wine-devel

wine-devel@winehq.org

84463 discussions

debugger startup
by eric pouech 07 Apr '01

07 Apr '01

someone (IIRC Gérard) had issues when starting the debugger when an exception occured. the setup was that the AeDebug registry key pointed to a shell script, which in turn, launched the real debugger (used to set up different context options, and such) however, what happens is: - the main program (the one generating the exception) tries to launch the debugger using the AeDebug - the process creation forks, and exec:s the debugger program - at the unix level, the debugged program starts, which runs the shell script, which then starts the debugger - at the windows level, since the shell script isn't a Wine process, the call fails (even if an Unix process is created) - the exception handler then reports the error in process creation and stops the main program, whereas the debugger has been started for this program either we decide that shell scripts shouldn't be used (easy solution) or we could do something like : - if the CreateProcess fails, then wait for a certain amount of time. If after that time the current process isn't currently debugged, then report the error. Otherwise assume what did happen is what has been described above comments are welcome A+ -- --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) "The future will be better tomorrow", Vice President Dan Quayle

1 0

Re: [PATCH] deferred trace
by Alexandre Julliard 07 Apr '01

07 Apr '01

eric pouech <eric.pouech(a)wanadoo.fr> writes: > using the Wine debugger (or any program using the debugging API) may > have some drawbacks. As the API is designed, it's not possible to > get control of the debugged program between its creation and its > entry point entering this makes very difficult to let the Wine > debugger trigger the expected features by poking the memory in the > debuggee. That's not exactly correct, the debugger can intercept the dll loads and break in the dll init code. Since with dll separation all Wine inits will be done inside dlls, the debugger should be able to intercept everything from the point we connect to the wineserver. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: Nasty Wine/kernel bug, anyone?
by Ulrich Weigand 06 Apr '01

06 Apr '01

> If I were going to make more WAGs (Wild Assed guesses), > I'd hazard a guess that this was due to my Alsa sound system > (running on the AC97 codec on my Asus motherboard). > > I tried to walk the stack back, but didn't get very far > (after the first time of wrongly computing the hex offset > from a .s file, I gave off, stomped off to look > for a better disassembler and gave up). If you have the 'vmlinux' file, try objdump --disassemble vmlinux > My best guess is that we're coming through dentry_open > (xc012e7fe)/ chrdev_open (0xc012f66c) > /soundcore_open (0xc01bbf71). However, I wasn't > seeing a familiar EBP/EIP pattern, so I'm possibly just > pulling random symbols out of the air. Yes, this appears to the proper trace: c0108d7f system_call c012ea2c sys_open c012e732 filp_open c012e7fe dentry_open c012f66c chrdev_open c01bbf71 soundcore_open d0826a42 <module> d087ce3e <module> c0107aa4 __down_failed c010794f __down ........ schedule [I am assuming that you have 256MB main memory. Correct?] Now, the question is still, who is sleeping on that semaphore and why? >From the call trace, it would seem that some sound-related module is loaded, which does a down() on some semaphore in its open() routine. We even know which semaphore: cff267ac. (This does not help much, as this is would appear to be somewhere in kmalloc()ed memory ...) We can continue to track in two directions. First, we can try to locate the point within the Wine application where the open() syscall is performed. >From the trace we see that the system call instruction is located at 40303a24, and the stack pointer points to 405c6638. The system call itself is an open( 41610240, O_WRONLY|O_NDELAY ) [The stack pointer value is rather weird, I'd say ...] This particular open mode occurs in the Wine source only in dlls/winmm/wineoss/audio.c, which confirms the sound hypothesis. The next time you reproduce the bug, could you please add the output of /proc/<pid>/maps for the hanging Wine process? This way, we could further analyse these pointers to Wine user space ... The more promising way to track the problem, however, is to find out which module those d* addresses above reside in. If the modules are still loaded, the output of /proc/ksyms should at least give a hint. For a more refined symbol lookup, it is best to use the 'ksymoops' tool, which correlates the /proc/ksyms output with the ELF symbols read from the object modules themselves. To do so, write this: Call Trace: [<d0826a42>] [<d087ce3e>] to a file and pipe the file through ksymoops. In any case, this would seem to be clearly a bug in that sound module. I don't know, however, why only Wine would trigger that bug ... Bye, Ulrich -- Dr. Ulrich Weigand weigand(a)informatik.uni-erlangen.de

