April 2006 - Wine-devel - list.winehq.org

Re: MAPI32: class IProfAdmin, functions MAPIAdminProfiles & MAPIFindNext
by Dimi Paun 22 Apr '06

22 Apr '06

On Sat, April 22, 2006 9:10 am, Konstantin Petrov said: > changelog: > mapi.h : added comments to structure MapiFileDesc > mapix.h : added declaration of the interface (class) IProfAdmin > mapi32_main.c : implemented the function MAPIFindNext with FIXME message > created iprofadmin.c : MAPI Default IProfAdmin implementation. There are > implemented 12 functions of class IProfAdmin. Functions > IProfAdmin::AddRef, > IProfAdmin::QueryInterface and IProfAdmin::Release are realized, and other > functions return declared error. > iprofadmin.c : the function MAPIAdminProfiles implemented. > as consequence there were fixed Makefile.in and mapi32.spec Did you use a tab setting that is not 8 spaces, 'cause the indentation looks pretty weird. If you did, please reformat assuming an 8-space tab and resubmit. -- Dimi Paun <dimi(a)lattica.com> Lattica, Inc.

1 0

About the web forums and why it's just a waste of time
by Segin 22 Apr '06

22 Apr '06

This is a big of a nag, but mostly truth. Hear me out. Most of users' problem with the email is the lack of spiffy graphics. I don't understand why people rant and rave onver the email because it doesn't have spiffy graphics. Also, this extends to HTML formatting. They want to be able to use god awful formatting and colors and fonts that make your stomach churn, or at least hurt your eyes. Now that I did my mostly truthful rant, let's get to business: I have also heard about problems configuring their email client for certain features. Guess what? Forums require the exact same thing! If you want people to reply via email to a different address, you normally setup a reply-to field in your email client. This is usually done for better notification of replies. With forums, you register your email that you recieve notices with, then you have to set for every topic oyu start (or reply to) that you want email notification. This is expodentally more complex than email! While admin-set global stickies are something that only forums can do, that's basically the only thing going for them (and we could just instruct users to setup a sticky filter based on a string that only admins can put into subjects, save replies). User-set custom stickys - These are a pain to set with forums, that is if the forum software supports it, and 99% of forums software doesn't. Users like this, and switching to a forum can really piss them off with this. With an email client, you have more flexibility, cause depending on your client, you can setup these by user, subject, regex on the user or subject, if a attachment exists or not, body of message, size, time, date, who is in the 'cc' list (to hide those not sent directly to the list), and so on. The only forum software that I saw that did this would only do it if you selected a thread for a personal sticky -- this isn't really flexible compared to email.

1 0

SoC idea: get Sun's JRE working well
by Dan Kegel 22 Apr '06

22 Apr '06

A lot of applications rely on Sun's JRE, some even use it in their installers! (cf. http://bugs.winehq.org/show_bug.cgi?id=5040 ) Getting Sun's JRE working properly on Wine would be a great project. -- Wine for Windows ISVs: http://kegel.com/wine/isv

