December 2003 - Wine-Devel - list.winehq.org

Delayed debug tracing
by Mike Hearn Dec. 13, 2003

Dec. 13, 2003

Hi, I'd like to implement delayed debug tracing, by which I mean you can do a +relay,+seh,+tid,+win,+otherstuff and yet not create a log file so huge it's impossible to work with. I seem to recall seeing this feature in WineX, where you can press Alt-F12 to switch it on. Does anybody who knows the debugging/keyboard code have any particular preferences as to how this should be implemented? thanks -mike

5 8

Corrupted Wine source package on sf
by Vincent Béron Dec. 13, 2003

Dec. 13, 2003

Hi Alexandre, The 20031212 source .tar.gz on sourceforge is shorter (~400k) than the one on ibiblio, and when untarred yields an error. You probably want to reupload it. Vincent

2 1

Internally maintained end-to-end API documentation
by Joshua Walker Dec. 13, 2003

Dec. 13, 2003

Hello all, I have been a lurker, going on now a good part of three years. I have some ideas I wish to cast to the table. I am a writer, not a programmer by craft. This hindrance aside, I have kicked around my fair share of C code. I have also been known to throw gooey splatterings of VMS report data into Postgres, and somehow rationalize it into something coherent on the other side for pay. After reading Wine Weekly News Issue #200, (Congrats by the way), I stumbled across the Wine Status Pages. I particularly like how everything is grouped by major components. With a quick glance at a color-coded graph, one can see how well something has been implemented. There is one thing that really bugging the heck out of me. Most of the API documentation is possessed by MSDN. In my humble opinion, this is a very dangerous and volatile place to have it. I was once told that using the internet as a data resource is like watching The Library of Alexandria burn over and over again. Having attempted to be a code developer for Microsoft products in the past, I know how quick MSDN can be when it comes to pulling the data out from under your feet. I also know what it's like to go hunting for some bizarre functionality, only to have MSDN look at you back with nothing but a blank stare on it's face. Sometimes it's the "little things" like the deep inner workings of interprocess communication. Better yet, why on earth is my stack is being messed with after a function call. I think it's about time to take the various Windows APIs and place them onto a fishbowl for safe keeping. Having lurked here for as long as I have, I have found two constants about announcing projects on wine-devel that I'd like to share. 1) In order to have Alexandre change direction, you better have a very good reason. 2) Many ideas have been announced here with no real clear goal or steps of execution. They soon die of atrophy shortly after. As for myself, I'm thinking to taking up the mantle for documenting the Windows API, or at least setting up public facilties for myself and others to use. The problem I have is how about will it be documented? First goal is to figure out where on earth to *put* the documentation.. I'm thinking of some kind of universal data holder where API information can aggregated, parsed, and rendered to one's choosing. As I shook my magic hate-ball I asked, "What do you think about storing the API data in some kind of XMLish-thingy format?" All signs pointed to yes. This is quite tragic as my knowledge of XML is really weak. As far as I know, it's that HTML superset, isn't it? Stepping back a bit, I started thinking of tossing an SQL database on one of my machines and writing some kind of dynamic HTML front end for it. The SQL idea is nice because I can write the front end to spit out what I have in any data format I can dream of, including XML. It also allows people to come to my website, pop in an API function that wan't there, and move on with their lives. That seems plausible. The second hurtle is what to put as the actual fields. I'm really only used to seeing function rendered an a "Unix man style" format. Name, synopsis, description, return value, errors, bugs. See "man fork" for example. We will also need kind of a DLL overview, (man stdio?). What about something a little more UML compatible. Anyone have any ideas? Admittedly this is where my coherence of the project ends. There might be someone out there a little more apt at content than I am. Let me close with a story; When I was 12 my mother bought a brand new washing machine. My mother gave me the old one and a set of wrenches, and then proceeded to tell me to go nuts. In the span of five hours there was washing machine all over our front lawn. I learned more about about washing machines at the age of 12 by tearing that thing apart more than most adults know now. It's kinda what I want to do with Windows. -Joshua

