November 2001 - Wine-devel - list.winehq.org

Re: Which is correct? listview or treeview?
by Gerard Patel 22 Nov '01

22 Nov '01

At 01:58 PM 21/11/2001 -0800, you wrote: >With regard to my problem that I was getting infinite redraws in my listview >following Alexandre's change to server/user/message.c: > >I eventually tracked it down to the fact our code used >SetRedraw(FALSE)/SetRedraw(TRUE) on a list control. (So the bug can now be >demonstrated easily on e.g. the MFC RowList sample with a two-line change). >Now I see that in the case of LISTVIEW_SetRedraw DefWindowProc is called and >then RedrawWindow. >In the TREEVIEW_SetRedraw case it simply sets a flag within the treeview's >data. > >Which is (more) correct? I have seen the same problem with the plain listbox I have an app (Eudora) that is using WM_SETREDRAW on a listbox and found that it was not possible to reproduce the app's behaviour without doing directly a RedrawWindow. Andi commented it was ugly and unefficient (IIRC) and Alexandre Julliard ignored the patch anyway. So I intended at some point to look at Listview and Treeview (with the nasty intent to prove that I am right and they are wrong of course :-)) If you find that doing an explicit RedrawWindow is the right way (TM) I am all for it. OTOH, I have myself - at least ;-( - one regression pending for this patch. The attached hack (on dlls/x11drv/winpos.c) fixes it, maybe with a lot of luck it could fix your problem too ? I'm not too hopeful as it does not seem related but you never know. Gerard --- winpos.c.orig Tue Nov 13 23:02:20 2001 +++ winpos.c Thu Nov 22 00:24:35 2001 @@ -1464,6 +1464,14 @@ /*********************************************************************** * X11DRV_ConfigureNotify */ + +Bool predicate(Display *display, XEvent *event, char *arg) +{ + Window w = *((Window *) arg); + return ((event->type == ConfigureNotify) && + (event->xany.window == w)); +} + void X11DRV_ConfigureNotify( HWND hwnd, XConfigureEvent *event ) { HWND oldInsertAfter; @@ -1472,8 +1480,20 @@ RECT rect; WINDOWPOS winpos; int x = event->x, y = event->y; + XConfigureEvent ev; if (!(win = WIN_GetPtr( hwnd ))) return; + + while (XCheckIfEvent(event->display, (XEvent *)&ev, + predicate, (XPointer)&event->window)) + { + event->x = ev.x; + event->y = ev.y; + event->width = ev.width; + event->height = ev.height; + event->above = ev.above; + event->send_event = ev.send_event; + } data = win->pDriverData; /* Get geometry */

2 1

thanksgiving prayer
by lee 22 Nov '01

22 Nov '01

...wishing everyone a special thanksgiving filled with great joy, special friends and loved ones as we move forward to a brighter tomorrow, carrying with us the best of yesterday and the promise of today- lee

1 0

Crashing menus, event handling broken?
by Jukka Heinonen 22 Nov '01

22 Nov '01

While playing Civilization 2, I noticed that Wine crashes if I click menu bar twice while the game is thinking. The reason for this is that raw hardware messages get processed before menu event loop starts. Raw event processing turns raw events into cooked events and because menu event loop has not started yet, game window is not capturing events and thus clicks on menu bar are turned into nonclient messages. These cooked messages are put back into message queue. When menu event loop starts, it receives the second nonclient button press event and passes it to DispatchMessage. Default window procedure creates a new menu event loop and since the nonclient message was still in the event queue (it would have been removed if DispatchMessage had returned), we get a nice stack overflow. The real problem is not in menu event loop but in the fact that Wine event handling forgets that we can't determine whether a mouse event is nonclient event or not unless we have already called window procedure for all previous mouse events. This is because one of those calls might have for example captured mouse input after which we no longer report nonclient events. Since event handling is quite complicated and this flaw looks more like a design error than coding bug, I would like to get some discussion about what should be done to fix this problem. -- Jukka Heinonen <http://www.iki.fi/jhei/>

1 0

Re: Packing in mmsystem16 and supporting broken GetWindowLong use cases
by Joshua Thielen 21 Nov '01

