- Wine-Devel - list.winehq.org

Re: PerformanceCounterFrequency fix
by Michael Ost Jan. 27, 2005

Jan. 27, 2005

On Tue, 2005-01-25 at 14:05, wine-devel-request(a)winehq.org wrote: > Message: 1 > Date: Tue, 25 Jan 2005 19:30:04 +0100 > From: Andreas Mohr <andi(a)rhlx01.fht-esslingen.de> > To: Rein Klazes <wijn(a)wanadoo.nl> > Cc: Lionel Ulmer <lionel.ulmer(a)free.fr>, wine-patches(a)winehq.com, > wine-devel(a)winehq.org > Subject: Re: PerformanceCounterFrequency fix. > > Hi, > > On Tue, Jan 25, 2005 at 06:44:04PM +0100, Rein Klazes wrote: > > On Mon, 24 Jan 2005 15:08:56 +0100, you wrote: > > > > > > How bad is it to use the gettimeofday() method? > > > > > > In my opinion, the RTDSC method should be suppressed from the code and we > > > should always use the 'gettimeofday' method (despite the penalty hit of a > > > syscall). > > > > I was more concerned about the accuracy of gettimeofday (not > > incrementing in usec's). So I did a small test and I find it behaves > > very nicely. > > > > That was the only reason I could see to justify the rdtsc method, so > > here it goes. As the cpuHz variable is not used anymore, we might as > > well move it to ntdll. > (sorry for not replying earlier - no time :-\) > I'm not sure why you'd want to base it on gettimeofday(). > This is a terrible idea IMHO. > I'm quite certain that many programs use that function for extremely time critical code > (games, anyone??), and that thus the Windows function is equally highly optimized, > certainly much less slow than a gettimeofday() call. > > This should remain based on rdtsc IMHO, or on equally suitable and fast methods > (ACPI counter, ...). > > Or did you actually test it with programs calling it a large number of times, > or test its performance behaviour on Windows? > > Andreas Mohr Our system uses the performance counter all the time in extremely time critical code. If the call is anything more than a few cycles with absolutely no chance of blocking, we are hosed! The application is an embedded audio plugin player. The audio is processed with SCHED_FIFO and needs to be as deterministic and fast as possible. I hope this fix/change doesn't jeopardize our product's use of Wine... mo =================================== Michael Ost, Software Architect Muse Research, Inc. most(a)museresearch.com

3 4

Re: shdocvw: properly implement DllCanUnloadNow
by jacek Jan. 26, 2005

Jan. 26, 2005

> HRESULT WINAPI SHDOCVW_DllCanUnloadNow(void) > { > - FIXME("(void): stub\n"); > - > - return S_FALSE; > + return SHDOCVW_refCount != 0 ? S_FALSE : S_OK; > } I think you should try to call Mozilla's DllCanUnloadNow before returning S_OK. Your patch may cause unload of mozilla while some instances of IWebBrowser still exist.

1 0

NPTL and Wine threads
by Luke Kenneth Casson Leighton Jan. 26, 2005

Jan. 26, 2005

it's come to my attention that NPTL cannot cope with jumping out of a cancellation handler. i thought you might like to be made aware of this in case you weren't already because this has been discovered to have an impact on the way that FreeDCE operates. there is some code in FreeDCE which expects to be able to jump out of a cancellation handler: i noticed in the "exception" comments that whilst it is not okay to do a "goto" from inside a TRY block, it says it's okay to do that in the EXCEPT block. you may wish to double-check this when running Wine under NPTL, and if it doesn't work please bitch like hell at the designers of NPTL for me because it's needed for FreeDCE as well. ta, l. -- -- <a href="http://lkcl.net">http://lkcl.net</a> --

1 0

more d3d9
by Oliver Stieber Jan. 26, 2005

Jan. 26, 2005

Hi, I've been tinkering for a week or so and D3d9 is almost their (well except for shaders) Along with Pirates I can now run Axis and Alies and Myst4 is looking promising Most of the demos from here http://www.codesampler.com/dx9src.htm Swap Chains only runs in one window, Shadow Volume is still a very flaky and some of the textures have strange offsets, but everything runs. Apart from the slightly bad texture offsets wine should be able to run more DirectX 9 applications than Cedega if only I could get them installed! I'm going to do some more debugging, cleaning up and post some screenshots on a web site and start sending patches against head (in smaller chunks that the last lot!) ___________________________________________________________ ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com

2 1

Re: Does wine have a posix threads "emulation" with which TRY/EXCEPT/CATCH will work?
by Luke Kenneth Casson Leighton Jan. 26, 2005

