Package Details: wine-git 9.19.r0.g7ee99608f46-1

Git Clone URL: https://aur.archlinux.org/wine-git.git (read-only, click to copy)
Package Base: wine-git
Description: A compatibility layer for running Windows programs (git version)
Upstream URL: https://www.winehq.org/
Keywords: windows wine
Licenses: LGPL-2.1-or-later
Conflicts: bin32-wine, wine, wine-wow64
Provides: bin32-wine, wine, wine-wow64
Replaces: bin32-wine
Submitter: None
Maintainer: dbermond
Last Packager: dbermond
Votes: 86
Popularity: 0.000028
First Submitted: 2007-07-18 16:01 (UTC)
Last Updated: 2024-10-06 15:02 (UTC)

Dependencies (99)

Required by (378)

Sources (3)

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 7 8 9 10 .. 15 Next › Last »

dbermond commented on 2018-07-03 17:12 (UTC)

@Enverex If you are building with avx support enabled then you're modifying the default Arch Linux build flags. By doing this you're pretty aware of the issues that it can cause (like the one you're having). Users wanting the most vanilla experience should 1) not modify the build flags, and/or 2) build the package in a clean chroot. It's user reponsibility to control the build flags, unless some major intervention is required.

Forcedly disabling avx support is not a good idea. Firstly, because it's not intended to use avx by default in any way (due to the default Arch build flags being x86_64 generic ones). Secondly, because it will impact users that want to benefit from avx support. Thirdly, because this is affecting only your application. If it would be a broad wine issue affecting a huge amount of applications, then well, we could disable avx here, but this is not the case.

I suggest you to fill a bug report in upstream wine saying that you're having issues in your application when avx is enabled, instead of asking for forcedly disabling avx here.

Enverex commented on 2018-06-29 12:49 (UTC)

@dbermond - Wine will literally just crash if it tries to use it, there is no benefit to leaving it enabled. You have 2 possible situations:

  1. The user doesn't have an AVX capable CPU, the flag won't do anything because AVX was never supported in the first place.

  2. The user has an AVX capable CPU. As soon as Wine tries to use AVX, Wine will crash. Adding the flag negates this situation.

Essentially leaving AVX enabled will do nothing other than cause Wine to crash whenever it's used (universally, until it's fixed upstream, if that will actually happen). There is zero positive to enabling the flag (which will be enabled by default for anyone using native or CPU specific compiler flags).

But it's up to you.

dbermond commented on 2018-06-05 15:26 (UTC)

@Enverex I don't think it's a good idea to force users to compile without avx support. It can be a user choice to use avx.

Enverex commented on 2018-06-04 14:32 (UTC)

@dbermond - I've actually tracked down the real reason now. It will technically affect anyone with an AVX capable processor but only triggers in very rare circumstances: https://bugs.winehq.org/show_bug.cgi?id=45289

May be worth adding "-mno-avx" to C/XXFLAGS in the build section somewhere to negate it.

dbermond commented on 2018-06-03 02:28 (UTC)

@Enverex I'm glad that you found the answer. But yes, this package is the same as wine from the official repositories, except for being the development version.

Packages from the official repositories are built in a clean chroot to avoid linking against undesired libraries and other interferences that may occur. For example, in a chroot it will not link against any nvidia specific library that the user might have installed. In order to better reproduce the official repository package, the user should build it in a clean chroot. And as you know, official repository packages are built with generic x86_64 compiler flags, and this can also cause runtime differences.

Enverex commented on 2018-06-02 22:25 (UTC) (edited on 2018-06-02 23:07 (UTC) by Enverex)

Cemu crashes in Mario Kart 8 with this (even when set to the 3.9 release commit specificall) but not the stock repo Wine.

The actual build looks pretty much the same between this and the stock Arch repo version. Is there anything else different in the build script here that could be responsible for this different behaviour?

EDIT: Looks like it may be a GCC issue. It only happens if compiled with -march=native (i7-8700K).

dbermond commented on 2017-12-11 19:26 (UTC)

@EndlessEden Segfault at compile time? Package is building fine for me.

EndlessEden commented on 2017-12-11 09:03 (UTC)

i segfault with this. idk why...

dbermond commented on 2017-10-10 20:35 (UTC)

I'm maintaining this package now. After some cleanups, it's now building and working fine. Enjoy.

Vash63 commented on 2017-09-28 06:09 (UTC)

Follow up - #2 below was fixed, I can now build and use wine-git when removing the Flex.patch and with no other changes.

« First ‹ Previous 1 2 3 4 5 6 7 8 9 10 .. 15 Next › Last »