[Wine] Rainmeter desktop display, skins not transparent

[Caledfwlch]

Firstly, I've filed a bug about this transparency issue here ( http://bugs.winehq.org/show_bug.cgi?id=18144 ) as well as another bug completely unrelated to the threads in the vein of gdiplus ( http://bugs.winehq.org/show_bug.cgi?id=18142 ).

The bug I'm interested in for the purpose of this thread is the transparency issue displayed when Rainmeter loads a skin. Rainmeter (as well as Rainlendar2 when used in WINE), shows skins, but they lack a transparent background. This essentially kills the purpose of loading any skin in Rainmeter, as each individual skin displays a black box w/ it's surroundings (short of using a very squared skin, which presumably would work).

The purpose of this thread is to a.) bring attention to the bug, firstly, but b.) find out if the bug itself can actually be remedied through some workaround.

I am uncertain of the date of this particular bug's existence. I have searched for Rainmeter within the buglists, and short of the bugs I have personally filed, I have found nothing.

If there is anyone interested in this particular bug or the workability of Rainmeter within Linux under WINE, I'm interested in hearing from you.

[jay]

I can confirm the first bug. Haven't tried to reproduce the second one yet.

I used Rainmeter 0.14 for testing ( http://www.softpedia.com/progDownload/Rainmeter-Download-20289.html )
on
wine-1.1.20
on
compiz/xfwm4/metacity
on
Ubuntu 9.04, amd64 [2 cores]

Didn't get any really useful error messages either.

Code:

$ wine "C:\Program Files\Rainmeter\Rainmeter.exe"
fixme:win:EnumDisplayDevicesW ((null),0,0x32f544,0x00000000), stub!
fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION
fixme:dciman:DCICreatePrimary 0x328 0x75129c




It seems like the skin is shown correct first, but then the black corners appear.
Furthermore this bug only occures when the wine-version for Rainmeter.exe is set to a Windows 2000 or newer, with ME and below it seems to be displayed fine.

Furthermore I noticed that the cpu load is (dependent on the pretended windows OS type) either persistent 0% (when OS is some NT -> fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION ) or something like 1078919324% (no fixme: message for that) and permanently increasing.


"fixme:win:EnumDisplayDevicesW ((null),0,0x32f544,0x00000000), stub!" doesn't seem to show up on some win-versions btw.

Well, got other things to do now, but I hope I could give you some useful info or at least didn't waste your time too much.

Jay

ps: oh and btw... the application locked up when testing different version settings while winecfg was running, didn't happen in virtual desktop mode though.

[Caledfwlch]

Firstly, jay, thanks for replying.

I can confirm what Jay is saying to a tee. Using the Arch Linux latest snapshot of the repos, everything on my system is up-to-date.

Post by jay
It seems like the skin is shown correct first, but then the black corners appear.

I, too, notice this behavior. It's as if the skin 'wants' to appear correctly, but then fails after a half second or so.

Post by jay
Furthermore this bug only occures when the wine-version for Rainmeter.exe is set to a Windows 2000 or newer, with ME and below it seems to be displayed fine.

Setting Rainmeter to run in Win98, I can confirm this behavior. It's also notable that Rainmeter's menu "Transparency >" option is now grayed upon using Win98, forcing any used Rainmeter skin to 0% transparency.

It's also useful to note that, when using Win98, Rainmeter no longer bounces all over the screen when switching Compiz virtual desktops.

When moving the skin in Win98, you'll also note a reload of Rainmeter all together may need to occur due to the wallpaper background not changing.

Post by jay
I hope I could give you some useful info or at least didn't waste your time too much.

Believe me when I say your reply worked wonders. Thank you for trying!