2 1

Re: [PATCH] Name Thread Exception support in Wine debugger
by eric pouech 06 Apr '01

06 Apr '01

James Hatheway wrote: > > Hi guys, > > As I mentioned in an earlier email to the Devel mailing list, I discovered the > application I am supporting started raising more exceptions than usual. These > turned out to be an undocumented signal to the MS VC6 debugger to name the > current thread. You can read more about this at: > http://www.vcdj.com/upload/free/features/vcdj/2001/03mar01/et0103/et0103.asp > or searching google.com or deja.com for "0x406D1388". This patch fixes > the WINE debugger to handle this exception, just like MS VC6 debugger does. something seems wrong in your patch... the szName field from the exception field is never copied from the debugged program space to the debugger space (only the address is) so, you should use ReadProcessMemory instead (something like ReadProcessMemory(pThread->name, sizeof(pThread->name), DEBUG_CurrThread->process->handle, pThreadName->szName); ) also, I don't like the fact that you put the #define in a public header file and the exception struct definition inside the file it would make more sense to me if both were either public or private (I'd go for private, since those values don't seem to be defined in Windows' standard headers) A+ -- --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) "The future will be better tomorrow", Vice President Dan Quayle

2 1

WWN
by eric pouech 06 Apr '01

06 Apr '01

for personal reasons, I will no longer be able to write the WWN (and I haven't in the last couple of weeks) so, if someone wishes to take the job over, I'd be happy to help him (her) to set it up A+ -- --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) "The future will be better tomorrow", Vice President Dan Quayle

1 0

Re: [PATCH] deferred trace
by gerard patel 06 Apr '01

06 Apr '01

At 03:00 PM 14/03/2001 -0800, you wrote: <snip> >that the wine debugger can resolve, it is possible to meet the needs that >have been expressed. Expressed ? What I wanted is something short and simple, so I could help people on the news group. The use of my patch could be explained in *one* line and it would have worked with Wine right 'out of the box' :-)... something the internal debugger does not do anymore since a long time. I am afraid that this thing will be more complex to setup and to use that the debugger - something already too complex for many people. And well, I almost never use the debugger myself because it crashes at the slightest provocation so I don't feel much about saying people to use it. So in short, after all my 'needs' (helping ppl on the news group) are indeed specific :-) Gerard

2 1

Re: Make PostScript driver less loquacious
by Dmitry Timoshkov 04 Apr '01

04 Apr '01

"Ian Pilcher" <ian.pilcher(a)home.com> wrote: > dlls/wineps/init.c, ppd.c > > Change some PS driver FIXMEs/MESSAGEs to WARNs I think it's not the time to do it. Those messages/fixme's help a lot to add missing page and bin type into psdrv: they are not complete now. And without message about missing fonts in ppd/afm it would be nightmare to track down problems with ps fonts at least for me.

3 8

Newbie question...
by Elan Feingold 04 Apr '01

04 Apr '01

Hi, I have a 3rd party x86 DLL file, that I would like to use under Linux. I have read the documentation on the specmaker tool, which would seem to provide some help. The thing that makes my case a bit different is that I would like to use the DLL within a Linux native program, not a WineLib program (its just a command line application and I need it to be very lightweight). Im pretty sure the Wine code can help me out, but Im not sure what the easiest way to proceed is, and what parts I need to use. Thanks in advance, -elan

2 1

T@x2001: Linuxport via Wine
by Joerg Mayer 04 Apr '01

04 Apr '01

Just found the following at the SuSE portal pages (German only): http://portal.suse.de/de/content/reports/tax2001.html Rough translation on the introductory paragraph: The program t@x by Buhl Data has been around for the Windows platform for some years. This year, the current version T@x2001 has been made available for the Linux OS for a first time. In order to do this the Windows emulator Wine was used. Wine allows using Windows applications running Linux. This emulator has been modified to work with t@x2001. The summary as far as Wine is concerned: The program should be more stable and faster. Does anyone know which modifications were made to wine? Ciao Jörg -- Joerg Mayer <jmayer(a)loplof.de> I found out that "pro" means "instead of" (as in proconsul). Now I know what proactive means.