2 1

More deadlocks + backtrace
by Shachar Shemesh Dec. 13, 2003

Dec. 13, 2003

I ran "spyxx" from MSVC on my deadlocking program, and everything was ok. I then quite the program itself (Microsoft Digital Image Pro was run with +relay, spy was run without any debug values). I left spy running for a while (~2 days), as I had other things to do, and I wanted the window handles it collected. At some point, the console that launched spy began showing the following: $ wine SPYXX err:font:ReadFontDir Can't open directory "/win/windows/fonts" fixme:mdi:MDIRefreshMenu partially function stub fixme:mdi:MDIRefreshMenu partially function stub err:toolbar:TOOLBAR_NotifyFormat wrong response to WM_NOTIFYFORMAT (0), assuming ANSI err:win:CreateWindowExA bad class name "msctls_updown" err:ntdll:RtlpWaitForCriticalSection section 0x40380d68 "?" wait timed out in thread 0011, blocked by 0000, retrying (60 sec) err:ntdll:RtlpWaitForCriticalSection section 0x40380d68 "?" wait timed out in thread 0011, blocked by 0000, retrying (60 sec) err:ntdll:RtlpWaitForCriticalSection section 0x40380d68 "?" wait timed out in thread 0011, blocked by 0000, retrying (60 sec) err:ntdll:RtlpWaitForCriticalSection section 0x40380d68 "?" wait timed out in thread 0011, blocked by 0000, retrying (60 sec) err:ntdll:RtlpWaitForCriticalSection section 0x40380d68 "?" wait timed out in thread 0011, blocked by 0000, retrying (60 sec) Here are the backtraces: Wine-dbg>walk process pid threads parent executable >00000010 2 00000000 'Y:\sources\Transfer\RemoteDebug\SPYXX.EXE' Wine-dbg>walk thread process tid prio 00000010 (D) Y:\sources\Transfer\RemoteDebug\SPYXX.EXE 00000012 0 <== 00000011 0 Wine-dbg>bt 0x11 Backtrace: =>0 0x4010e8ab (NTDLL.DLL._memccpy+0x5202b in libc.so.6) (ebp=40032c60) 1 0x401f4002 (NTDLL_wait_for_multiple_objects+0xbf(count=0x0, handles=0x0, flags=0x8, timeout=0x40032d10) [sync.c:578] in NTDLL.DLL) (ebp=40032cf8) 2 0x401f1edd (usr1_handler+0x4e(__signal=0xa, __context=0x8f) [signal_i386.c:1123] in NTDLL.DLL) (ebp=40032d1c) 3 0x4006e498 (NTDLL.DLL.toupper+0x6208 in libc.so.6) (ebp=4080f7e0) 4 0x401f4002 (NTDLL_wait_for_multiple_objects+0xbf(count=0x1, handles=0x4080f8cc, flags=0xc, timeout=0x4080f900) [sync.c:578] in NTDLL.DLL) (ebp=4080f878) 5 0x401f40ae (NTDLL.DLL.NtWaitForMultipleObjects+0x74 in NTDLL.DLL) (ebp=4080f8a0) 6 0x401f40fe (NtWaitForSingleObject+0x42(handle=0x88, alertable=0x0, timeout=0x4080f900) [sync.c:611] in NTDLL.DLL) (ebp=4080f8c4) 7 0x401cda35 (RtlpWaitForCriticalSection+0x127(crit=0x40380d68) [critsection.c:193] in NTDLL.DLL) (ebp=4080f964) 8 0x401cdc25 (RtlEnterCriticalSection+0x51(crit=0x40380d68) [critsection.c:255] in NTDLL.DLL) (ebp=4080f97c) 9 0x404f2555 (wine_pthread_mutex_lock+0x34(mutex=0x101593c8) [pthread.c:182] in KERNEL32.DLL) (ebp=4080f98c) 10 0x0011273a (__pthread_mutex_lock+0x25(mutex=0x101593c8) [kthread.c:588] in wine-kthread) (ebp=4080f99c) 11 0x40f409ea (_end+0x75bb6 in libX11.so.6) (ebp=4080f9bc) 12 0x40f0d799 (_end+0x42965 in libX11.so.6) (ebp=4080fa1c) 13 0x40e8b39e (X11DRV_InitKeyboard+0x84(key_state_table=0x40b66880) [keyboard.c:1332] in X11DRV.DLL) (ebp=4080fadc) 14 0x40e8c1c4 (X11DRV_MappingNotify+0x35(event=0x4080fb94) [keyboard.c:1635] in X11DRV.DLL) (ebp=4080fafc) 15 0x40e83647 (.L164+0xb [event.c] in X11DRV.DLL) (ebp=4080fb7c) 16 0x40e82e65 (process_events+0x71(data=0x40380d40) [event.c:143] in X11DRV.DLL) (ebp=4080fbf8) 17 0x40e82ffa (X11DRV_MsgWaitForMultipleObjectsEx+0x17b(count=0x1, handles=0x40380c6c, timeout=0xffffffff, mask=0x0, flags=0x0) [event.c:184] in X11DRV.DLL) (ebp=4080fd38) 18 0x40ad5f96 (GetMessageW+0x287(msg=0x46becc, hwnd=0x0, first=0x0, last=0x0) [message.c:2363] in USER32.DLL) (ebp=4080fdf4) 19 0x40ad6020 (GetMessageA+0x33(msg=0x46becc, hwnd=0x0, first=0x0, last=0x0) [message.c:2380] in USER32.DLL) (ebp=4080fe14) 20 0x6c3be6e9 (MFC42.DLL.5307+0x13 in MFC42.DLL) (ebp=0046becc) 21 0x0000000f (ebp=000100d6) Wine-dbg>bt 0x12 Backtrace: =>0 0x4010e8ab (NTDLL.DLL._memccpy+0x5202b in libc.so.6) (ebp=4030fc60) 1 0x401f4002 (NTDLL_wait_for_multiple_objects+0xbf(count=0x0, handles=0x0, flags=0x8, timeout=0x4030fd10) [sync.c:578] in NTDLL.DLL) (ebp=4030fcf8) 2 0x401f1edd (usr1_handler+0x4e(__signal=0xa, __context=0x8f) [signal_i386.c:1123] in NTDLL.DLL) (ebp=4030fd1c) 3 0x4006e498 (NTDLL.DLL.toupper+0x6208 in libc.so.6) (ebp=419eb450) 4 0x401f4002 (NTDLL_wait_for_multiple_objects+0xbf(count=0x1, handles=0x419eb53c, flags=0xc, timeout=0x419eb530) [sync.c:578] in NTDLL.DLL) (ebp=419eb4e8) 5 0x401f40ae (NTDLL.DLL.NtWaitForMultipleObjects+0x74 in NTDLL.DLL) (ebp=419eb510) 6 0x404fcb20 (WaitForMultipleObjectsEx+0x1d5(count=0x1, handles=0x419eb670, wait_all=0x0, timeout=0x32, alertable=0x0) [sync.c:167] in KERNEL32.DLL) (ebp=419eb648) 7 0x404fc8b8 (WaitForSingleObject+0x3c(handle=0x64, timeout=0x32) [sync.c:95] in KERNEL32.DLL) (ebp=419eb668) 8 0x004232fa (SPYXX.EXE..text+0x222fa in SPYXX.EXE) (ebp=40aab4df) 9 0x0168ec81 (SPYXX.EXE..rsrc+0x1221c81 in wine-kthread) (ebp=53e58955) *** Invalid address 0x53e58955 (_end+0x121f453d) I hope this helps someone. -- Shachar Shemesh Open Source integration & consulting Home page & resume - http://www.shemesh.biz/

