn't seem to cause the segmentation faults anymore.
Revision: r7348: fix ppc "jump to address" textbox and some symbol menu items.
Revision: r7347: Fix DInput rumble.
Revision: r7346: DX11: Don't apply culling to lines and points. Fixes Metroid Prime beam-charging effects.
Revision: r7345: Fix DInput and SDL.
Revision: r7344: pthread_mutexattr_t's must be initialized before the type can be set. Fixes issue 4243 .
Revision: r7343: Move translation po files to an external svn repository. The location is [Devi essere iscritto e connesso per vedere questo link] I have given some of the primary dolphin committers write access to that repository. If you are a
dolphin-emu committer, and I didn't give you access, and you want it, just let
me know.
Now po files can be updated as often as we want without bumping up the svn
revision numbers of dolphin itself. I left the dolphin-emu.pot and primary
build files in the primary dolphin-emu svn. The dolphin-emu.pot file will still
need to be updated on occasion to reflect changes to strings in the program.
The gettextize script will still work as before for this.
Also added Arabic translations to the build thanks to mansoor.
Revision: r7342: It seems like CoreTiming's "external event" mutex is being used recursively. This fixes dsp lle jit on thread for me. YMMV
Revision: r7341: Yikes, I have no idea how this happened.
Revision: r7340: DX11: Support thick lines and large points with geometry shaders. Lines still aren't drawn with correct thickness, see Twilight Princess map. But the map corners are drawn with correct shadow "blobs" for the first time!
Revision: r7339: Clean up "ControllerInterface" a bit.
Revision: r7338: AudioCommon: Improve pad silence when ppc does not keep up with realtime.
Uses the last sample from the ppc buffer to fill the samples the ppc didn't deliver data for, avoids clicking on underruns.
Revision: r7337: Attempt to fix issue 4234.
Revision: r7336: Core/DSP/Jit: Implement Saturation/Limiting when reading ACM* in 40 bit mode.
Revision: r7335: The data from strdup must be deallocated with free, not delete. I got rid of that ugly strdup and replaced the A though Z cdrom drive letter searching with Windows API stuffs.
Revision: r7334: Re-enable the std::mutex try_lock on linux and use it for the Host_GetKeyState keystate_lock mutex. This reduces the number of application crashes on linux during emulation.
Revision: r7333: Fix things so that those who wish to compile with -std=c++0x can. To active
this with the cmake build add CXXFLAGS="-std=c++0x" before cmake on the command line, or export that variable. This enables the experimental features like
std::thread, std::mutex, etc., that are provided by g++ instead of using the
implementation in dolphin
Revision: r7332: Probably should use the IOFile built in mechanism to check if the file is open.
Revision: r7331: Make sure the frame dumping file is closed when emulation stops as it should be
for the non avi dump build.
Revision: r7330: Make the non-HAVE_LIBAV case compile.
Revision: r7329:Hack up hack number 7328, just to get things to compile.
Revision: r7328: Wrapped fopen/close/read/write functions inside a simple "IOFile" class. Reading, writing, and error checking became simpler in most cases. It should be
near impossible to forget to close a file now that the destructor takes care of
it. (I hope this fixes Issue 3635) I have tested the functionality of most
things, but it is possible I broke something. :p
Revision: r7327: windows build wiki: wee, vs2010 express has problems with windows sdk...
Revision: r7326: Remove the need for the less than standard strnlen()/strndup().
Revision: r7325: Fix OS X Lion build:
Lion no longer includes wxWidgets 2.8, so be prepared to fall back to
Externals/wxWidgets3 even when no wx-config command is found at all.
There are a few new and deprecated symbols in the 10.7 SDK that
conflict with wxWidgets/Dolphin, so just use the 10.6 SDK for now.
Revision: r7324: windows build wiki: Added link to VS2010 SP1
Revision: r7323: add an option to not eject/reinsert memcards on save/load state
when off certain games like ZTP ZWW maybe other cannot save to a different
memory card after loading from a savestate, but for TAS the game will no longer
think that the card was ejected to use add ReloadMemcardOnState = false to dolphin.ini
Revision: r7322: Xcode 4 only includes the 10.6 SDK. No matter - version-min=10.5 is sufficient for binary compatibility.
Revision: r7321: Fix OS X nowx.
Revision: r7320: Fix the default setting for the fullscreen resolution on linux.
Revision: r7319: DX11: Add real XFB mode support. May not work perfectly yet. I'm working on a set of fairly big VI-related changes. When I'm done, it should improve animation smoothness for all the backends, especially when virtual or real XFB mode is enabled.
Revision: r7318: Fix a couple apparent typos. a) Don't pass ASCII string to Unicode function. b)
If you type "(128 / 1000)" in C++, it will be treated as "0" because the integer
division gets truncated. Fixed.
Revision: r7317: Fix the no gui build for the recent std event changes.
Revision: r7316: Shut up recent versions of GCC about the OpenMP pragmas.