21 Nov '01

Jukka Heinonen wrote: >The second patch is because Wine mmsystem16 header does not use >correct packing which results into 48 byte structure taking >52 bytes and overwriting a pointer. This makes Civilization 2 >quite unhappy. Took almost a day to debug the game and >find out that the bugger who messes the pointer is really >waveOutGetDevCaps16... > Yessss! Thanks for finding this bug! It would have taken me ages to find it. Finale was choking awfully after calling midiInGetDevCaps. It would go into an endless loop, trashing my hard-drive and forcing me to do a hard reset on my box. Your patch fixes the problem. Cheers! Josh Thielen

1 0

Which is correct? listview or treeview?
by Bill Medland 21 Nov '01

21 Nov '01

With regard to my problem that I was getting infinite redraws in my listview following Alexandre's change to server/user/message.c: I eventually tracked it down to the fact our code used SetRedraw(FALSE)/SetRedraw(TRUE) on a list control. (So the bug can now be demonstrated easily on e.g. the MFC RowList sample with a two-line change). Now I see that in the case of LISTVIEW_SetRedraw DefWindowProc is called and then RedrawWindow. In the TREEVIEW_SetRedraw case it simply sets a flag within the treeview's data. Which is (more) correct? (And how does calling RedrawWindow once result in the repeated repainting of the treeview anyway?) I'm prepared to do the fix if no-one else can but I must admit that I am not really up on the details of the invalidation/paint etc Bill

1 0

New level of Shlwapi (parts 1 & 2)
by Guy L. Albertelli 21 Nov '01

21 Nov '01

Most of the shlwapi upgrades have been driven by my desire to be able to run Outlook Express under Wine. (I know that there are other mail packages, but the rest of the home uses Windows - anyway it is a great test vehicle, uses everything ever implemented). The current status is as follows: With V4 or IE and OE - with builtin shlwapi,commctrl,comctl32 . IE displays page, can even do www.cnn.com, even runs javascript. + Certain menu popups either fail to display, or bomb. + Needs work in comctl32 rebar for extra space. . OE seems to basicly work. + Can display messages, change folders, and create messages. + Cannot invoke address book, but can request address list when creating new message. V5 of OE - requires native commctrl, comctl32: . With builtin comctl32 + OE fails to startup . With native commctrl, comctl32: + OE comes up but is mostly non-functional. Bombs on shutdown due to shlwapi.208 and shlwapi.210 V5 of IE - Does not start up yet. Requires native shlwapi . With native shlwapi, builtin comctl32 + IE has black toolbars with white lines. Menu line real small. . With native shlwapi and comctl32: + IE has black bitmaps for button bar. Mostly functional, however cannot access dropdown arrow for address bar. <sarcasm> I really appreciate the MS developer that decided to implement menus in OE5 with objects and put some of his "common" code in shlwapi. </sarcasm> Guy Albertelli <<galberte(a)neo.lrun.com>>