2 2

REALLY old backwards compatability and Wine
by Segin 22 Apr '06

22 Apr '06

This is a email that serves multiple purposes, and you all may find intersting. Wine, as we all know, is able to load and run native programs that the real McCoy can't run, but these programs are also close to 20 years old, so nobody really cares anyways. That's right, programs for Windows 2.0 I recently downloaded a full distrib (bin+src) of one such program. It is a xeyes clone with color. Orignally, it worked in Wine, showing a large window with a pair of eyes that followed the cursor, with a blue-green background (the part where the "eyes" aren't drawn in the window) This program was running in a window, and this appears to be broken, as under the real McCoy, version 3.1 (*gasp!*), it runs minimized, cannot be restored, and modifies it's own icon to preform the eyes deal (with a grey "background", as described above). I went to make sure that Wine wasn't broken, so I re-exec'd the binary, but only to get this: First chance exception: page fault on write access to 0x000f0400 in 32-bit code (0xb7df2e97). Register dump: CS:0073 SS:007b DS:007b ES:007b FS:003b GS:0033 EIP:b7df2e97 ESP:7fbff570 EBP:7fbff5b0 EFLAGS:00010246( - 00 -RIZP1) EAX:00000000 EBX:7eee7ff0 ECX:00000023 EDX:00000000 ESI:000f0400 EDI:000f0400 Stack dump: 0x7fbff570: 00000000 7eed0e49 000f0400 00000000 0x7fbff580: 0000008c 7ffd50f0 00010000 7ffb190b 0x7fbff590: 7ffdcc24 7fc94988 00000000 0000ffff 0x7fbff5a0: 0000f30f 7eee7ff0 00000000 00000001 0x7fbff5b0: 7fbff5bc 7eed158d 7eee7ff0 7fbff5dc 0x7fbff5c0: 7eee51c4 7eec0000 00000001 00000000 Backtrace: =>1 0xb7df2e97 memset+0x37 in libc.so.6 (0xb7df2e97) 2 0x7eed158d DllMain+0x22 in winedos (0x7eed158d) 3 0x7eee51c4 in winedos (+0x251c4) (0x7eee51c4) 4 0x7ffb0a2d call_dll_entry_point+0x15 in ntdll (0x7ffb0a2d) 5 0x7ffb11f2 in ntdll (+0x211f2) (0x7ffb11f2) 6 0x7ffb13c5 in ntdll (+0x213c5) (0x7ffb13c5) 7 0x7ffb2d8c LdrLoadDll+0x5e in ntdll (0x7ffb2d8c) 8 0x7fc6bab1 in kernel32 (+0x2bab1) (0x7fc6bab1) 9 0x7fc6bb36 LoadLibraryExW+0x4f in kernel32 (0x7fc6bb36) 10 0x7fc6bbe5 LoadLibraryExA+0x2d in kernel32 (0x7fc6bbe5) 11 0x7fc6bc0b LoadLibraryA+0x1b in kernel32 (0x7fc6bc0b) 12 0x7fc1f219 main+0xb7 in winevdm (0x7fc1f219) 13 0x7fc1fb0b in winevdm (+0xfb0b) (0x7fc1fb0b) 14 0x7fc730f3 in kernel32 (+0x330f3) (0x7fc730f3) 15 0xb7ee654b wine_switch_to_stack+0x17 in libwine.so.1 (0xb7ee654b) 0xb7df2e97 memset+0x37 in libc.so.6: repe stosl %es:(%edi) Now, I am no debugger expert, but it appears that Wine is trying to write to a NULL pointer using memset. (I could be reading the stack dump wrong, bear with me here) I hope this helps.

3 3

SoC idea: improve dsound/winmm
by Roderick Colenbrander 22 Apr '06

22 Apr '06

Hi, One of the things users complain about is wine's sound quality. Problems which you experience range from crackling sound in media players to buffer underruns and high latencies in games. For a part the problem is caused by crappy audio drivers and soundcards. None the less are there things in wine which can be improved to reduce the issues. I don't know this part of the code well but I believe the alsa driver is far from optimal and second the dsound contains some bugs too. In this SoC project a student would need to look into the causes of wine's bad sound quality and try to improve it by fixing winealsa and other parts. Regards, Roderick Colenbrander -- Echte DSL-Flatrate dauerhaft f�r 0,- Euro*! "Feel free" mit GMX DSL! http://www.gmx.net/de/go/dsl

7 7

Re-proposal: web forums
by Dan Kegel 22 Apr '06

22 Apr '06

Mike Hearn wrote: > [Let's set up web forums for wine users] I dislike the thought of having to monitor multiple forums. Worse still, if a particular web forum goes offline, all its content is lost; not so for mailing lists, which are archived in multiple locations and so are much more reliable and long-lived. Let's keep it all in wine-users, please. If your goal is to make sure that the average person can get answers to his or her questions about Wine, that's a noble enterprise. I would prefer a solution that doesn't have the reliability and usability problems of the current crop of web forums. I would also prefer an evidence-based approach to demonstrating the problem(s) with the current situation. "Some people prefer web forums" is not sufficient evidence. As I've said before, Mike is brilliant but misguided :-) - Dan -- Wine for Windows ISVs: http://kegel.com/wine/isv

