Getting a lockup when app loses focus for 5+ seconds on Windows 10 when using SDL_INIT_JOYSTICK. Only happens when running the .exe from file manager. Reported the bug here:
Not yet. I plan to do a pass and go through all of these, but not for 2.0.9.
Thanks for taking the time to write this up. I went through and reviewed the patches and accepted them all for 2.0.9.
Do you have specific bugs that arenât already reported?
I went ahead and commented on the two WASAPI bugs in bugzilla:
https://bugzilla.libsdl.org/show_bug.cgi?id=4171
https://bugzilla.libsdl.org/show_bug.cgi?id=4066
No I donât, but the issue is that Iâm getting some buffering problems (most likely underrun) in some rare occasions when using the callback. This is not acceptable, since Iâm developing a music program.
On some machines it happens way more often. Increasing the buffer size doesnât decrease the cracks/pops at all, which is very strange! Since I read that other people had glitchy problems with WASAPI on SDL2 as well, I concluded that itâs an SDL2 problem and not my code (which has no issues when using directsound, by the way).
Some bugzilla entries that might be of interest for 2.0.9, some
of which have proposed patches:
https://bugzilla.libsdl.org/show_bug.cgi?id=2954
https://bugzilla.libsdl.org/show_bug.cgi?id=3589
https://bugzilla.libsdl.org/show_bug.cgi?id=4066
https://bugzilla.libsdl.org/show_bug.cgi?id=4171
https://bugzilla.libsdl.org/show_bug.cgi?id=4231
https://bugzilla.libsdl.org/show_bug.cgi?id=4236
https://bugzilla.libsdl.org/show_bug.cgi?id=4238
https://bugzilla.libsdl.org/show_bug.cgi?id=4241
https://bugzilla.libsdl.org/show_bug.cgi?id=4242
https://bugzilla.libsdl.org/show_bug.cgi?id=4244
https://bugzilla.libsdl.org/show_bug.cgi?id=4246
https://bugzilla.libsdl.org/show_bug.cgi?id=4255
And some very recently reported ones, some of which may already
be under investigation:
https://bugzilla.libsdl.org/show_bug.cgi?id=4272
https://bugzilla.libsdl.org/show_bug.cgi?id=4278
https://bugzilla.libsdl.org/show_bug.cgi?id=4284
https://bugzilla.libsdl.org/show_bug.cgi?id=4285
https://bugzilla.libsdl.org/show_bug.cgi?id=4286
https://bugzilla.libsdl.org/show_bug.cgi?id=4287
Great. Iâve updated my Go language bindings at
Could you please add them to the list of Go bindings on the page
https://www.libsdl.org/languages.php
MSB
Changeset 12286 seems to have broken compiling on the Raspberry Pi. Eariler changesets compile fine.
> /usr/src/SDL-75abb8e13a91/src/render/software/SDL_render_sw.c: In function âSW_CreateRendererForSurfaceâ:
> /usr/src/SDL-75abb8e13a91/src/render/software/SDL_render_sw.c:154:13: error: âSDL_Renderer {aka struct SDL_Renderer}â has no member named âSetTextureColorModâ
> renderer->SetTextureColorMod = SW_SetTextureColorMod;
> ^~
> /usr/src/SDL-75abb8e13a91/src/render/software/SDL_render_sw.c:155:13: error: âSDL_Renderer {aka struct SDL_Renderer}â has no member named âSetTextureAlphaModâ
> renderer->SetTextureAlphaMod = SW_SetTextureAlphaMod;
> ^~
> /usr/src/SDL-75abb8e13a91/src/render/software/SDL_render_sw.c:156:13: error: âSDL_Renderer {aka struct SDL_Renderer}â has no member named âSetTextureBlendModeâ; did you mean âSupportsBlendModeâ?
> renderer->SetTextureBlendMode = SW_SetTextureBlendMode;
> ^~
> /usr/src/SDL-75abb8e13a91/src/render/software/SDL_render_sw.c:161:13: error: âSDL_Renderer {aka struct SDL_Renderer}â has no member named âUpdateViewportâ; did you mean âUpdateTextureâ?
> renderer->UpdateViewport = SW_UpdateViewport;
> ^~
Thank you!
Anyone else noticed that vsync doesnât work with SDL2.0.8 on macOS Mojave? Is it because of lack of OpenGL, and can it be fixed?
Hi @slouken!
Iâm very happy to see a dedicated sensor API that I needed for my future games. Iâm in hurry to test.
However Iâm always wondering why SDL only bump the patch version when adding new features like that. The minor version should be updated instead. For example SDL 2.1.
Cheers!
the minor version is not bumped as long as the API and ABI isnât broken (changed in in a backwards-incompatible way)
We usually bump minor version when API changed are made in backward compatible way (e.g. new features). And patch versions mostly refer to bugfixes. ABI breakage is another stuff that is handled at shared library level (using SONAME) and not at the version at all. This is the most common versioning system.
This was off on a branch, which happened to be the âtipâ revision if you pulled at the wrong time.
(but that branch should now work on the Raspberry Pi again, too.)
âryan.
Excellent, thank you for looking into this.
I take the view that if only the âpatchâ figure is stepped there ought to be a very high likelihood that an application requires no code changes in order to work with the new version of SDL2. But thatâs not the case in 2.0.9: for example according to the prerelease announcement the default behavior of the Android âbackâ button has been changed, necessitating the addition of a hint to restore the previous behavior. It has been true of earlier releases too, such as when the default Windows audio driver was changed to WASAPI. They deserve a change to the âminor versionâ, in my opinion.
The default behavior of the Android back button hasnât changed - you can just now specify a hint to get the back button as a key press event.
The WASAPI change was intended to be transparent, and any case where that isnât true is a bug.
I get the back button as an SDL_KEYDOWN event now, using SDL 2.0.5, (ev.key.keysym.sym = SDLK_AC_BACK
as documented here) so Iâm confused about what the new hint actually changes.
The iOS Demo xcode project Demos.xcodeproj is missing the information to link against the CoreBluetooth.Framework.
Same probably in the Template example project file.
Also Xcode 10 complains about âTraditional headermap style is no longer supported; please migrate to using separate headermaps and set âALWAYS_SEARCH_USER_PATHSâ to NO.â
If you set this to no, it compiles fine.