1 0

Re: Improving DIB support for SHM
by Gavriel State 21 Nov '01

21 Nov '01

Doh! I guess we've been too lax about integrating some of our non D3D work back into WineHQ. We did this in WineX some time ago, but with a couple of differences: 1) We don't restrict using XShmGetImage to the whole-screen case and 2) We save and restore the XImage mask fields around the call to XShmGetImage, which seems to like to write over those fields to no useful purpose. IE: /* We must save and restore the bmpImage's masks in order * to preserve them across the call to XShmGetImage, which * decides to eliminate them since it doesn't happen to know * what the format of the image is supposed to be, even though * we do. */ saveRed = bmpImage->red_mask; saveBlue= bmpImage->blue_mask; saveGreen = bmpImage->green_mask; XShmGetImage( gdi_display, descr->drawable, bmpImage, descr->xSrc, descr->ySrc, AllPlanes); bmpImage->red_mask = saveRed; bmpImage->blue_mask = saveBlue; bmpImage->green_mask = saveGreen; We'll probably get around to doing some merging back into WineHQ after our next release. -Gav Jukka Heinonen wrote: > > The following patch halves the time it takes to move units > from one square to another in Civilization 2. If SHM is supported, > the patch uses XShmGetImage instead of XGetSubImage when filling > the whole XImage (which is probably way too careful since SHM is > much faster than the alternative...). Hopefully this > version is paranoid enough to work with older XFree86 servers > where XShmGetImage might change the underlying XImage format > into XYPixmap. > > Changelog: > Try to use XShmGetImage instead of XGetSubImage if SHM is supported. > Do a sanity check before using XShmPutImage. > > Index: wine/graphics/x11drv/dib.c > =================================================================== > RCS file: /home/wine/wine/graphics/x11drv/dib.c,v > retrieving revision 1.82 > diff -u -r1.82 dib.c > --- wine/graphics/x11drv/dib.c 2001/11/05 23:52:18 1.82 > +++ wine/graphics/x11drv/dib.c 2001/11/21 12:39:42 > @@ -4473,7 +4473,7 @@ > descr->xSrc, descr->ySrc, descr->xDest, descr->yDest, > descr->width, descr->height); > #ifdef HAVE_LIBXXSHM > - if (descr->useShm) > + if (descr->useShm && descr->image) > { > XShmPutImage( gdi_display, descr->drawable, descr->gc, bmpImage, > descr->xSrc, descr->ySrc, descr->xDest, descr->yDest, > @@ -4519,9 +4519,26 @@ > TRACE("XGetSubImage(%ld,%d,%d,%d,%d,%ld,%d,%p,%d,%d)\n", > descr->drawable, descr->xSrc, descr->ySrc, descr->width, > lines, AllPlanes, ZPixmap, bmpImage, descr->xDest, descr->yDest); > - XGetSubImage( gdi_display, descr->drawable, descr->xSrc, descr->ySrc, > - descr->width, lines, AllPlanes, ZPixmap, > - bmpImage, descr->xDest, descr->yDest ); > + > +#ifdef HAVE_LIBXXSHM > + if (descr->useShm && descr->image && > + descr->xDest == 0 && descr->yDest == 0 && > + descr->width == bmpImage->width && > + lines == bmpImage->height && > + XShmPixmapFormat( gdi_display ) == ZPixmap) > + { > + XShmGetImage( gdi_display, descr->drawable, bmpImage, > + descr->xSrc, descr->ySrc, AllPlanes ); > + if (bmpImage->format != ZPixmap) > + ERR( "XShmGetImage changed image format!\n" ); > + } > + else > +#endif > + { > + XGetSubImage( gdi_display, descr->drawable, descr->xSrc, descr->ySrc, > + descr->width, lines, AllPlanes, ZPixmap, > + bmpImage, descr->xDest, descr->yDest ); > + } > > TRACE("Dib: depth=%2d r=%lx g=%lx b=%lx\n", > descr->infoBpp,descr->rMask,descr->gMask,descr->bMask); > > -- > Jukka Heinonen <http://www.iki.fi/jhei/> -- Gavriel State, CEO TransGaming Technologies Inc. http://www.transgaming.com gav(a)transgaming.com