1 0

Re: Interesting Magic # exception understood by the MS debugger...
by Alexandre Julliard 03 Apr '01

03 Apr '01

"James Hatheway" <james(a)macadamian.com> writes: > OK, I don't know too much about the internals of the WINE debugger, so > here goes. Is there a field for this already, or should I make a new > field in DBG_THREAD structure? (I think that seems like the right place > for it) Yep, you can add a field there. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: Interesting Magic # exception understood by the MS debugger...
by Alexandre Julliard 03 Apr '01

03 Apr '01

"James Hatheway" <james(a)macadamian.com> writes: > For now, I put the little hack I attached to this mail in my tree to > save my sanity. My question to the list is, (A) is there a way to name > threads in Linux/Unix (I don't think so, but of course I could be wrong) No, and we don't need that. > and (B) should we support MS's hack somehow in the official tree? Yes, in the Wine debugger, which should store the name and simply continue execution, just like the MSVC debugger does. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: Speeding up wineserver syncronization objects with shared memory
by David Howells 03 Apr '01

03 Apr '01

Okay, I've thought about how to make the wine kernel module easier to use for Wine, and I've started adding another API to it - one that interprets Wine server messages. When I designed the module initially, I decided to make it possible to add as many API's as I liked, so in theory it's not too difficult. I'm currently performing a major rearrangement of the files that comprise the module: * Splitting the current win32 API bits out of the same files that contain the actual object class implementations. * Putting the core and the API's in separate subdirectories. * Abstracting name support, so that the internal API takes an "object name" which the user API is responsible for extracting from userspace and converting from WCS/MBS as necessary. * Fixing introduced bugs. This does not mean that I'm getting rid of the Win32 API I've already got! It will just be optional. Calling the new API will be by means of either: union generic_request msg; syscall(__NR_win32,WINESERVER_WINE_SERVER_MESSAGE,&msg); Or: union generic_request msg; int fd = open("/proc/wineserver"); struct WineserverIoctl wioc = { WINESERVER_WINE_SERVER_MESSAGE, &msg }; ioctl(fd,WINESERVER_IOCTL,&wioc); David

1 0

Interesting Magic # exception understood by the MS debugger...
by James Hatheway 02 Apr '01

02 Apr '01

Hi guys, I encountered something interesting while debugging a newer build of my application. I noticed that all of a sudden I was receiving approximately 10 new exceptions in the debugger on startup that I wasn't receiving before. After digging a little, I found out that the app coders were making use of a 'hidden' feature of MS VC that allowed you to name your threads (thus allowing for easier debugging). To do this, you call RaiseException with the code of 0x406D1388. (Search on deja.com or google.com for "0x406D1388" for details.) The windows debugger seems to be on the lookout for this code. Of course, what it does in wine is just breaking into the debugger many times and getting on my nerves. :-) For now, I put the little hack I attached to this mail in my tree to save my sanity. My question to the list is, (A) is there a way to name threads in Linux/Unix (I don't think so, but of course I could be wrong) and (B) should we support MS's hack somehow in the official tree? -James -- James Hatheway Software Designer - Macadamian Technologies, Inc. james(a)macadamian.com ~ http://www.macadamian.com "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier" "One could be happy if the air were as pure as the beer"

