- wine-gitlab - winehq.org

[PATCH 0/1] MR3503: ntdll: Try to raise the file limit above OPEN_MAX on MacOS.
by Stefan Dösinger (＠stefan) 03 Aug '23

03 Aug '23

OPEN_MAX is 10240, which is usually lower than the kern.maxfilesperproc sysctl value. Said value from sysctl works on Mojave to Monterey (and most likely earlier Mac OS versions, but I can't test). Since Big Sur we can successfully set the reported rlim_max value. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/3503

2 1

[PATCH 0/1] MR3494: include: Update SECURITY_LOGON_TYPE and SECURITY_LOGON_SESSION_DATA in ntsecapi.h.
by Vitaly Lipatov (＠lav) 03 Aug '23

03 Aug '23

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/3494

3 4

Re: [PATCH v5 0/2] MR3473: whoami: Refactor and add more commands.
by Maxim Karasev (＠mxkrsv) 03 Aug '23

03 Aug '23

On Thu Aug 3 07:27:22 2023 +0000, Hans Leidekker wrote: > Can you give an example where wprintf() doesn't do the right thing? Printing a non-english username. `wprintf()` by default doesn't print non-english letters at all (I remember testing various system LC_ALL variants, like `C.UTF-8`, `en_US.UTF-8`, `ru_RU.UTF-8`, and none of that took an effect). With `setlocale(LC_ALL, something)` it does print the letters, but 1. Only of a language that is set, so I presume it chooses something like a `cp1251` equivalent (an 8-bit language-specific encoding?). 2. It isn't used in any of other wine programs, while that `ConsoleWriteW()` hack is used at least in present whoami and xcopy implementations (probably more). This implementation, in turn, behaves exactly like the Windows whoami. Also when testing the `klist` I noticed that on Windows it was printing `?` signs instead of `Server` string (somehow hooked up a translation?), so `ConsoleWriteW()` seems to be necessary for any translatable console programs if one wants to avoid `setlocale()`. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/3473#note_41167

1 0

[PATCH 0/6] MR3502: LDAP session key
by Hans Leidekker (＠hans) 03 Aug '23

03 Aug '23

This can happen when credentials are cached. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/3502

2 6

[PATCH 0/2] MR289: spirv: Clamp ftou output.
by Zebediah Figura (＠zfigura) 03 Aug '23

03 Aug '23

SPIR-V says this is undefined behaviour, but Direct3D actually specifies that it should clamp. Most drivers do clamp, but llvmpipe does not. Hence this fixes a couple of tests with llvmpipe. This does of course add overhead to the ftou instruction, but I cannot imagine that it is a common instruction to execute. This also is not the first time we perform such checks; cf. udiv. -- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/289

3 7

[PATCH v5 0/3] MR3472: secur32: Convert "Negotiate" to an LSA package.
by Dmitry Timoshkov (＠dmitry) 03 Aug '23

03 Aug '23

.Net assumes that Negotiate LSA package has index 0: https://github.com/dotnet/runtime/blob/f1d2be8f443a599a142506d622a35e1c2e68… -- v5: secur32/tests: Add some tests for "Negotiate" LSA package. secur32: Add a better stub for LsaApCallPackageUntrusted(NegGetCallerName). secur32: Convert "Negotiate" to an LSA package. https://gitlab.winehq.org/wine/wine/-/merge_requests/3472

3 4

[PATCH v2 0/5] MR290: hlsl: Replace bare list pointers with hlsl_block, part 7.
by Zebediah Figura (＠zfigura) 03 Aug '23

03 Aug '23

-- v2: vkd3d-shader/hlsl: Use hlsl_block_add_instr() in clone_block(). vkd3d-shader/hlsl: Clean up the static_initializers block when the context is destroyed (Valgrind). vkd3d-shader/hlsl: Pass a hlsl_block pointer to dump_instr_list(). vkd3d-shader/hlsl: Store the "instrs" field of struct hlsl_attribute as a hlsl_block. vkd3d-shader/hlsl: Pass an hlsl_block pointer to add_load_component(). https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/290

3 6

Re: [PATCH v5 0/2] MR3473: whoami: Refactor and add more commands.
by Hans Leidekker (＠hans) 03 Aug '23

03 Aug '23

Hans Leidekker (@hans) commented about programs/whoami/main.c: > +#define WIN32_LEAN_AND_MEAN > #define SECURITY_WIN32 > #include <windows.h> > #include <security.h> > +#include <sddl.h> > +#include <stdarg.h> > > #include "wine/debug.h" > > WINE_DEFAULT_DEBUG_CHANNEL(whoami); > > -static int output_write(const WCHAR* str, int len) > +/* Behaves like the regular wprintf, but is able to fall back to the OEMCP > + * when redirected. > + * Most importantly, it prints fully in Unicode by default, > + * so any language characters are shown regardless of the current locale. Can you give an example where wprintf() doesn't do the right thing? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/3473#note_41157

1 0

[PATCH v4 0/3] MR3472: secur32: Convert "Negotiate" to an LSA package.
by Dmitry Timoshkov (＠dmitry) 03 Aug '23

03 Aug '23

.Net assumes that Negotiate LSA package has index 0: https://github.com/dotnet/runtime/blob/f1d2be8f443a599a142506d622a35e1c2e68… -- v4: secur32/tests: Add some tests for "Negotiate" LSA package. secur32: Add a better stub for LsaApCallPackageUntrusted(NegGetCallerName). secur32: Convert "Negotiate" to an LSA package. https://gitlab.winehq.org/wine/wine/-/merge_requests/3472

3 5

[PATCH v2 0/1] MR3491: winegstreamer: Leave media source critical section before unlocking workqueue.
by Bernhard Kölbl (＠besentv) 03 Aug '23

03 Aug '23

Today, if media_source_Shutdown is called around the same time as a work item is put on the async_commands_queue, we end up in a deadlock if Shutdown enters media source's cs first, as it waits for the queue's callback to finish, which, in turn, waits for the object's cs to be released. To avoid this leave the cs, before unlocking the workqueue, to let any callback on the queue finish running. Signed-off-by: Bernhard Kölbl <bkoelbl(a)codeweavers.com> Blocks !3351 -- v2: winegstreamer: Leave media source critical section before unlocking workqueue. https://gitlab.winehq.org/wine/wine/-/merge_requests/3491

3 2

2025

2024

2023

2022

wine-gitlab