64156fc039
Expose glDrawBuffersEXT and allow GL_EXT_draw_buffers
...
The GL_EXT_draw_buffers extension defines the glDrawBuffersEXT method
which can be aliased to the glDrawBuffers method.
2017-05-22 07:51:45 -07:00
16bc831c16
Protect against recursive/re-entrant swapbuffers calls
...
* E.g. internally SwapBuffers might call wglSwapBuffers which we also
hook, and we'd process it twice redundantly. Only process the first
one we come across, whichever that is.
2017-05-22 11:00:33 +01:00
188481e881
Fix implementation of round_z in D3D debugging
2017-05-22 11:00:33 +01:00
007846ac67
Fix patch list count off by one error
2017-05-22 02:56:26 -07:00
1dcd39afd5
Add support for GL_ETC1_RGB8_OES compressed texture format
2017-05-22 02:55:54 -07:00
70514fc5ac
Store and load data of compressed textures on GLES
...
Since GLES does not support glGetCompressedTexImage function, the
compressed data cannot be extracted. For solving this, we store
them at glCompressedTexImage* calls, and use these data when
required (however it is not necessarily correct).
2017-05-22 02:55:54 -07:00
9ef362c28e
Fix isSRGB detection for EGL.
...
glBindFragDataLocation not allowed for gl_ outputs.
Fixes found when using Mesa.
2017-05-22 02:55:42 -07:00
8b059987ad
GL_EXT_texture_buffer and GL_OES_texture_buffer
2017-05-22 02:55:19 -07:00
b0c07fdd01
GL_OES_geometry_shader and GL_EXT_geometry_shader support
2017-05-22 02:55:19 -07:00
6a66492f83
GL_OES_tessellation_shader and GL_EXT_tessellation_shader support
2017-05-22 02:55:19 -07:00
5552e22431
GL_OES_sample_shading support
2017-05-22 02:55:19 -07:00
6566426195
GL_EXT_draw_buffers_indexed support
2017-05-22 02:55:19 -07:00
f93b0d6fb7
GL_OES_copy_image and GL_EXT_copy_image support
2017-05-22 02:55:19 -07:00
87ec279f1c
Update some forgotten Copyright years to 2017
2017-05-19 17:28:41 +01:00
1f5f34e16c
Bump version info to v0.35
2017-05-19 17:27:00 +01:00
ec4899ec72
Default current view in buffer viewer to VS In.
2017-05-19 16:24:43 +01:00
9b41534daa
Make the linux timeout even more generous still
...
* It's not too bad to wait a while since we have a progress bar that
will pop up if the operation takes a while.
2017-05-19 15:57:58 +01:00
57831f1926
If a callstack is valid but no resolver is found, return one empty entry
...
* This is a complete hack really but it signifies this status to the UI.
2017-05-19 15:56:49 +01:00
24f2c460f5
Pass GL_RED_INTEGER instead of GL_RED to glClearBufferData
...
* Mesa requires that we precisely specify it, and won't convert from
'floating point' GL_RED. Although I'm not sure what that means with
GL_RED and GL_UNSIGNED_INT...
2017-05-19 15:46:13 +01:00
79a61c5700
Disallow docked panels in a shader viewer from becoming float windows
2017-05-19 14:44:32 +01:00
a8a7e32cc0
Set reasonable bounds and steps for projection matrix guess controls
2017-05-19 14:41:42 +01:00
5f05c4309f
Fix a mistaken string format that merged two arguments accidentally.
2017-05-19 14:41:23 +01:00
d1db941275
Tweak a couple of lines to refer to renderdoc.org instead of github
2017-05-19 14:24:27 +01:00
abb07a5be1
Add missing mention of D3D12 support in a couple of places.
2017-05-19 13:52:16 +01:00
e36204da09
Bump the MSI version to 3.1, since the VS2015 redist requires it
...
* 3.1 is still available everywhere we care about - I think even 4.5
ships with win7.
2017-05-19 13:27:12 +01:00
04433f4c3a
Remove search for suffix on git commit hash in C# UI
...
* Instead we just add a comment next to a 'false' return, and it can be
replaced in a build script.
2017-05-19 12:51:15 +01:00
0c9806f30d
Restore linking against breakpad libraries in release builds.
