October 2001 - Wine-Devel - list.winehq.org

Re: wine/memory virtual.c
by Ove Kaaven Oct. 24, 2001

Oct. 24, 2001

On Tue, 23 Oct 2001, Alexandre Julliard wrote: > Modified files: > memory : virtual.c > > Log message: > Try to use read instead of mmap for files on removable media, so that > we don't need to keep the mapping handle open. OK, I see that you did this part too, which I hadn't submitted yet... but there's still one thing missing; the .exe file is still open inside the wineserver, still having two outstanding references: - main_exe_file in scheduler/process.c - process->exe.file in server/process.c Since main_exe_file is apparently never used again anyway after init_process_done, I've simply used the following hack to close the .exe file on the wineserver side, assuming the server file object only needs to exist to detect sharing violations, not for actually ever accessing the file again: --- server/process.c 2001/03/05 22:55:56 1.1.1.9 +++ server/process.c 2001/10/19 16:40:52 1.4 @@ -802,6 +802,8 @@ { if (process->exe.file) release_object( process->exe.file ); process->exe.file = file; + /* close the unix file, so the media the file was on can be changed */ + remove_select_user( (struct object *)file ); } generate_startup_debug_events( current->process, req->entry ); set_event( process->init_event ); Maybe you have a better way to do it?

1 0

Latest in the homeworld/wineoss saga.
by Simon Britnell Oct. 24, 2001

Oct. 24, 2001

I've attacked homeworld for a couple of hours with IDA free and none of the places that set the value of the NULL pointer are remotely close to a waveOutWrite call. I've run -debugmsg +wave with and without the synchronous wodWrite patch. The results are fairly large so I won't send them unless you feel they're particularly useful. To summarise: Both have 2 wodOpen calls. Both produce "Oooch buffer underrun" messages. The synchronous wodWrite version does only one wodWrite and a large number of Oooch messages. The asynchronous one produces one Oooch per wodWrite. Here's a more detailed description of the fault, followed by a couple of ideas on what's really going wrong: -- DESCRIPTION OF STARTUP -- The game starts with a 2D animated Sierra online logo followed by an animated relic logo and then drops into the game menus which are 2D graphics. mode. There are 3 distinct behaviours by the game: 1) Immediately prior to playing the Sierra logo, the game produces an exception report complaining of a page fault referencing 0x0000000. The animation starts and plays behind the exception dialog, but there is no sound. If the exception dialog is ignored, it disappears when the game menu is drawn. The menus can be navigated until either the graphics mode is changed or the actual game is started at which time wine hangs. top shows 30% of CPU taken by each of two wine processes and a wineserver. Tracing with remote gdb shows WaitForSingleObject in the stack traces, when they can be seen (Usually by running gdb as root). I usually have trouble getting the correct symbols loaded. 2) As per (1) above, but no exception dialog appears. 3) The game starts and runs with sound. There is no hang. While writing this, I had an idea. More to follow in a later post :) __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com

2 2

Lindows
by Uwe Bonnes Oct. 24, 2001

Oct. 24, 2001

Hallo, does anybody know more about "Lindows" than the sparse information they annouce on their webpage and the press releases? I don't think they reimplemented the win32 api, so I guess they use wine code. Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

2 1

datacache bug
by giorgian Oct. 24, 2001

Oct. 24, 2001

hi all, as i wrote few days ago, there's a problem with the DataCache_GetAdvise function in dlls/ole32/datacache.c: the call (this->sinkInterface)->lpVtbl->QueryInterface(this->sinkInterface, &IID_IAdviseSink, (void**)ppAdvSink); crashes since this->sinkInterface is NULL. executing the program with the --debugmsg +relay option i've seen that there are no calls to DataCache_SetAdvise before this call, the sinkInferface field is not initialized. but the program works well under M$window$. looking here and there in the source, i tried to replace this call with this->lpvtbl2->QueryInterface(this->lpvtbl2, &IID_IAdviseSink, (void**)ppAdvSink); i don't know why, but it works. ok, what should i do now to let the world know this? thanks -- Pietro Giorgianni, Icube s.r.l. (via Ridolfi, 15 - Pisa (Italy)) giorgian(a)icube.it