Revision: r7315: Check if we are just displaying the empty game list before assuming that the ratings column is present. Fixes issue 4183.
Revision: r7314: - Little fix on CPUDetect.cpp about the correct detection related to new AVX extensions (bug/fix reported&suggested by debian.micove).
Revision: r7313: Core/DSP/Jit: problem: TryAcquireSRWLockExclusive requires Windows 7 :/. solution kinda: use std::recursive_mutex (CriticalSection on windows) where try_lock is used.
Dolphin should work on Vista x64 again.
Revision: r7312: Core/DSP/Jit: Fix register cache state tracking in bloop and loop emitters Fixes dsp lle jit crashes with dkcr.
Revision: r7311: Somewhat experimental. If the SYSCONF file is not found or is invalid then generate one. For now ask the user if this should be done. The generated SYSCONF file works for me with the games I have tested, but this needs further testing. Unfortunately, most users will never use this as they will have the SYSCONF file anyway. Eventually we could perhaps remove the distributed SYSCONF, and generate it the first time dolphin is run.
Revision: r7310: Use clang for compiling the few .c files, including minilzo-2.04
which appears to trigger a bug in the version of llvm-gcc included with some releases of Xcode prior to 3.2.5.
Revision: r7309: * Italian translation update
+ Overall and general inspection before...
+ ...removal all "string-leashes" (now you should see 100%(?) of translated strings)
Revision: r7308: Fix Host_GetKeyState in a more effective way.
Revision: r7307: Fix preprocessor prob.(recursive mutexes were being used where not necessary)
Hopefully fix Common::Event to handle spurious wakeups since it uses condition
variables now.
Revision: r7306: Fix GCPad rumble. Fix issue 4212.
Revision: r7305: Fix Common::Barrier and fix the deadlock by making Common::Event check to see if an event has already been set or not.
Revision: r7304: Fix the Host_GetKeyState deadlock caused by Billiard's latest changes.
Revision: r7303: DX11: Eliminate most int-to-float conversions and use float math in more places.
GPUs are more efficient with float operations. This gives a slight speed-up.
Revision: r7302: * Italian translation update
+ minor GUI adjustements, CPUID Detect cleanup... video profile selection is disabled if a game is running.
Revision: r7301: Make OpenMP optional on linux. For now it is on by default, but you can disable it by adding -DOPENMP=OFF to the cmake command line.
Revision: r7300: more fixes for opengl sing openmp, i hope is the last
Revision: r7299: more fixes for my openmp commit.
Revision: r7298: Fix OS X build.
Revision: r7297: some fix of Chinese translation
Revision: r7296: Build fix of linux build fix.
Revision: r7295: Missed a critical section.
Revision: r7294: Replaced Common::CriticalSection with a std::mutex implementation. 64bit Windows builds now use SRWLocks and ConditionVariables(requires Vista/7, x64 builds will no longer work on Windows XP x64). Tell me if you hate that. Removed Common::EventEx. Common::Event now uses a std::condition_variable impl.(using ConditionVariables on Windows x64, Events on x86, or posix condition variables elsewhere). I experience slight speed improvements with these changes.
Revision: r7293: Move the cdb.vff file creation out of the wx code.
Revision: r7292: deleting a channel via the wii sysmenu works now Implement ES_DELETETICKET & ES_DELETETITLECONTENT
Revision: r7291: Get rid of the annoying (and invalid issue creating) 3+ minutes to create cdb.vff now creation is handled by the gui and takes < 1 second
Revision: r7290: Replace install wii menu option tools to install wad Add the option to clear a title from the NANDContentLoader and attempt to reload it This allows the using the system menu immediately after installing rather than requiring a relaunch
Revision: r7289: some fixes for my last commit.
Revision: r7288: Linux build fix.
Revision: r7287: Minor cleanup.
Revision: r7286: reconnect all wiimotes on eslaunch
same hack as r7090 update ES_GETSTOREDTMDSIZE, ES_GETSTOREDTMD to work when called with only 1 inbuffer and no outbuffer
Revision: r7285: Fix a typo in BPFunctions causing a PanicAlert in SW:RS2. Really minor bugfix in DX9.
Revision: r7284: First Revert my changes to VertexLoader.cpp, i don't own the games that get error so i revert the changes until i can test it myself.
Second:
A experiment. implemented parallelization in texture decoding using openmp. is
most a experiment to test the performance in different os/plataforms. in my
system (windows x64 amd 1055t) give a speedup in large textures, but i tested in
in intel dual core and gives a slowdown. o i limited the use for large textures
and cpus with more than 3 cores.
please test an let me know if it improves or degrades the speed.
please for linux and osx user. to enable this you will have to enable your
compiler support for openmp to test this code.
Revision: r7283: Merge the log window CreateGUIControls and LoadSettings methods. This allows the settings from the ini file to be applied when the controls are created rather than setting a default, and then changing the settings later. In
particular word wrap is applied when the text control is created. This works
around the crash at application start that users are reporting in issue 4196.