2017-05-19 11:46:36 +01:00
4eda46fd96
Don't print to stdout/stderr in captured programs
...
* Otherwise it could interfere with normal functioning. Especially if we
run a script that then runs ls or dirname or pwd or something, and we
start printing logging messages! Oops.
2017-05-18 15:57:52 +01:00
5ddf450ce2
Only update the vertex data we're using in the PickVBBuf
2017-05-18 15:18:27 +01:00
5b74b66493
Don't construct a QFont globally (as a class static)
...
* It causes a crash when Qt is statically linked.
2017-05-18 15:05:03 +01:00
c5b79e9382
Crashhandler should only be enabled on *windows* official builds
2017-05-18 13:58:26 +01:00
7f5a56912b
Hook the nvapi wrappers around D3D11CreateDevice
...
* We used to hook nvapi.dll's call to d3d11.dll, but we don't hook that
dll anymore. So we'd miss this call and so miss the whole device.
2017-05-18 12:44:25 +01:00
6a0538dcd4
Fix mismatched serialising on remote server code, std::vector both sides
2017-05-18 11:18:45 +01:00
481dfe994b
Ensure each element in an array is vec4-aligned in cbuffer. Refs #618
2017-05-18 09:46:49 +01:00
3d20a6be1f
Fix GL pipeline state viewer brokenness. Refs #617
2017-05-17 19:50:49 +01:00
b94d13b838
Make sure linux log files don't get deleted while still logging to them
...
* We need to use file locks to ensure the last process knows it can
delete the log, as well as delete any inherited file descriptors when
forking to launch a new executable.
* To accomplish this, tracking the single logfile handle is now part of
the OS-specific code instead of common logging code. It makes little
difference either way as we don't support multiple log files.
2017-05-17 19:41:39 +01:00
6df6d152f0
Increase maximum wait time to find the target control port after launch
...
* Sometimes a process can take a while to start, this change waits up
to 1+2+4+8+16+32+64 = 127 ms which is still not that much.
* Also log a message if we still can't find it, so it's clear what the
problem was.
2017-05-17 17:45:58 +01:00
61dc8806bd
Add message about NVIDIA 378.x series driver crash issues
2017-05-17 17:10:37 +01:00
5dc10bcc3d
Make sure to call parent event handlers when we override to emit signals
2017-05-17 16:42:37 +01:00
1dc5f1b7ed
Fix vulkan drawindirect - not incrementing arg pointer each draw
2017-05-17 16:35:12 +01:00
f6cef9ea4a
Don't setFont(Consolas) - it's windows only. Fix monospace for scintilla
2017-05-17 15:55:42 +01:00
c867784647
Don't assign function ptrs in gl*GetProcAddress if we have one already
...
* In theory, gl*GetProcAddress can return different function pointers
at different times, but in practice this does not happen. Also we have
many more problems if there was an implementation that did this.
* In reality what can happen if we don't preserve existing function
pointers is that any call to gl*GetProcAddress will overwrite
required emulated functions for e.g. EXT_dsa if the implementation
doesn't support the extension but does return valid pointers for the
query.
2017-05-17 15:15:36 +01:00
f8d10850f8
Fix ambiguous QVariant casts from uint64_t on linux
2017-05-17 13:59:17 +01:00
9831a8eb24
Fix compile error comparing enums
2017-05-17 13:20:51 +01:00
3fff6fa86a
Ensure that in-memory buffer created serialisers have right version
...
* When flattening a log m_SerVer would be uninitialised which mean that
if it was less than 0x32 we'd start aligning buffers when we shouldn't
have.
2017-05-17 11:59:17 +01:00
b69a71c885
Don't try to set line number on disposed python shell
2017-05-17 11:57:10 +01:00
8a1304f6b4
Don't enable D3D12 debug layer in release builds on replay
2017-05-16 23:34:46 +01:00
9259055286
Handle 32-bit depth as floating point values
2017-05-16 18:01:30 +01:00
2ebcb9a337
Print a more helpful error when failing to convert component
2017-05-16 17:47:08 +01:00
c070c3e643
Ignore CoreMessaging.dll - it causes bizarre crashes in PeekMessage
2017-05-16 15:56:51 +01:00
Peter Gal