1 0

cleaning up pthreads implementation - ABI problem
by Peter Hunnisett 21 Nov '01

21 Nov '01

Howdy, I was just looking at cleaning up the pthreads implementation but I don't think that it's possible to have our pthreads implementation without providing binary compatibility with the pthread implementation we're hijacking. The reason is that the mutex and rwlock objects are supposed to be opaque to the programmer but in fact they are not opaque to the compiler. According to the standard, one doesn't have to call pthread_{mutex,rwlock}_init before using a {mutex,rwlock} as would be required with a totally opaque type. The standard allows for static initialization of the mutex and rwlock using PTHREAD_{MUTEX,RWLOCK}_INITIALIZER if you don't want any special attributes. libc does just this with its rwlock for gettext's global data. If you look at the ream of FIXME:pthread_rwlock_{rd,un}lock that sometimes come out starting wine you'll see that there's no call to pthread_rwlock_init. Of course, this stuff is safe because there's generally no writers wanting the pthread_rwlock_wrlock. So our current pthreads mutex implementation works in Linux by the skin of its teeth, becuase the first field in the mutex is reserved and initialized to 0 (in both the static and pthread_mutex_init cases), provided the version of LinuxThreads is new enough and a pointer is the same size as an int. Now I assume that this would mean it fails curiously for things that don't use LinuxThreads. This would also seem to mean, since we need binary compatibility, that things could even fail if someone upgrades their libc :( BTW, our check is presently basically only on __GLIBC__. Flavour me ignorant, and lazy, but I would assume that this not a Linux only symbol. So in fact the wine pthreads implementation might be already used incorrectly. Reader/Writer locks can be made to work the same way, as the mutex, since there are several fields which are initialized to NULL and we can then HeapAlloc our own structure off that field. We can then come up with weird and wonderful exceptions for the wonky cases that are sure to exist in some implementation of pthreads. Can anyone see a nice way out of this that I'm remaining ignorant to? I can't even think of a runtime way of detecting if our implementation is going to work. All in all I'm not too enthusiatic about the potential maintenance nightmare that this may provide but it would seem the only way is to have different mutex and rwlock structures depending on the pthreads implementation we're hijacking. Some code if people are more interested in details: >From glibc 2.2.4 /* Mutexes (not abstract because of PTHREAD_MUTEX_INITIALIZER). */ /* (The layout is unnatural to maintain binary compatibility with earlier releases of LinuxThreads.) */ typedef struct { int __m_reserved; /* Reserved for future use */ int __m_count; /* Depth of recursive locking */ _pthread_descr __m_owner; /* Owner thread (if recursive or errcheck) */ int __m_kind; /* Mutex kind: fast, recursive or errcheck */ struct _pthread_fastlock __m_lock; /* Underlying fast lock */ } pthread_mutex_t; #define PTHREAD_MUTEX_INITIALIZER \ {0, 0, 0, PTHREAD_MUTEX_TIMED_NP, __LOCK_INITIALIZER} What WINE uses (scheduler/pthread.c): /* assume that pthread_mutex_t has room for at least one pointer, * and hope that the users of pthread_mutex_t considers it opaque * (never checks what's in it) */ typedef struct { CRITICAL_SECTION *critsect; } *wine_mutex; rwlocks on the other hand are not so lucky as there's no unused bits: >From glibc-2.2.4 /* Read-write locks. */ typedef struct _pthread_rwlock_t { struct _pthread_fastlock __rw_lock; /* Lock to guarantee mutual exclusion */ int __rw_readers; /* Number of readers */ _pthread_descr __rw_writer; /* Identity of writer, or NULL if none */ _pthread_descr __rw_read_waiting; /* Threads waiting for reading */ _pthread_descr __rw_write_waiting; /* Threads waiting for writing */ int __rw_kind; /* Reader/Writer preference selection */ int __rw_pshared; /* Shared between processes or not */ } pthread_rwlock_t; # define PTHREAD_RWLOCK_INITIALIZER \ { __LOCK_INITIALIZER, 0, NULL, NULL, NULL, \ PTHREAD_RWLOCK_DEFAULT_NP, PTHREAD_PROCESS_PRIVATE } Ciao, Peter hunnise(a)yahoo.com __________________________________________________ Do You Yahoo!? Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1

2 1

Re: _dev_t for MSVCRT
by Marcus Meissner 21 Nov '01

21 Nov '01

On Tue, Nov 20, 2001 at 10:16:14PM +0100, Uwe Bonnes wrote: > Changelog: > include/msvcrt/sys/types: > _dev_t in a unsigned integer > -- > Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de > > Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt > --------- Tel. 06151 162516 -------- Fax. 06151 164321 ---------- > Index: wine/include/msvcrt/sys/types.h > =================================================================== > RCS file: /home/wine/wine/include/msvcrt/sys/types.h,v > retrieving revision 1.2 > diff -u -r1.2 types.h > --- wine/include/msvcrt/sys/types.h 2001/10/22 18:59:24 1.2 > +++ wine/include/msvcrt/sys/types.h 2001/11/20 19:44:21 > @@ -15,7 +15,7 @@ > #endif > > > -typedef unsigned short _dev_t; > +typedef unsigned _dev_t; Dann sollte da aber 'unsigned int' stehen ;) Ciao, Marcus

1 1

Re: (no subject) (msvcrt.getenv)
by Francois Gouget 20 Nov '01

20 Nov '01

On Tue, 20 Nov 2001, Uwe Bonnes wrote: > > Changelog: > dlls/msvcrt/environ.c: MSVCRT_getenv > Compare for the length of the key and return NULL in case of failure Good catch. Another thing to keep in mind: IIRC environment strings are not case sensitive on Windows. This means that normally getenv("PATH")==getenv("path") and that putenv("Path") would overwrite that same environment variable... This does not appear to be the case in the current code. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ Dieu dit: "M-x Lumière". Et la lumière fut.

1 0