November 2007 - Wine-Devel - list.winehq.org

Patch Follow up

by Alistair Leslie-Hughes

Hi, Are there any reason they following patches where missed? Patch 1 of 2 - Return E_INVALIDARG on bad parameter Patch 2 of 2 - Enable Test for Bad Argument / Fix Memory leak clusapi: Corrected value in GetNodeClusterState Best regards Alistair Leslie-Hughes

Nov. 11, 2007

Re: %fs, %gs on AROS hosted

by Marcus Meissner

On Sat, Nov 10, 2007 at 12:52:24PM +0100, Staf Verhaegen wrote: > On Thu, 08 Nov 2007 09:08:29 +0100, Marcus Meissner > <marcus(a)jet.franken.de> wrote: > > >> But the use of %fs by wine does not seem to conflict in any way with a > >> possible usage of that segment register by the OS where wine is running > >> on. Or is there some OS specific code that is handling it ? > > > > No, it does not conflict. Your syscall / scheduler entry/exit points > > however need to take care to save/restore it correctly, otherwise > > you get interesting crashes ;) > > Already big thanks. > > Do you know where this code for these entry/exit points handling is > located for the different OSes in the wine source code? These are in the Operating System itself. Linux: arch/i386/kernel/entry.S > Does Windows/wine also have such entry/exit points so I could freely use > %fs,%gs in AROS hosted on Windows? See above. Ciao, Marcus

Nov. 11, 2007

Re: Bugzilla rights

by Jan Zerebecki

On Tue, Nov 06, 2007 at 09:09:14PM +0000, L. Rahyen wrote: > On Tuesday November 6 2007 19:45, Vitaliy Margolen wrote: > > Bugzilla admin, please disallow people to add/remove e-mail addresses > > to/from bugs unless they've been given that right. Or at least have rights > > to modify any aspect of the bug. > I think that we should remove the right to add/remove e-mail addresses ONLY > from user who abuse his/her rights (like Toby Brommerich). All other users > should have this right - at least the right to add others to CC list. Sorry, neither of these seem to be possible with bugzilla as it currently is. I suppose in this case it was just an error from which something was learned. If someone doesn't learn, we still can disable their account. Jan

Nov. 11, 2007

Re: What the hell is up with searching google for wine-devel and wine-patches posts?

by Jan Zerebecki

On Wed, Nov 07, 2007 at 01:36:01AM -0500, Steven Edwards wrote: > I used to just be able to > search as follows > > wine-patches Steven Edwards > > and see every patch I ever submitted to wine. For searching mailinglists I found http://search.gmane.org/ (e.g. use gmane.comp.emulators.wine.patches as group) very useful. Jan

Nov. 10, 2007

Re: Linux/Windows app connection, how?

by Jan Zerebecki

> This comes up from time to time, the solution is always compile a > winelib app with winegcc then use sockets or something to communicate. > In your server app you can use windows and linux code mixed together. If you don't need to use linux APIs in your windows application you can also make a normal windows executable with a windows compiler of your choice. So winegcc/winelib is not a requirement for this style of client/server solution. (You could use winsock to communicate over tcp with the linux application.) Jan

Nov. 10, 2007

Re: alsa->pulseaudio->alsa directsound hang

by Jan Zerebecki

I now tested wine (and mplayer) with pulseaudio through the alsa plugin and it worked just fine. I'm positively impressed by pulseaudio. The bug with alsa where it played random noise at the start because the buffer was not correctly cleared also doesn't appear with this setup. The winmm wave test hangs at some varying point (but does not with plain alsa), but I didn't investigate that, yet. On Tue, Oct 30, 2007 at 01:24:20AM -0700, Al Tobey wrote: > Did anybody actually write a patch to directly support pulseaudio or > was the earlier discussion devoid of code? Nobody showed up with any code. And AFAIK nobody who actually codes on wine requested a winepulseaudio. > My registry mods and .asoundrc for using the alsa/pulseaudio plugin > are attached. Getting pulseaudio is left as an exercise to the > reader ;) I didn't notice it the first time I read your mail, but you didn't set pcm.!default and thus you manually need to select that for alsa applications. Though you did that in the registry I suspect it didn't work and wine tried the default. Pulseaudio obviously blocks your soundcard as it requests a direct hw device without dmix from alsa. Jan

