#wine-staging

ehoover: ping ?

slackner: pong!

repeat: slackner: wow, he took the fd caching patch!

lol

ehoover: yep, lol xD

ehoover: although i broke PowerPC support xD

we could submit a patch to use locking on PPC

ehoover: i've already prepared such a patch, but i can't test it :/

slackner: who needs PowerPC anyway?

Bradan: no idea xD

slackner: hmm, i can check my wife's old mac and see if it's PPC

ehoover: i wanted to ask because of the recent regressions, at least the one i already debugged a bit seems to be related to https://github.com/wine-compholio/wine-staging/...

slackner: I might have an idea *turnsNintendoWiiOn* ... let's try it there ... xD

ehoover: unfortunately no upstream bug is specified, so i'm not sure for what the patch was originally written for

slackner: shell32-Default_Folder_ACLs ?

ehoover: yep

slackner: one sec, i think i remember which app that's for

it's Finale Notepad 2012 (https://bugs.winehq.org/show_bug.cgi?id=34406)

04Bug 34406: normal, P2, ---, wine-bugs, NEW , Finale Notepad 2012 doesn't copy/create user files on program start

ehoover: the last comment says its already fixed, which means we could remove/disable this patch, and probably fix the regression

slackner: hmm, what's the regression?

(these folders should have inheritable attributes)

ehoover: https://bugs.wine-staging.com/show_bug.cgi?id=265

04Bug 265: normal, ---, ---, bugs, CONFIRMED , Sqlite3 reports 'database disk image is malformed'

ehoover: from what i found out so far it has to do with inheritance in combination with this patchset

its probably also related to https://github.com/wine-compholio/wine-staging/...

slackner: hmm, maybe

ehoover: also, bela found a couple of other regressions - might be possible that those are unrelated

slackner: i would recommend that we just disable it for now

ehoover: but i somehow fear that its related to the ACL rework

ehoover: now attributes don't get lost anymore xD

slackner: what do you mean by that?

ehoover: in 1.7.41 i changed a lot of the ACL stuff because i noticed that tests failed randomly - i spotted a couple of errors which had the effect that the wineserver SD and the XATTR got out of sync

hmm

so you think there's some regressions because we're properly keeping all the attributes?

ehoover: i haven't figured out what exactly goes wrong (didn't find any explicit ACL checks), but yes, thats what i think

ehoover: the main issue are probably the additional test failures in https://github.com/wine-compholio/wine-staging/...

slackner: well, we probably want to figure out a good way to do FILE_ATTRIBUTE_READONLY then

we should be able to handle that with or DOS attributes paches

*patches

ehoover: i'll do some more tests with this sqlite3 app, if i can't figure out more details i'll disable the other patchset until we have fixed the remaining test failures

[06staging/master][05e0390d0](00http://is.gd/5QY8wy) 02Sebastian Lackner: Removed various patches (accepted upstream).

mogorva: btw, i've figured out why the uplay "play" button doesn't work for demos - it tries to open an url with steam:// which obviously can't work when steam is not installed

mogorva: with steam installed it seems to work, but don't ask me how to download demos in uplay directly O_o not sure if they removed that feature

mogorva: can't test with windows unfortunately, so not sure how it is supposed to behave ^^

ehoover, DarkPlayer: hm, upstream wine doesn't compile anymore lol

hmm?

desktop.o:(.data.rel.ro+0x0): undefined reference to `GUID_NULL'

ccache bug ?

can someone else reproduce ?

guilty patch -> 58775c878ae7907812824628c5320f9a0239fdd7

slackner: [06wine](00http://is.gd/ua0f4O) 02Nikolay Sivov: explorer: Added IDispatch support for ShellBrowserWindow instance.

nsivov: i assume you didn't get any build failures with that patch? ^

Hello

any help ? http://pastebin.com/wQK23K1Y I'm lost

[06staging/master][053f7b489](00http://is.gd/bdO96p) 02Sebastian Lackner: Compiler_Warnings: Added patch to fix recently introduced build failure with gcc4.9.

