Created attachment 249874 [details] Update to 8.0-5 Update to version 8.0-5, it brings also two more patches, needed for compilation. It needs some tests, I tested it only with two games.
I think gettid could be replaced by pthread_getthreadid_np. No objections otherwise.
Created attachment 249879 [details] Update to 8.0-5 with pthread_getthreadid_np A new version of the patch with the suggestion of the maintainer in comment #1. Using now pthread_getthreadid_np. As far I see, it is used only in debug messaging, so I think it shouldn't break anything (yes, famous last words :) ). Need some tests, I just tested compilation on FreeBSD 14.0 and a very short run.
Approved. (In reply to Bartek Jasicki from comment #2) > Need some tests, I just tested compilation on FreeBSD 14.0 and a very short run. To be honest, it's probably not worth thoroughly testing.
(In reply to Alex S from comment #3) "Sometimes" I miss reaction buttons here, so I don't need to spam with "I agree". ;) That's true, especially with the last change to the patch. Well, some time passed, and seems like the patch is working. I think that is enough tests. Now we only need to find someone who will commit it.
Note this part -.if ${LLVM_DEFAULT} == 11 +.if (${LLVM_DEFAULT} == 10 || ${LLVM_DEFAULT} == 11) is not correct. This was simplified to no longer take LLVM 10 into account a while ago, and the patch (presumably unintentionally) reverts it.
Created attachment 250139 [details] Update to 8.0-5 with pthread_getthreadid_np and fixed LLVM check (In reply to Gerald Pfeifer from comment #5) Good catch, thank you for checking it. Here is the new version of the patch.
I'll have a look and see to commit this...
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=cb4943510ff163609a697f02398b79d6c8a8d441 commit cb4943510ff163609a697f02398b79d6c8a8d441 Author: Bartek Jasicki <thindil@laeran.pl.eu.org> AuthorDate: 2024-04-26 08:53:16 +0000 Commit: Gerald Pfeifer <gerald@FreeBSD.org> CommitDate: 2024-04-26 08:53:16 +0000 emulators/wine-proton: Update to 8.0-5 This brings two new patches on top of upstream (to maintain compilation). PR: 278288 Approved by: maintainer emulators/wine-proton/Makefile | 4 ++-- emulators/wine-proton/distinfo | 6 +++--- .../files/patch-dlls__ntdll__unix__virtual.c (new) | 18 +++++++++++++++++ .../files/patch-dlls__winevulkan__vulkan.c (new) | 23 ++++++++++++++++++++++ 4 files changed, 46 insertions(+), 5 deletions(-)
I pushed the patch two weeks ago. One recommendation, Bartek, to make things easier to upstream. In files/patch-dlls__winevulkan__vulkan.c, instead of replacing the code for other platforms, it's probably better to add a configure check for pthread_np.h and then conditionally include it as already the case below with #ifdef HAVE_SYS_SYSCALL_H Then use the same pattern in signaller_worker as with HAVE_SYS_SYSCALL_H.
The suffix np stands for non-portable, so that part in fact should only have platform-specific ifdefs (__linux__, __FreeBSD__, etc).
(In reply to Gerald Pfeifer from comment #9) Thank you for your advice, I will try to remember it. :) I think in this package (wine-proton), removing a part of a code may be allowed. The upstream (Valve) doesn't accept any patches, plus they don't care about anything except Linux. Our patches often remove such platforms' specific code. Perhaps these patches will not work with (or not needed in) the standard version of Wine, the differences between both versions are too big.