Nov. 10, 2007

Re: USER32: Patch 1 of 6: Fix returned value of LB_DIR

by James Hawkins

On 11/10/07, James Hawkins <truiken(a)gmail.com> wrote: > On 11/10/07, Alex Villacís Lasso <a_villacis(a)palosanto.com> wrote: > > Tests reveal that (on an unsorted listbox), LB_DIR returns the value of the > > highest index inserted, as if it had sorted the list of files before starting > > the insertion. This patch attempts to replicate this behavior. > > > > Changelog: > > * Make LB_DIR return an index value as if it had inserted filenames in sorted > > order. > > > > Tests? Tests that aren't in the test suite that should be to verify > the patch is correct? > I'll go ahead and eat this one :) -- James Hawkins

Nov. 10, 2007

Re: USER32: Patch 1 of 6: Fix returned value of LB_DIR

by James Hawkins

On 11/10/07, Alex Villacís Lasso <a_villacis(a)palosanto.com> wrote: > Tests reveal that (on an unsorted listbox), LB_DIR returns the value of the > highest index inserted, as if it had sorted the list of files before starting > the insertion. This patch attempts to replicate this behavior. > > Changelog: > * Make LB_DIR return an index value as if it had inserted filenames in sorted > order. > Tests? Tests that aren't in the test suite that should be to verify the patch is correct? -- James Hawkins

Nov. 10, 2007

How can we handle undcomented shell functionsAW, that MS declared now (WSTR version only)

by Detlef Riekenberg

Hi There are a some undocumented Functions in shell32, which have one entrypoint that use STR in win9x, but WSTR in NT: - dlls/shell32/undocshell.h - http://www.geocities.com/SiliconValley/4942/paths.html Some of them are now declared without A/W suffix (include/shobj.h from SDK 2003), but since win9x is no longer supported by MS, it's only for the Wide version. Examples: PathProcessCommend, PathResolve, PathQualify Question for include/shlobj.h: What is the correct way to declare such functions? In undocshell.h, we use functionAW with LPVOID for the first two, but for PathQualify, we use: // BOOL WINAPI PathQualifyA(LPCSTR path); BOOL WINAPI PathQualifyW(LPCWSTR path); #define PathQualify WINELIB_NAME_AW(PathQualify) BOOL WINAPI PathQualifyAW(LPCVOID path); // And PathResolve in exported by name (see shell32.spec) ############################################################# Next question for shell32.spec: Are we fixed to "ptr" for such functions? ("str", "wstr" and "long" are all used and need to be fixed) Is it possible / useful to add support for "awstr" (or a similar name)? What about support for UNICODE-Strings ("ustr")? -- By by ... Detlef

Nov. 10, 2007

Re: %fs, %gs on AROS hosted

by Staf Verhaegen

On Thu, 08 Nov 2007 09:08:29 +0100, Marcus Meissner <marcus(a)jet.franken.de> wrote: >> But the use of %fs by wine does not seem to conflict in any way with a >> possible usage of that segment register by the OS where wine is running >> on. Or is there some OS specific code that is handling it ? > > No, it does not conflict. Your syscall / scheduler entry/exit points > however need to take care to save/restore it correctly, otherwise > you get interesting crashes ;) Already big thanks. Do you know where this code for these entry/exit points handling is located for the different OSes in the wine source code? Does Windows/wine also have such entry/exit points so I could freely use %fs,%gs in AROS hosted on Windows? greets, Staf.

Nov. 10, 2007