Jan. 26, 2005

thank you rob. i'm attaching some example code - it's from client/perf.c. after looking at these closely, i believe it would be straightforward to #define CATCH to __EXCEPT etc. and to rewrite the FreeDCE __EXCEPT argument as a function rather than a structure. dcethreads provides a CATCH macro - you are expected to provide a catcher for _one_ particular exception type. you're expected to roll these one after another: the macro expands to "} else if exc_matches(...) { " basically, i think the way that the Wine code does it, the CATCH possibilities are "rolled" into one EXCEPT - is that correct? or can you do more than one EXCEPT (even if the function says "i accept more than one type of exception - visa _or_ marstercaaard") so if the FreeDCE CATCH exceptions were turned into EXCEPT-style functions, on a "#ifdef / ./configure --with-wine-exceptions" basis, i think that'd work. this raises a further question, which i apologise for not being able to work out from looking at the macros. is it possible to do this: __TRY { /* do something horrible */ } __EXCEPT(exception_1) { printf("this is exception 1\n"); } __EXCEPT(exception_2) { printf("this is exception 1\n"); } ENDTRY secondary question: does Wine's exception handling cope with "stacked" exceptions and "fall-through"? if there are "nested" __TRY / __EXCEPT blocks, and an exception is not covered by the "inner" exception handlers, will it go through to the "outer" ones? ta rob. l. On Wed, Jan 26, 2005 at 03:11:36PM -0600, Robert Shearman wrote: > Luke Kenneth Casson Leighton wrote: > > >hi, > > > >i would like to check something which, if it is the case, > >may make porting FreeDCE to wine a no-brainer. > > > >i notice that there exists pthread emulation in Wine: it > >says so at the top of wine/dlls/kernel/pthread.c > > > >can i therefore expect the following things: > > > >1) the attached test program to compile and work under wine, > >with a few macros and/or simple wrapper functions e.g. redirect > >RAISE to RaiseException? > > > >2) a program that uses posix threads to also at the same time > >be able to use Wine Exception routines/macros, try/except/catch? > > > > > >... wouldn't it be ironic if Wine's pthreads emulation turned > >out to provide exactly the posix draft 4 threading semantics > >that FreeDCE programs expect? > > > >l. > > > > > > No, that program won't compile or run. > Windows exception handling (SEH) requires you to have an exception > filter function so that you don't get, say, a STATUS_ACCESS_VIOLATION > exception when you are expecting an RPC_E_DISCONNECT exception. > > You can see an example of Wine exception handling here: > http://cvs.winehq.org/cvsweb/wine/dlls/user/lstr.c?rev=1.31&content-type=te… > > Rob -- -- <a href="http://lkcl.net">http://lkcl.net</a> --

1 0

Does wine have a posix threads "emulation" with which TRY/EXCEPT/CATCH will work?
by Luke Kenneth Casson Leighton Jan. 26, 2005

Jan. 26, 2005

hi, i would like to check something which, if it is the case, may make porting FreeDCE to wine a no-brainer. i notice that there exists pthread emulation in Wine: it says so at the top of wine/dlls/kernel/pthread.c can i therefore expect the following things: 1) the attached test program to compile and work under wine, with a few macros and/or simple wrapper functions e.g. redirect RAISE to RaiseException? 2) a program that uses posix threads to also at the same time be able to use Wine Exception routines/macros, try/except/catch? ... wouldn't it be ironic if Wine's pthreads emulation turned out to provide exactly the posix draft 4 threading semantics that FreeDCE programs expect? l. -- -- <a href="http://lkcl.net">http://lkcl.net</a> --

2 1

Re: [OLE #50] IRemUnknown Proxy/Stub Implementation
by Vincent Béron Jan. 26, 2005

Jan. 26, 2005

Le mar 25/01/2005 à 11:51, Robert Shearman a écrit : > Hi, > > This time without alloca. > > Rob > > Changelog: > Proxies/stubs for IRemUnknown methods, based on code written by Ove Kåven. > ---- > > diff -u -p -N -r -x '*~' -x '.#*' -x CVS -x Makefile -x '*.o' -x '*.spec.def' dlls/ole32_49/oleproxy.c dlls/ole32/oleproxy.c > --- dlls/ole32_49/oleproxy.c 2005-01-24 16:12:03.000000000 -0600 > +++ dlls/ole32/oleproxy.c 2005-01-25 10:47:23.011171080 -0600 > @@ -2,6 +2,7 @@ > * OLE32 proxy/stub handler > * > * Copyright 2002 Marcus Meissner > + * Copyright 2001 Ove KÃ¥ven, TransGaming Technologies Not sure we want UTF-8 in the source files... At least we don't have it (yet). Vincent

4 5

Re: Disable antialiasing in palette mode (updated)
by Jakob Eriksson Jan. 26, 2005