1 0

Re: RegSetValueEx problem?
by davep 02 Apr '01

02 Apr '01

On Sun, 01 Apr 2001, you wrote: > At 01:59 PM 31/03/2001 +0100, you wrote: > > <snip> > >Compiling and running Winemine under Windows 95 this problem doesn't occur. > >Win95 truncates the value at the first null. > > Under NT4 SP6, if I write (for example) '47' + 0 + 13 bytes of garbage, I get > exactly that when querying with the same buffer length (16). > If I use a longer buffer, one binary 0 is appended. > If I use a shorter buffer, nothing is copied and the function does not succeed > (ERROR_MORE_DATA). > > Of course under Regedit the garbage is not displayed; regedit displays only '47' > in this case. Ah, I didn't know Regedit was not giving the true value. That complicates matters even more. It seems I'll have to do some experiments. -- Dave Pickles

1 0

Re: event flags using pipes
by Alexandre Julliard 01 Apr '01

01 Apr '01

Mike McCormack <mike_mccormack(a)looksmart.com.au> writes: > Why? By changing the client side implementation of SetEvent and > ResetEvent the use file descriptors and implementing a client side > unix fd/handle cache, it should be possible to eliminate server calls > for these operations and improve the performance of wineserver. I don't see how you can get the right semantics this way, but maybe I missed something. How will you be able to wake a single thread for auto-reset events, but all of them for manual events? How are you going to implement PulseEvent? How can you do WaitForMultipleObjects for an auto-reset event without a race? -- Alexandre Julliard julliard(a)winehq.com

1 0

event flags using pipes
by Mike McCormack 01 Apr '01

01 Apr '01

Hi, i've written an implementation of windows event flags using pipes. (The modifications are confined to server/event.c for the moment.) Why? By changing the client side implementation of SetEvent and ResetEvent the use file descriptors and implementing a client side unix fd/handle cache, it should be possible to eliminate server calls for these operations and improve the performance of wineserver. Anyway, any comments are welcome. Mike ------------------------------------------ mailto:Mike_McCormack@looksmart.com.au ph +82 16 430 0425 __________________________________________________________________ Get your own free web email at http://www.looksmart.com.au LookSmart Australia

1 0

RegSetValueEx problem?
by davep 31 Mar '01

31 Mar '01

I've been trying to track down why WineMine does not save my best scores ;-( On exit, WineMine saves the current user settings to the registry with code like this: char data[16]; wsprintf( data, "%d", p_board->pos.x ); RegSetValueEx( hkey, "Xpos", 0, REG_SZ, (LPBYTE) data, sizeof( data )); The problem here is that the 'count' parameter passed to RegSetValueEx is not the string length but the size of the entire buffer. Wine dutifully writes into the registry the entire buffer including the trailing garbage, and also adds a null to the end if there isn't one there. When WineMine tries to read the value back it is now 17 characters long and too big to fit in the 16-byte buffer. Compiling and running Winemine under Windows 95 this problem doesn't occur. Win95 truncates the value at the first null. So RegSetValueEx needs fixing. However there are already comments in there about the differences between 95 and NT in the handling of REG_SZ values, and I don't have an NT system to carry out checks. Can someone with experience propose a solution? -- Dave Pickles

1 0

Sendmessage fix..
by crazney 31 Mar '01

31 Mar '01

Hey, im writing cause i presently use Andreas Mohr's sendmessage fix (titled TEB fix, because it stop the TEB errors).. basically, what the patch does, is clean up after a killed SendMessage if it didnt clean itsself up.. (well something like that, iirc).. anyway, i run the LHL howto page, and this patch is very useful to running Half-Life and several other games/programs.. so itd be nice to have it in CVS.. which it isnt presently.. when Andreas Mohr first wrote the patch, he told me that alexandre denied it, and neither him, ove or myself can work out why.. so yer, i'd just like to know a) why was it refused.. and b) can you put it in now, or write something which does just this now.. thanks in advance ps, can you please email me about this, because im not subscribed to this list. cheers David -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ David Hammerton aka crazney Victoria, Australia http://crazney.net/ HOWTO maintainer and Linux Half-Life http://lhl.linuxgames.com/ email: http://crazney.net/contact.html -> Note: this email address expires after one week.