5 6

undefined reference to `GetCharABCWidthsI'
by Gerald Pfeifer 22 Apr '06

22 Apr '06

After the following change to dlls/gdi/font.c revision 1.32 date: 2006-04-19 18:16:36 +0000; author: julliard; state: Exp; lines: +49 -0 Jeff Latimer <lats(a)yless4u.com.au> gdi: Added implementation of GetCharABCWidthsI. I now get the following build error on FreeBSD 5.4: ../../tools/winegcc/winegcc -B../../tools/winebuild -shared ./gdi32.spec dispdib.spec.o gdi.exe.spec.o wing.spec.o bidi16.o dispdib.o env.o gdi16.o metafile16.o wing.o bidi.o bitblt.o bitmap.o brush.o clipping.o dc.o dib.o driver.o enhmetafile.o enhmfdrv/bitblt.o enhmfdrv/dc.o enhmfdrv/graphics.o enhmfdrv/init.o enhmfdrv/mapping.o enhmfdrv/objects.o font.o freetype.o gdi_main.o gdiobj.o icm.o mapping.o metafile.o mfdrv/bitblt.o mfdrv/dc.o mfdrv/graphics.o mfdrv/init.o mfdrv/mapping.o mfdrv/objects.o mfdrv/text.o painting.o palette.o path.o pen.o printdrv.o region.o version.res -Wl,--rpath,\$ORIGIN/`../../tools/relpath /tmp/WINE/lib/wine /tmp/WINE/lib` -o gdi32.dll.so -L../../dlls -ladvapi32 -lkernel32 -lntdll -L../../libs -lwine -L../../libs/unicode -lwine_unicode -L../../libs/port -lwine_port font.o(.text+0x3f28): In function `GetCharABCWidthsI': /sw/test/wine/dlls/gdi/font.c:2390: undefined reference to `WineEngGetCharABCWidthsI' winegcc: /usr/bin/gcc failed. SUSE Linux 10.1 works fine. WOuld you mind having a look? Gerald

2 2

git & sharing patches (was: SoC idea: improve dsound/winmm)
by Molle Bestefich 22 Apr '06

22 Apr '06

Mike Hearn wrote: > Do the artifacts change or disappear if you use my thread priority patch? I'd like to try out your patch with a couple of apps where sound is shoddy. What's the correct GIT way to apply your patch (and still be able to pull changes from the Wine tree)? Apologies if it's a stupid question, but git --help seems to give a plethora of commands and I've absolutely no idea which are meant for common use and which are completely useless or will corrupt my local tree...

2 1

msvcrt: sharing violation caused by duplicated hande in _get_osfhandle
by qingdoa daoo 22 Apr '06

22 Apr '06

Hi, wine's implementation of _get_osfhandle is found to cause sharing violation in some apps. The author commented this piece of code with a block of FIXME warning about possible leakage. It turns out it's no just a matter of leakage. A simple test program ------------------------------------------------------------ #include "windows.h" #include "io.h" #include "fcntl.h" #include "share.h" int main() { int fd, bWritten; fd = _sopen("foo", _O_CREAT|_O_RDWR,_SH_DENYRW, 0); WriteFile((HANDLE)_get_osfhandle(fd), "bar", 3, &bWritten, NULL); _close(fd); fd = _open("foo", _O_RDONLY, 0); _write(1, fd==-1? "F\n" : "S\n", 2); } ------------------------------------------------------------ The second _open fails with builtin msvcrt, not with native one. The comment says that the duplication is to make an inheritable hande. With the recent change in msvcrt's io init procedure, that shouldn't be a problem anymore. --------------------------------- Blab-away for as little as 1¢/min. Make PC-to-Phone Calls using Yahoo! Messenger with Voice.

2 1

Re: kernel: fix determination of using locale
by Dmitry Timoshkov 22 Apr '06

22 Apr '06

"Sergei Butakov" <butsergej(a)yandex.ru> wrote: > This patch fix for me two bugs: > 1. I'm using locale > # export LANG="POSIX" > # export LC_CTYPE="ru_RU.CP1251" > # export LC_COLLATE="ru_RU.CP1251" > with which in many programs some Russian glyphs look wrong or as square. > 2. in many programs English/Russian glyphs are ugly. > > There are workaround by > # export LANG="ru_RU.CP1251" > but it not suit for me. Probably you just need to set LC_MESSAGES to en_US. -- Dmitry.

3 10