Jan. 26, 2005

I'm just a peon but wouldn't it be better for the ReactOS project if you put the code in gdi instead? They have to solve it someplace else anyway, since they don't have x11drv. Glenn Wurster wrote: >This is an updated patch based on Alexandre's comments. It uses >GetObjectW instead of GDI_GetObjPtr. > >Description: > >According to the docs for CreateFont: > >"Windows 95 with Plus!, Windows 98/Me: The display must greater than >8-bit color, it must be a single plane device, it cannot be a palette >display, and it cannot be in a multiple display monitor setup. In >addition, you must select a TrueType font into a screen DC prior to >using it in a DIBSection, otherwise antialiasing does not happen." > >Windows 2000 has been tested and it also does not antialias fonts when >writing into an 8 bit DIB. > >This patch disables antialiasing when a palette is involved. You've >seen this before in previous incarnations. The docs specify to >disable antialiasing and I've got an application which depends on >antialiasing being disabled. I've tried putting the change in the gdi >code, as well as in the x11drv code. I think it looks cleaner in the >x11drv code (no change to function prototypes) so that's where this >attempts to put it (with the caching split off into a different patch >this time). > >Author: > * Glenn Wurster (gwurster(a)scs.carleton.ca) > >Changelog: > * Disable antialiasing of fonts while working in a palette. > >Index: dlls/x11drv/xrender.c >=================================================================== >RCS file: /home/wine/wine/dlls/x11drv/xrender.c,v >retrieving revision 1.56 >diff -u -r1.56 xrender.c >--- dlls/x11drv/xrender.c 24 Jan 2005 19:32:54 -0000 1.56 >+++ dlls/x11drv/xrender.c 26 Jan 2005 04:15:27 -0000 >@@ -1013,6 +1013,25 @@ > INT *deltas = NULL, char_extra; > HRGN saved_region = 0; > UINT align = GetTextAlign( hdc ); >+ BOOL disable_antialias = FALSE; >+ AA_Type antialias = AA_None; >+ >+ /* Do we need to disable antialiasing because of palette mode? */ >+ do { >+ HBITMAP hBitmap; >+ DIBSECTION bmp; >+ >+ hBitmap = GetCurrentObject( physDev->hdc, OBJ_BITMAP ); >+ if( GetObjectW( hBitmap, sizeof(bmp), &bmp ) != sizeof(bmp) ) { >+ TRACE("bitmap is not a DIB\n"); >+ break; >+ } >+ >+ if (bmp.dsBmih.biBitCount <= 8) { >+ TRACE("Disabling antialiasing\n"); >+ disable_antialias = TRUE; >+ } >+ } while (0); > > TRACE("%p, %d, %d, %08x, %p, %s, %d, %p)\n", hdc, x, y, flags, > lprect, debugstr_wn(wstr, count), count, lpDx); >@@ -1307,14 +1326,16 @@ > > EnterCriticalSection(&xrender_cs); > entry = glyphsetCache + physDev->xrender->cache_index; >- formatEntry = entry->format[entry->aa_default]; >+ if( disable_antialias == FALSE ) >+ antialias = entry->aa_default; >+ formatEntry = entry->format[antialias]; > > for(idx = 0; idx < count; idx++) { > if( !formatEntry ) { >- UploadGlyph(physDev, glyphs[idx], entry->aa_default); >- formatEntry = entry->format[entry->aa_default]; >+ UploadGlyph(physDev, glyphs[idx], antialias); >+ formatEntry = entry->format[antialias]; > } else if( glyphs[idx] >= formatEntry->nrealized || formatEntry->realized[glyphs[idx]] == FALSE) { >- UploadGlyph(physDev, glyphs[idx], entry->aa_default); >+ UploadGlyph(physDev, glyphs[idx], antialias); > } > } > assert(formatEntry); >@@ -1354,7 +1375,7 @@ > wine_tsx11_lock(); > XSetForeground( gdi_display, physDev->gc, textPixel ); > >- if(entry->aa_default == AA_None) { >+ if(antialias == AA_None) { > for(idx = 0; idx < count; idx++) { > SharpGlyphMono(physDev, physDev->org.x + x + xoff, > physDev->org.y + y + yoff, > >

2 1

Re: [winecfg] Show detected audio systems, and offer selection
by Rob Shearman Jan. 26, 2005

Jan. 26, 2005

Paul van Schayck wrote: >Hey, > >Never thought a simple popup with a few radio boxes would be this hard... > > Instead of generating the dialog box manually, why don't you store it in a resource? It will then be easier to localise and should be easier to use. You could also disable the radio boxes of drivers that cannot be used instead of hiding them and then you wouldn't have to do complicated rearranging. Rob

3 3

Q GUIDS
by Oliver Stieber Jan. 26, 2005

Jan. 26, 2005

Hi, I've had to generate some GUIDs for wined3d is there a wine policy on internal GUIDs or should the ones I've generated be ok. secondly, nothing to do with GUIDS, what's the 'best' thread safe IUnknown enumerate to use for storing a heap of unique elements that will be searched for by the IUnknown pointer as well as emerated. It would be usefull if the enumerate didn't reference count because I may also need to use a call back or sorts so that when the ref count of the IUnknown gets to zero it's release can tell the enumerate to remove it from the list. ___________________________________________________________ ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com

2 1