2 3

Added support for inter-process GetWindowLong on the window extra bytes
by Phil Krylov Dec. 13, 2003

Dec. 13, 2003

Hello, This change prevents IBM Translation Manager from working correctly. It has a dialog named "Export Folder" which used to contain three buttons and a tabbed notebook with various options. After this update, the pages of the tabbed notebook are no longer displayed (and created). Hope I can provide some more details like a debug log if someone will tell me which channels should be traced. AJ> ChangeSet ID: 10222 AJ> CVSROOT: /opt/cvs-commit AJ> Module name: wine AJ> Changes by: julliard(a)wine.codeweavers.com 2003/11/26 16:15:41 AJ> Modified files: AJ> windows : win.c AJ> server : window.c trace.c protocol.def AJ> include/wine : server_protocol.h AJ> Log message: AJ> Added support for inter-process GetWindowLong on the window extra AJ> bytes. AJ> Patch: http://cvs.winehq.com/patch.py?id=10222 AJ> Old revision New revision Changes Path AJ> 1.226 1.227 +132 -111 wine/windows/win.c AJ> 1.25 1.26 +36 -4 wine/server/window.c AJ> 1.185 1.186 +7 -3 wine/server/trace.c AJ> 1.87 1.88 +6 -0 wine/server/protocol.def AJ> 1.88 1.89 +7 -1 AJ> wine/include/wine/server_protocol.h -- Best regards, Phil mailto:phil(a)newstar.rinet.ru