1 0

Re: [PATCH] HOOK_CallHook
by David Elliott 30 Mar '01

30 Mar '01

Francois Jacques wrote: > Group, This patch comes from investigation of a bug that was observed in > RedHat 7.0 and not observed in RedHat 6.1. After a debugging session that > involved Stephane and I, we found out that it was a compiler issue (which > I feared from the start, but wasn't considering it seriously... > arghhh). The RedHat 6.1 installation has gcc version 2.91.66 and the > RedHat 7.0 installation has gcc version 2.96-20000731 (RedHat Linux 7.0) > with latest updates from RedHat. What's happening : 1) HOOK_CallHook is > being called (notice wParam and lParam arguments)2) Local variables > wParamOrig and lParamOrig are set to wParam and lParam values3) Mapping > function is called (MapFunc) with the *addresses* of wParam and lParam so > those get changed before we call the hook4) After MapFunc is called, > wParamOrig and lParamOrig values also changed to the new values of wParam > and lParam (!) Explanation : Down to the assembly level code, the compiler > doesn't create wParamOrig and lParamOrig on the stack. Instead, it > directly use wParam and lParam. Fix : add volatile to wParamOrig and > lParamOrig to prevent optimizations on those variables (aka. force their > presence on the stack) This kind of begs the question: how many more places in Wine should the volatile keyword be used? I find it hard to believe that this is the only spot in a large very low-level project like Wine. -Dave

6 9

Re: [PATCH] HOOK_CallHook
by François Jacques 30 Mar '01

30 Mar '01

>OK, I must have misunderstood you then. Sorry ... It's ok :) >I've managed to reduce the problem to a simple testcase, which is >appended below. It fails on RedHat gcc-2.96-69 (as of 12/2000) ... > >Culprit *appears* to be the global common subexpression elimination, >although this is not quite definite. I haven't tracked the problem >down yet. Are you thinking of doing more investigation on that? >However, when switching off GCSE (-fno-gcse), correct code is produced. >I'm not sure I'd like to do a complete Wine build without GCSE though; >this might significantly reduce code quality ... Oh the runtime behavior should not change that much but it's still non-negligeable. I like the "disable-gcse-if-you-probe-that-gcc-version" approach. >I'll write a configure check to switch off gcse, using the testcase >below. We should probably contact RedHat (I think Jacub Jelinek >handles 2.96 bugs ...) to confirm that GCSE is really at fault. If you want to post it with the test program, go ahead :) I never posted to gcc bug list so I'm not sure how much attention would be given to the bug. I think the mailing list should be cc'ed on the postings so everybody can follow up how it turns out. Ciao! Francois

1 0

Re: [PATCH] HOOK_CallHook
by François Jacques 30 Mar '01

30 Mar '01

>To be honest, I didn't read the patch and simply assumed this was a simple case >of forgetting to add volatile and getting bitten by a compiler that optimizes >more agressively. > >Either way it still is a valid point to be concerned about not only compiler >bugs but also new compiler features. 100% agree on this point. The community needs some ways to keep up to date with those new compiler versions. I know it's hard since we work on different projects which usually needs some kind of stability (in the product and in the development process) so we tend to stick with a reference compiler that we know works. However, if we don't keep up to date with those new compiler version, well fall in some intensive bug chasing session... and debugging at the assembly level is not that fun. And it's a cycle, we test the new compiler version, we find bugs, we get an improved compiler which, with a nice side-effect, may solve other bugs. >That would be a good idea. Of course I don't think the volatile keyword in >this case is hurting anything except maybe speed (negligble). Hopefully this >can get fixed soon in gcc. Looking at the gcc project page, it seems that 3.0 >is very close to be done. 3.0 ? That's good news in terms of technology advance... however I fear some late nites chasing bugs generated from that new version :) Cheers! Francois

