baldurk
ebe442d1d8
Update tips dialog and add a menu item for it.
v0.26
2015-09-25 21:09:12 +02:00
baldurk
b5f73cbb63
Don't call TextureBinding where enum might be GL_NONE. Refs #156
2015-09-24 18:40:11 +02:00
baldurk
f3abf31148
Catch FinishCommandList calls on immediate context, and throw an error
2015-09-24 18:30:14 +02:00
baldurk
7fdadc52f7
Fix register/variable highlighting sometimes getting busted
2015-09-23 13:30:12 +02:00
baldurk
51f8e949a1
Add back some code that was deleted by accident
2015-09-23 13:29:54 +02:00
baldurk
edccd1f9af
Remove topologies, index type, offsets, etc from GL drawcall strings
...
* This was mostly just useless noise that cluttered up the event list -
it's all implicitly displayed on the mesh output or pipeline state
2015-09-20 13:02:07 +02:00
baldurk
a5ba0e4597
Make injecting into a GL app safer (still does nothing if after GL init)
...
* If we haven't seen a context create anywhere, refuse to do any work -
this is pretty much definitely because of being injected into the
program after init time. Unlike D3D11 where all the API calls are via
interface pointers that were fetched and not wrapped, GL starts being
hooked immediately, which is bad!
2015-09-20 11:09:49 +02:00
baldurk
8678fa9e2d
Call SetLastError(0) to mute any errors after RenderDoc work
...
* Obviously not ideal, but better than passing spurious error codes
back to the application where presumably it wasn't getting them
before.
2015-09-18 23:31:34 +02:00
baldurk
4242c12226
Don't call into SwapBuffers if no window is associated with the DC
2015-09-18 23:26:11 +02:00
baldurk
1b8d291874
Avoid pointless multiple file passes, now that seeking isn't cheap
2015-09-18 22:37:44 +02:00
baldurk
b090092d73
Don't apply renderdoc hooks to renderdoc.dll. Things go... badly
2015-09-18 22:27:46 +02:00
baldurk
a9166ffb38
Fix logic for GetBuffer() UUID check
2015-09-18 22:11:15 +02:00
baldurk
84a89c1e32
Linux compile fix
2015-09-16 08:52:48 +02:00
baldurk
822cbd879d
Convert GL_CLAMP to GL_CLAMP_TO_EDGE
2015-09-16 08:37:17 +02:00
baldurk
1936a7a0b4
Get sized format for render buffers (not sure if that's even valid...)
2015-09-16 08:28:53 +02:00
baldurk
0c42d195c2
Fix copy-paste error with DRAW/READ framebuffer
2015-09-16 08:27:33 +02:00
baldurk
fd85a981ed
Get sized stencil formats
2015-09-16 08:27:23 +02:00
baldurk
e58c7aff13
Add extra bounds check just in case trace has no steps.
2015-09-16 01:18:14 +02:00
baldurk
f021aed31b
Handle unexpected UUIDs in IDXGISwapChain::GetBuffer a bit more nicely
...
* Had some crash reports hitting the assert indicating some other UUID
was passed. Most likely is ID3D10Texture2D or similar, which we
can't support. Any other just print an error and the UUID.
2015-09-16 01:12:33 +02:00
baldurk
c18b1547ac
Distribute renderdoc_app.h in packages
2015-09-15 23:11:07 +02:00
baldurk
60081dcfe7
Add fallback path to use d3dcompiler from renderdoc folder. Refs #153
...
* This should be a very rare case, but e.g. if the D3D11 runtime hasn't
been loaded ever, d3dcompiler won't be in the system dll path.
2015-09-15 23:11:02 +02:00
baldurk
822a235349
Make sure not to try and use compressed reader for non-compressed logs
2015-09-15 18:52:43 +02:00
baldurk
2100344636
Fix logic to ensure we don't silently drop manual update checks
2015-09-15 18:43:06 +02:00
baldurk
1f3a83cf38
Don't make UpdateDialog TopMost, that's globally always-on-top
2015-09-15 18:42:42 +02:00
baldurk
219bd2d434
Refactor serialisation format with new version
...
* The main improvement is that instead of just callstack resolving DB
and frame capture data, some arbtirary number of sections can be
included, for e.g. adding notes or saving the bookmarks or whatever
use the user wants.