Phrk_: the output does not contain much useful information unfortunately - do you know if it should work ?

slackner: thx that exactly what i think, useless informations

on winehq it's gold

so yes it "should" work

Phrk_: try to test with a version which is very similar to the upstream report

Phrk_: as you're using POL it should be easy to switch versions

I already have same error

Phrk_: are there any other hints on appdb?

:'(((

Phrk_: also, i would suggest you test with a fresh wineprefix

Phrk_: based on the name i assume you created the steam prefix orginally for a different game

Disable steam overlay, i already have

Phrk_: and installing too many games into one prefix is never a good idea, there are even compatibility issues on windows

Well i have 3 other games, and they work very well

And they recent

Phrk_: well, but to get those other games working you probably installed native overrides?

native overrides ?

yes, native windows dlls to workaround specific bugs

they might work well with one specific app, but not in general

I think i created a new prefix with nothing than wine-stagging

but it's a old prefix so yo maybe right

i will try with a new one

slackner, no, I usually try to beat the urge to send if it doesn't build

what broke exactly?

nsivov: see above, linking raises a "undefined reference to 'GUID_NULL'" error. i've fixed it with http://is.gd/bdO96p

gcc -v?

nsivov: http://ix.io/ihb

(on arch)

mine is close, 4.9.2 from deb, how do you build it? from top dir usual way?

nsivov: if you mean wine, yes, just running the regular makefile - i also ran "make clean", but the error remains

nsivov: i haven't tested yet if its a ccache bug, but rather unlikely, it was working very well all the time ^^

nsivov: let me quickly run it without ccache, just to be sure

make clean on explorer itself does trigger it too?

hm, I tried both, and it works here

I don't see how 'make clean' on top level could affect that though

nsivov: the problem also occurs without ccache - either my build tree is somehow corrupted or its caused by compiler differences

can someone else in here test ?

slackner: i compiled today's git on Fedora 21 without problems, gcc version 4.9.2 20150212 (Red Hat 4.9.2-6) (GCC)

meh, 3 weeks older

slackner, interesting, I do get it now

nsivov: broken for me too (gcc version 4.8.2 (Ubuntu 4.8.2-19ubuntu1))

nsivov: hm, what did you change? ^^

slackner, I applied another patch to this file, unrelated to this change

slackner: IT WORKS WOOOT THX YOU !!!

Phrk_: nice :) np

nsivov: does https://github.com/wine-compholio/wine-staging/... work for you, and do you think thats the best approach to solve it?

slackner: [06staging/master][053f7b489](00http://is.gd/bdO96p) 02Sebastian Lackner: Compiler_Warnings: Added patch to fix recently introduced build failure with gcc4.9.

slackner, I think I'll add DEFINE_GUID instead

DarkPlayer: about a macro that turns a string into wine's way of expressing a WCHAR string, i only know of a few macro's or template tricks thatll do smth like it

the macro ones are basically recursive macro's, limiting the string length by how deep a compiler can handle macro recursion

the template ones are c++11

afaik

learn_more: templates won't be an option since wine uses C only

:(

writing com with c

that's not healthy

almost all solutions for this use a template

and/or c++11

and for the macro solution it depends on the compatibility between compilers and recursion depth

i had an old thing that did it with macro's only, but limits the length to ~80 or so

its original purpose was to obfuscate a string compile time

maybe it can be fixed to do just this

i will try to make it work :)

not sure if it got accepted: https://gcc.gnu.org/ml/gcc-patches/2008-03/msg0...

u'c-char-sequence'char16_t character literal (UTF-16)

learn_more: a solution using c++ and/or templates will not help, wine tries to use a very old C standard ^^

but most probably this only works with gcc

ya that wont work with msvc

afaik

it should be possible to map it with a macro to L"..."

hm

in debug build it might incur overhead lol

slackner, there's another cleaner way, I'll send a patch

nsivov: kk

yuk

forgot how ugly that macro was