#wine-staging

did i read right the other day that there will be problems building wine due to archlinux' move to gcc 5.1?

[06wine](00http://is.gd/9s6lQX) 02Nikolay Sivov: dwrite: Implement GetGlyphOrientationTransform().

[06wine](00http://is.gd/o5pjfx) 02Nikolay Sivov: riched20: Convert from twips to points when tomUsePoints is used.

[06wine](00http://is.gd/ZSwKxt) 02Nikolay Sivov: riched20: Switch font unit is not allowed.

[06wine](00http://is.gd/QROLUx) 02Nikolay Sivov: riched20: Added GetStoryType().

[06wine](00http://is.gd/FgdLCJ) 02Nikolay Sivov: riched20: Implement SetFont().

[06wine](00http://is.gd/V4We7W) 02Nikolay Sivov: riched20/tests: Test more return values and released cases.

[06wine](00http://is.gd/XFgovF) 02André Hentschel: ntdll: Add stub for __chkstk on ARM.

[06wine](00http://is.gd/Dpt6gd) 02André Hentschel: ntdll: Fix returning from various ARM assembler functions.

[06wine](00http://is.gd/BteDUu) 02Alexandre Julliard: krnl386.exe: Release the Win16 lock when loading a 32-bit dll.

[06wine](00http://is.gd/sIT8Y0) 02Damjan Jovanovic: qcap/tests: Fix SmartTeeFilter reference leaks from the pins in its test.

[06wine](00http://is.gd/U6qrJL) 02Damjan Jovanovic: qcap/tests: Add many more SmartTee filter tests.

[06wine](00http://is.gd/gCh1WT) 02Damjan Jovanovic: qcap/tests: Add media tests for the SmartTee filter.

[06wine](00http://is.gd/YDY9Xd) 02Damjan Jovanovic: qcap/tests: Test SmartTee filter aggregation.

[06wine](00http://is.gd/JTeYID) 02Damjan Jovanovic: qcap: Implement a stubbed SmartTee filter.

[06wine](00http://is.gd/9gzB41) 02Damjan Jovanovic: qcap: Implement the SmartTee media type functions.

[06wine](00http://is.gd/3mGZ2X) 02Daniel Lehman: msvcr120: Add erf functions.

[06wine](00http://is.gd/2dtjl7) 02Iván Matellanes: msvcirt: Share exceptions implementation with msvcp.

[06wine](00http://is.gd/gtmWkf) 02Matteo Bruni: wined3d: Remove point size scaling hack.

[06wine](00http://is.gd/eZtzIB) 02Matteo Bruni: wined3d: Introduce a get_pointsize_minmax() function.

[06wine](00http://is.gd/BYvDdG) 02Matteo Bruni: wined3d: Introduce a get_pointsize() function.

[06wine](00http://is.gd/5H2L8w) 02Matteo Bruni: wined3d: Avoid the builtin uniforms for point size when using the fixed function vertex pipeline.

[06wine](00http://is.gd/5TyBA9) 02Alexandre Julliard: Release 1.7.44.

JackWinter: there should be ArchLinux Wine maintainers on this channel

so you have higher chances asking here is gcc-5.1 is severely broken for Wine

JackWinter1: <puk> JackWinter: there should be ArchLinux Wine maintainers on this channel

<puk> so you have higher chances asking here is gcc-5.1 is severely broken for Wine

well i'm one of them (wine-rt), but i'll just have a go. been asking a few times already without getting an answer :)

heh

JackWinter: there are various report that compiling with gcc 5.1 results in broken binaries, but you'll have to try it yourself to find out how critical it is ;)

slackner: just copy protected stuff breaks?

or whole Wine?

aka winecfg / winemine?

various reports with 5.0 were talking about whole wine is broken, including prefix setup

for 5.1 at least a couple of apps are broken

well, .0 is basically beta now for gcc

even officially

slackner: that console/editline bug you gave me turned into 3 bugs :/

hmc: ugh ... well, that happens almost all the time when i try to fix sth ^^

:D

hmc: you didn't fix them yet?

;-P

no

it took me 20 minutes to decipher the code

there's no indentation

:)

autoformatter

hmc: you have luck that it isn't shell32 code ;)

Bradan: no, not allowed for wine

slackner: but for viewing code

format it, understand it, undo it.

Bradan: thats possible of course xD

hmmm. i wonder if Alexandre would agree to format the codebase

hmc: no, he doesn't

hmc: there were various requests in the past to do that

pity. some of it really messy

*is really messy

but it would take forever...

i once did this in wine staging for some shell32 file, but it will never get upstream: https://github.com/wine-compholio/wine-staging/...

shell32 has an ugly mix of tabs and spaces

DarkPlayer: nice cleanup

slackner: how would you touch it? by writing or reading?

Andre_H: reading should be sufficient, its a guard page so you'll get an exception for all kind of memory access

Andre_H: the main problem is that its not properly documented how exactly the function works. does it access N bytes below the current stack pointer, or is there some additional offset involved? ^^

fwiw, it seems wine built for multilib at the least, so i guess it's ok for building. let's see how broken it is for running... what are you guys gonna do for archlinux? wait with building binaries, or ship binaries and see what bug reports come in?

JackWinter: for our builders i've locked the gcc package, but i am also aware that this is no suitable solution in the long term

JackWinter: will probably package an old version of gcc and use that for building

JackWinter: and: building isn't really the interesting part, this always worked. having broken builds is the much more critical issue

i suppose i could probably downgrade gcc in my chroot for building

yeah agreed. so far i've only seen the 3.14 kernel having problems building

slackner: happen to know if gcc is the only thing that needs "locking", or does it go hand in hand with libtools or something else?

I've locked gcc, gcc-* and *-gcc-*

so gcc-libs too

it worked when i pushed our builds for 1.7.43, but i fear you'll have to remove the lock (or downgrade various additional libs)

i fear the same ;)

i hate it when archlinux does this with gcc... would much prefer that they kept it a long time in testing

JackWinter: well, for critical packages as gcc it would be nice to have just multiple versions available in the repos

:-P I had the same problem with wpa_supplicant and clang35/36

but they downgraded wpa_supplicant back to 2.3 from 2.4

whee the only program i normally use does this :)

wine: Unhandled page fault on write access to 0x00000000 at address 0x7f0a7a22ee7f (thread 000b), starting debugger...

err:module:DelayLoadFailureHook failed to delay load user32.dll.LoadStringW

wine: Call from 0x7b84bd77 to unimplemented function user32.dll.LoadStringW, aborting

guess i'll downgrade again, and leave wine-rt in the aur at 1.7.43

wat?

with 1.7.43 the app starts.

JackWinter: sounds like its still worse than just a couple of broken apps xD

now you make me upgrade to F22 for the gcc-5.1 experience

go archlinux. i remember the last time they did a major change of gcc, and the kernel broke all the out of tree modules. seems like they forget to test things well before making the plunge

i started with Red Hat Linux 4.1 back in the days

and work for Red Hat

think the dev built the kernel on his system running testing, instead of doing it in a chroot, so they had to upgrade gcc and then rebuild a load of packages

so I'll stick to what I know

fedora is by all accounts a nice distro, never ran it though (did run some kinda red hat 15-20 years ago though)

well, my first Linux usage was Slackware

and I've run Debian for a while

as I didn't like RHL 5.0

joined back at RHL6.1

wonder if i should write something on the aur that i won't be upgrading the build script until arch figures out what gcc 5.1 broke B)

think that i'd probably go for fedora if i moved off archlinux