* Sections (namely the frame capture data) can now be (de)compressed on
I/O, which leads to much smaller file sizes.
* Also the default internal alignment has changed from 16 to 64 to be
more future proof (I decided not to go with completely arbitrary
alignment).
* A path for loading previous version logs remains and will be removed
at some point in the future, maybe a version or two down the line.
2015-09-14 09:25:46 +02:00
baldurk
011178cdf1
Update LZ4 to latest
2015-09-13 21:21:14 +02:00
baldurk
e6621c8956
Don't recreate callstacks, instead re-set stack in consistent instance
2015-09-13 21:20:29 +02:00
baldurk
c136dcfcfe
Disable SPIR-V disassembly in non-release as it's non-functional atm
2015-09-13 14:06:38 +02:00
baldurk
5fccf87d43
Only check that DDPF_FOURCC flag is set, not that it's the only flag
2015-09-13 14:06:13 +02:00
baldurk
a01b3badfa
Create initial state for textures/VAOs that don't have it. Refs #148
2015-09-12 18:18:45 +02:00
baldurk
d16a8024d5
Make sure we have some fallback for locating position element. Refs #149
...
* If no position element is selected at all that's really confusing, so
worst case we just pick the first attribute.
* Making it intuitive/obvious that you can change which attribute is
displayed as mesh - that's a different problem.
2015-09-12 17:30:35 +02:00
baldurk
96b43aa54d
Allow the queued capturing of frame 1 immediately. Refs #150
2015-09-12 17:23:30 +02:00
baldurk
193aa5addd
Do lazy disassembling of shaders, only when needed. Refs #136
...
* This is a slightly simpler implementation of the above pull request.
Only the disassembly is done lazily since everything else is pretty
much always requested during log reading.
* I didn't measure much difference of this but I think it depends on
having a lot of shaders with heavyweight shader debug info.
2015-09-12 17:18:49 +02:00
baldurk
4f67e825cf
Store DXBCs in WrappedShader for input layouts. Refs #136
...
* This will allow lazy shader interpretation to just happen in one place
2015-09-12 14:41:13 +02:00
baldurk
1ca3685668
Don't autofit if no texture is viewed (e.g. if a buffer is up)
...
* This fixes a bug from a crash upload
2015-09-11 11:52:02 +02:00
baldurk
a35d18467d
Make sure to set default capture options
2015-09-10 21:36:55 +02:00
Baldur Karlsson
311c7486fd
Manually thumbnail images in README
...
* They don't render correctly in firefox the old way
2015-09-03 20:30:07 +02:00
baldurk
a63c0217dc
Tweak empty array debug prints and add newline
2015-09-02 11:38:08 +02:00
baldurk
65dbac28f0
Change array serialisation functions to explicitly be uint32/uint64.
2015-09-02 09:00:00 +02:00
baldurk
a763656302
Compile fix for linux
2015-09-02 10:36:18 +02:00
baldurk
d2fc251b58
Add a serialise function that calls Serialise on each array element
...
* Basically for the case where an array element isn't POD and needs
special handling for serialisation
2015-08-31 17:53:48 +02:00
baldurk
b9ba4099ba
Rename Serialise() overload that does array to SerialisePODArray()
...
* To emphasise that it's serialised as a blob of memory, and that
Serialise isn't called on each member.
* Restore the ToStr()
2015-08-31 17:39:49 +02:00
baldurk
c1e8df576c
Compile fix
2015-08-31 17:23:39 +02:00
baldurk
277b97efb6
Rename OS module loading function as it clashes with define on win32
2015-08-31 17:12:21 +02:00
baldurk
32b028e5a6
If elems count is 0 when reading array, set resulting pointer to NULL
2015-08-31 17:11:16 +02:00
baldurk
2d12d5abee
Add an update dialog that shows update notes & does upgrade smoothly
2015-08-30 00:11:48 +02:00
baldurk
efc5c92e1a
Add miniz reader functions to header and add wchat ver of file extract
2015-08-30 00:00:12 +02:00
baldurk
95be5454f9
Add a "Find in all files" dialog on ctrl-shift-f to shader viewer
2015-08-29 19:26:50 +02:00
baldurk
9f43f12135
Add a list of all files in shader viewer when there are more than 3
2015-08-29 17:18:28 +02:00
baldurk
24b0f04f17
Remove double loop increment
2015-08-28 22:04:06 +02:00