#wine-staging

mwhahahaha: https://dl.dropboxusercontent.com/u/195059/wine...

its part of the windows driver, and to use pcap there is no need to run that one

need to retitle it, but MWAHAHAHAHA!

ehoover: how many test failures? ;)

ehoover: and does it fix the original issue?

slackner: 0

and yes

oh, patchupdate finally finished!

:D

!

lol

what do you think I call this one? shell32-ShellExecute_with_Path?

ehoover: yep, should be fine

ehoover: is there also a test for this specific case ?

not yet, but there will be in a minute ;)

slackner: do we know if hearthstone uses these fields?

ehoover: no, can you read my mind? ^^

ehoover: was thinking the same

lol

ehoover: too bad i destroyed my battle.net acc xD

i'll download it ;)

ehoover: so you'll have to test it yourself ^^

wow, there are a lot of things we're not using in that structure

hmm, i have no idea how to test that this is what it really does without writing an app that tries to look at how it was called

ehoover: hmm

ehoover: maybe a test which copies the exe file ?

slackner: at what point does hearthstone fail?

ehoover: noone has figured out exactly why it fails yet - first an agent process is spawned

which checks for updates, afterwards a second agent process is spawned

(both using this function)

so, this is after the game is installed?

yes, installation works fine

it happens when starting the game

(needs a battle.net acc)

k, i'm downloading right now

slackner: so what are you thinking wrt. copying the executable?

ehoover: when i understand the problem correctly the problem (in this bug report) has to do with which executable is picked, either in the working dir or in the provided path

ehoover: you could probably copy the test executable to a subdirectory to test that

ehoover: so that both the cwd and the subdir contains the same exe

slackner: ah, i get your meaning

ehoover: and then the test should check which one is executed (can be determined with GetModuleFileName or similar)

i have another idea

lpDirectory = parentdir, lpFile = subdir/file.exe

no executable in lpDirectory

so the launch fails

(with old code0

)

yep, that would also work

ehoover: does it work? ^^

slackner: no

:(

what exactly, on windows or wine ?

it works for the other game, but not hearthstone

ah, i thought you mean the test

slackner: oh, the tests i've come up with so far work both ways - it's odd

ehoover: hm, i guess by default the current .exe directory is always in the search path :/

which means you would really have to copy it somewhere

that gives me an idea

wait, that's no different from what i'm trying already

hmm

slackner: it's a far more fundamental problem for stronghold

i need to go to bed, but it's either that RtlDosSearchPath_U needs to search the current path or that SearchPathW needs to prepend/append the current working directory to the path list

based on the docs at http://msdn.microsoft.com/en-us/library/windows... i suspect that it's up to SearchPathW

slackner: https://bpaste.net/show/9042b9e64b1d any ideas why adobe air doesn't appear to be working now that I've updated wine?

NP-Hardass: hm, with which version was it working before? such errors are always a bit difficult

because it could be either an error in one of the libX libraries, or maybe even a wine bug

or also a driver bug, when you are using an intel GPU

slackner: 1.7.33

slackner: I'll try switching wine versions to pinpoint it later

NP-Hardass: and which version is not working? 1.7.33 is the latest release

slackner: 9999

and i am not aware of any git changes in either wine or wine-staging which could make a difference

did you maybe also install other new packages?

slackner: none that I am aware of. I'll muck around and try to nail down an exact cause

slackner: two points, one works fine in 1.7.33, two, when 9999 starts, it complains abut being built without libcms support (which makes no difference to this bug)

slackner: you think it is a wine bug, not a wine-staging bug?

building with/without libcms could be the difference, i think there was a change recently

let me check

hm, no, that fix did not get upstream yet

slackner: I was claiming that it doesn't because I recompiled the 9999 with libcms support and still experience adobe air issues

hm, why does configure behave differently then ?

NP-Hardass: can you show the full terminal output ?

slackner: yeah, give me a few minutes. in the meantime, let me describe how air is misfunctioning

basically, when I launch an air application (including the airappinstaller the window doesn't load immediately, and sort of hangs, anything drawn over it stays on top of it when it regains forcus. I found out last night that if I left my air application running, eventually it would load. every action that I took basically took forever to happen. minutes to load, minutes to perform each subsequent

action in the application

slackner: actually, it looks like it is complaining about the libcms in 1.7.33 when it is functioning. red herring.

NP-Hardass: hm, weird

NP-Hardass: as a first step i would suggest to show the full output

NP-Hardass: if that doesn't help, it would be really useful if you could run some quick test builds with USE=-staging

is the software freely available? then i can also take a look myself in a couple of minutes

slackner: >___<

? xD

so apparently after numerous (at least 7-8 times) of experiencing that issue... I just re-emerged the (binary for) 9999, wineserver -k'd and it's functioning normally...

hm, wut? are you sure it works all the time now?

or is it a race-condition, that would be even worse than just a regression

just ran it 4 times, each without issue :(

I haven't a clue now :P

hm, please continue testing a bit ^^

maybe it was related to the upgrade, for example still an old wineserver version running ?

the wineserver version check just checks for an incrementing number, so its not completely unlikely that the number patches, but the versions are still incompatible

*number matches

if you start wineserver and some apps with 1.7.33, and others with -9999 that might cause trouble

when I experienced the problem initially, I killed the wineserver just in case. so that would seem unlikely

hm, yeah :/

just in case you wanted to try it on your system: http://blog.coursevector.com/downloads/Minerva-...

slackner: so if air takes forever to install that, and the window is wonky, you've reproduced it

have you used air in wine before?

nope

slackner: air is a little picky about paths

slackner: so you'll want to put the .air file in C:\Program Files (x86)\Common Files\ Adobe AIR\Versions\1.0\Resources

slackner: then cd to that directory

and run wine airappinstaller.exe blah.air

and that should work

i have a question. if you use windows and install steam game 1st time, does it always want to install dx9 and other stuff? or does it see that you already have them installed?

also if this happens with wine, does those installers override wine settings?

jarkko_: I suspect that it depends on the installer