Also change the for loops in SysConf to use iterators to placate godisgovernment
and billiard.
Revision: r7282: Move SConscript files out from the Src subdirectories to be consistent with the other build systems.
Revision: r7281: Fixed channel loading from the Wii System Menu. The DOL is not being loaded properly by the system probably due to caching. Wiimotes will need to be manually reconnected after the channel loads.
Also fixed a small bug in the PowerPC ResetRegisters function (clearing CRFast).
Revision: r7280: Prepare for the release of Xcode 4, which has working LTO with static libraries.
Revision: r7279: Remove a pointless cast. Name the DSP thread.
Revision: r7278: When reloading the sysconf file make sure that the m_Entries vector is cleared.
Revision: r7277:Reload sysconf file on stop in order to preserve changes committed during emulation
Revision: r7276: Add Serbian translations thanks to nikolassj. They are not very complete though.
Update most of the other translations thanks to the dolphin translation team.
Revision: r7275: Change YieldCPU to sleep(0) like on windows. Time trials show this to be the most effective. Also, when using usleep(1) and dsp on thread the emulator is painfully slow. With the old _mm_pause() the emulator hangs on shutdown. With sleep(0) the emulator is faster, and doesn't hang on shutdown.
Also fix a possible undefined operation (according to gcc).
Revision: r7274: Put the video prepare code back, and use the existence of the g_renderer to determine if video prepare has been run or not.
Revision: r7273: add a slider to control the wiimote speaker volume which is stored in sysconf.
Revision: r7272: Fix the wiimote speaker problem in the core and remove the HLE patch for it. Real wiimote tested on windows, untested but should work on linux, dunno about osx.
Revision: r7271: If the DSP LLE can not find the needed ROM files exit the emulator without crashing the application. Also reimplement soren's revision 7195 in a way that works on linux and windows. (This makes it easier to clean up the video backend
if the DSP emulator fails to initialize.)
Revision: r7270: DSP: implemented saturation in dsplle-int.
Revision: r7269: DX11: Simpler depth-fetch. All DX10+ cards should have enough float precision to make this safe.
Revision: r7268: Pass std::string's by reference when possible.
Revision: r7267: DX11: Implement the rest of the EFB encode formats, but beware: I don't have any games to test them.
Revision: r7266: More conversion from char * to std::string.
Revision: r7265: Convert GetUserPath to return a std::string instead of a const char *. This simplifies its usage in most cases.
Revision: r7264: Integer constants larger than int need a type suffix.
Revision: r7263: Make the break point window toolbar a real toolbar, and general debugger gui
cleanup.
Revision: r7262: DX11: Improve efb-encode shader some more.
According to AMD's GPU ShaderAnalyzer, most combinations of shaders have about 1
.5x-2x higher peak per-clock throughput after this commit. For those concerned
about performance, I do intend to make this at least as fast as the other
backends. This is one more step toward that goal.
Revision: r7261: Remove numerous video options from the ISO properties.
a little simplification for the dlist id and a little correction to
YieldProcessor in windows.
in windows sleep(0) behaves more like what is defined in the yield instruction
so use it instead. In my amd 6 core system brings a nice 8% speedup so please
test I'm interested in knowing the behavior in different systems.
Revision: r7260: Core/DSP: Access all registers except ST* through the regcache
No speed difference, probably because our basic blocks are
too small to contain more than one access to any register
or used too seldom for multiple accesses to make a difference.
Revision: r7259: Remove numerous video options from the ISO properties.
Use the new configuration profile system instead.
Thanks to kostamarino for migrating the game inis to the new system .
Revision: r7258: Improve DX11 efb-encoder shader code generation a little.
Revision: r7257: Fix a memory leak in the breakpoint window.
Parent message alerts by the active window. This way if another window is open
it isn't pushed behind the main window. We probably should parent the message
alerts by the calling window instead, but this may be good enough.
Make sure there is only one instance of some modeless windows (Cheats Manager
and Net Play).
Revision: r7256: Fix a brain-fart....
Revision: r7255: PS3 controllers have some input elements with very large value fields
that cause IOHIDValueGetIntegerValue() to smash the stack when trying
to convert them.
In practice, all relevant axes seem to also be available as either
8 or 16-bit values, so just ignore anything that doesn't look like
that (or a button).
Revision: r7254: add new files from r7253 to vs2010 projects.
Revision: r7253: Add EFB encode-to-RAM support in DX11 backend. It could probably be simplified a
lot, and not all the possible formats are implemented. I tried to use the
dynamic-linking feature of shader model 5, but Microsoft's HLSL compiler is
broken. "Dynamic mode" is implemented, but disabled for now.
Revision: r7252: Core/Core/Src/DSP: Drop the indirect addressing hackery from DSP Jit
Was made possible by the move of Core/DSPCommon to Core/Core.
This should not give a significant speed difference, but frees
another register for register allocator use on x64 and
generally simplifies the code.
Revision: r7251: Linux build fix.