2 3

Core DLL's Status Section
by Tom Dec. 13, 2003

Dec. 13, 2003

Hi, Here is my first draft anyone have any comments?, constructive criticism? You will see I have wineps, x11drv, winedos listed on this page now so ill remove those sections from the Core Status page. If everything looks good ill paste the three sections together and send Brian a copy for WWN inclusion. :) And while he is doing that ill add in the remaining doc links. and then it should be ready for wine-patches. I've also got a nice new list of todo's to add to that page as well. Happy Holidays, Tom

1 0

Re: gdi: AlphaBlend constants
by Alexandre Julliard Dec. 12, 2003

Dec. 12, 2003

Kirill Smelkov <kirr(a)landau.phys.spbu.ru> writes: > +/* AlphaBlend flags */ > +#define AC_SRC_NO_PREMULT_ALPHA 0x01 > +#define AC_SRC_NO_ALPHA 0x02 > +#define AC_DST_NO_PREMULT_ALPHA 0x10 > +#define AC_DST_NO_ALPHA 0x20 Where do these come from? I don't see them in the SDK headers or on MSDN, there's only an AC_SRC_ALPHA. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: ntdll / kernel32: #38
by Dmitry Timoshkov Dec. 12, 2003

Dec. 12, 2003

"Eric Pouech" <pouech-eric(a)wanadoo.fr> wrote: > +HANDLE WINAPI OpenConsoleW(LPCWSTR name, DWORD access, LPSECURITY_ATTRIBUTES sa, > + DWORD creation) > +{ > + static const WCHAR coninW[] = {'C','O','N','I','N','$',0}; > + static const WCHAR conoutW[] = {'C','O','N','O','U','T','$',0}; > + BOOL output; > + HANDLE ret; > + > + if (strcmpW(coninW, name) == 0) > + output = FALSE; > + else if (strcmpW(conoutW, name) == 0) > + output = TRUE; While you are moving the code around then you are probably reviewing it as well? So, the question is: shouldn't the above comparisons be case insensitive? -- Dmitry.

2 1

Re: menu: Fix for menu tracking problem version 2
by Alexandre Julliard Dec. 12, 2003

Dec. 12, 2003

Andrew de Quincey <adq_dvb(a)lidskialf.net> writes: > I wrote a program in windows which had owner drawn menu items. Whenever I > received the WM_DRAWITEM message: > > 1) I called GetCapture(), and compared it with the hwnd of main window > containing the menu. It pops up a message box if they are different. > 2) I then called SetCapture() with the HWND of the menu item itself. > > This approximately simulates what IDA is doing. > > The message box never popped up. This shows that even though the capture was > changed in one WM_DRAWITEM, windows had reset the capture to the window > containing the menu by the time of the next one. I also checked the messages > _were_ being sent using Spy++ (in case of a programming error). That's a good start, but it's not enough to know what Windows is really doing. You should also check the capture window right after you call SetCapture; and you should use GetGUIThreadInfo and print out all the details, especially the GUI_INMENUMODE flag. My suspicion is that the capture is not allowed to change at all in menu mode. -- Alexandre Julliard julliard(a)winehq.com

2 5

Problems with transparent windows
by Juan Antonio Boscá Lloret Dec. 12, 2003

Dec. 12, 2003

Hello, I've found a probable bug in Wine when running win32 applications. Wine does not repaint windows areas that are hidden by another window over them (I think this is a behaviour by design). This behaviour is not correct when the upper window is transparent (WS_EX_TRANSPARENT style). Is it a bug? Thank you. --------------------------------------------------------------- Juan Antonio Boscá Lloret Departamento Programación CAE, S.A. jabosca(a)edicurs.com ---------------------------------------------------------------

2 1