1 0

Re: PATCH: revert version change
by Francois Gouget Oct. 24, 2001

Oct. 24, 2001

On Tue, 23 Oct 2001, Marcus Meissner wrote: > Hi, > > There was following change to misc/version.c: > revision 1.44 > date: 2001/07/27 23:57:38; author: julliard; state: Exp; lines: +1 -1 > Bob Goodwin <goodie1(a)pacbell.net> > Fix the 16bit win95 version. > > This makes for instance InstallShield, that uses DOS3CALLs to check for win95, > report a Bad Windows Version. (Installer of Monkey Island 3) > > Reference wine-devel post is: > http://www.winehq.com/hypermail/wine-devel/2001/07/0085.html > > If there is a testprogram, I can still boot Windows 95... Ladislav Sladecek <lsla(a)post.cz> posted a program to check the 16bit version back in august: http://www.winehq.com/hypermail/wine-devel/2001/08/0275.html I think the revert is the right thing to do. And this is what I get on the Win95 computers I have access to. I also recently found that quite a few installers are not happy with 0x0304. I attached Ladislav's 16bit test application and also attached a application I wrote that dumps what the 32bit GetVersion and GetVersionEx return. Bob, how did you come up with 0x07000304? Was it from the Daggerfall application or from a windows computer? If the latter, then could you run the attached tests and report what the results? Here are the result of the tests I have done here: * Win95 (the original version): It does indeed return 0x5f03 in 16bit mode. The CSDVersion does not seem to match in 32bit mode: GetVersion: version = 0xc0000004 GetVersionInfoEx: size = 148 major = 4 minor = 0 build = 0x40003b6 platform = VER_PLATFORM_WIN32_WINDOWS CSDVersion= "" vs. "Win95" in Wine I don't know if this could cause problems. * Win98SE: Returns 0x5f03 in 16bit mode, like Wine. No problem in 16bit mode. In 32bit mode I get some discrepencies, probably because this is SE and not the original 98: GetVersion: version = 0xc0000a04 GetVersionInfoEx: size = 148 major = 4 minor = 10 build = 0x40a08ae vs. 0x40A07CE in Wine platform = VER_PLATFORM_WIN32_WINDOWS CSDVersion= " A " vs. "Win98" in Wine * NT4: Everything is fine. Tonight I hope to get results for Win2000 and WinXP. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ RFC 2549: ftp://ftp.isi.edu/in-notes/rfc2549.txt IP over Avian Carriers with Quality of Service

1 1

Re: usleep in winmm/wineoss audio
by Simon Britnell Oct. 24, 2001

Oct. 24, 2001

--- eric pouech <eric.pouech(a)wanadoo.fr> wrote: > after, rereading my old code, there's nothing wrong > (at first glance) here some invocations (like close, > reset) need to be synchronous > (the calling thread blocks until the rendering > thread is done with the message), some others (like > adding a wavehdr for play back) > are asynchronous, hence the difference I added a WaitForSingleObject late last PM as an experiment. The results were ... interesting. The first thing I noticed was that my problem attaching to the process went away, ie. winedbg started to work. The second thing that happened was that the problem I experience went from being intermittant to constant and now also invariably throws a page fault exception (in the same place it intermittantly threw an exception). Tracing back from that page fault is proving difficult as the page fault occurs with a value that's been stored on the heap at some prior point in the execution. Any tips for rapid location of where it's being stored? I suspect it's a null return from a wine system call. > so, back to square 1. can you at least trace the > program with > -debugmsg +wave (and still capture the hanging ?) Unfortunately, turning -debugmsg +wave on has so far made the problem go away. I'll have another couple of attempts this PM as its been quite intermittant. Also, it occurs to me to try -debugmsg +wave with the (broken) WaitForSingleObject patch I mentioned above. __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com

3 8

Re: Shlwapi functions _29, _30, _31
by Alexandre Julliard Oct. 24, 2001

Oct. 24, 2001

"Guy L. Albertelli" <galberte(a)neo.lrun.com> writes: > I assume that the differences may be due to revisions of the Unicode > standard, but that is an assumption only. Yes, there are some differences between our Unicode tables and the Windows ones. It shouldn't really matter, and if it does it's easy to fix. The important thing is to make sure everything uses the tables in libwine_unicode so that there is only one place to fix if necessary. -- Alexandre Julliard julliard(a)winehq.com

1 0

Shlwapi functions _29, _30, _31
by Guy L. Albertelli Oct. 24, 2001

Oct. 24, 2001

Alexandre, Thank you for simplifing these functions when you committed them. I thought you might want to know that they do act slightly differently from the versions in the native shlwapi.dll (version 5.50.4807.2300): _29 (iswspace) - extra TRUE for 0x1680, 0x202f _30 (iswblank) - missing TRUE for 0x3000 _31 (iswpunct) - missing TRUE for 0x00aa, 0x00b5, 0x00bb, 0x0192, 0x05ba - extra TRUE for 0x02df, 0x02ea-0x02ee, 0x0346-0x034e, 0x0361, 0x0488-0x0489, 0x058a, 0x0591-0x05a1, 0x05a3-0x05af, 0x05c4, 0x0653-0x0655 (+ probably many more - I stopped comparing here). I assume that the differences may be due to revisions of the Unicode standard, but that is an assumption only. Guy Albertelli <<galberte(a)neo.lrun.com>>

1 0

Re: The big winsock patch
by Alexandre Julliard Oct. 23, 2001

Oct. 23, 2001

Francois Gouget <fgouget(a)free.fr> writes: > Make winsock.h and winsock2.h independent from the Unix headers > Make them compatible with both the Unix C headers and the MSVCRT > headers > Ensure compatibility with the Unix headers via the USE_WS_PREFIX macro > Add WINE_NOWINSOCK: prevents winsock.h from being included > from windows.h when defined > Add ws2tcpip.h, move definitions to the right header I don't like the WS_API preprocessor abuse. It would be much better to declare function prototypes and typedefs explicitly, even if it requires a bit more typing. Your solution makes it hard to find out what gets defined when, and will break tools that extract information from header files. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: mediatype patch
by Alexandre Julliard Oct. 23, 2001

Oct. 23, 2001

Ove Kaaven <ovehk(a)ping.uio.no> writes: > Save a disk file's media type in the server object. OK for the storing part, but why do you return it in the get_file_info request? Shouldn't it rather be in something like get_mapping_info? -- Alexandre Julliard julliard(a)winehq.com

2 1