1 0

Re: controls/edit.c
by lawson_whitney＠juno.com 29 Mar '01

29 Mar '01

On Thu, 29 Mar 2001, Ulrich Czekalla wrote: > I forgot to plug a region leak before I sent the patch in. I have > attached an updated patch. > > /Ulrich > Good day! This message comes to you from Juno-2.0.11 without benefit of junopine: I am composing it with Juno's GUI composer using the new edit.c. Much less flicker. I still can't answer a letter this way: Juno chokes on a SendMessageCallback as soon as it gets anything in its inbox, but that is nothing to do with edit.c It is hot and heavy with ads now and I get some sluggishness because I am contending with the ad engine for the 47 bogoMIPS of the mighty p-120, but really it seems much better. I still don't like GUI though, and Juno might get annoyed if I delete all its ads, which I would have to do to really use this composer, so I will go back to Pine now, but it does seem much better nor it was. Lawson ---cut here Well, not quite without junopine. I hit alt-s meaning to send it, and saved it to the draft folder instead, and tripped on another SendMessageCallback trying to get it out, but I did compose it with a GUI editor. The first patch moved the cursor for each character of the password, but showed no character, but with the revised patch I see the xxxx strikover I am accustomed to see. I think it is an improvement, but I am hardly a GUI guru. First soldier into the minefield, maybe, and still alive. Regards, Lawson ---cut here. Really. "The Rest is Spam." ________________________________________________________________ GET INTERNET ACCESS FROM JUNO! Juno offers FREE or PREMIUM Internet access for less! Join Juno today! For your FREE software, visit: http://dl.www.juno.com/get/tagj.

1 0

Re: controls/edit.c
by Ulrich Czekalla 29 Mar '01

29 Mar '01

I forgot to plug a region leak before I sent the patch in. I have attached an updated patch. /Ulrich Ulrich Czekalla wrote: > > This patch modifies the edit control to keep the line format data > and simply update it as changes occur. It also uses the line data > to calculate a proper update region. The result is a significant > reduction in flicker. Since this is a big change to the format > engine of the edit control it should be tested for a while before > committing. > > Changelog: > Ulrich Czekalla <uczekalla(a)codeweavers.com> > Update format engine to reduce flicker > Index: controls/edit.c =================================================================== RCS file: /home/wine/wine/controls/edit.c,v retrieving revision 1.75 diff -u -w -r1.75 edit.c --- controls/edit.c 2001/03/13 23:31:08 1.75 +++ controls/edit.c 2001/03/29 17:31:36 @@ -63,6 +63,7 @@ INT net_length; /* netto length of a line in visible characters */ LINE_END ending; INT width; /* width of the line in pixels */ + INT index; /* line index into the buffer */ struct tagLINEDEF *next; } LINEDEF; @@ -172,7 +173,7 @@ /* * Helper functions only valid for one type of control */ -static void EDIT_BuildLineDefs_ML(WND *wnd, EDITSTATE *es); +static void EDIT_BuildLineDefs_ML(WND *wnd, EDITSTATE *es, INT iStart, INT iEnd, INT delta, HRGN hrgn); static void EDIT_CalcLineWidth_SL(WND *wnd, EDITSTATE *es); static LPWSTR EDIT_GetPasswordPointer_SL(EDITSTATE *es); static void EDIT_MoveDown_ML(WND *wnd, EDITSTATE *es, BOOL extend); @@ -263,6 +264,7 @@ static void EDIT_WM_Timer(WND *wnd, EDITSTATE *es); static LRESULT EDIT_WM_VScroll(WND *wnd, EDITSTATE *es, INT action, INT pos); static void EDIT_UpdateText(WND *wnd, LPRECT rc, BOOL bErase); +static void EDIT_UpdateTextRegion(WND *wnd, HRGN hrgn, BOOL bErase); LRESULT WINAPI EditWndProcA(HWND hWnd, UINT uMsg, WPARAM wParam, LPARAM lParam); LRESULT WINAPI EditWndProcW(HWND hWnd, UINT uMsg, WPARAM wParam, LPARAM lParam); @@ -1132,100 +1134,250 @@ * a soft return '\r\r\n' or a hard return '\r\n' * */ -static void EDIT_BuildLineDefs_ML(WND *wnd, EDITSTATE *es) +static void EDIT_BuildLineDefs_ML(WND *wnd, EDITSTATE *es, INT istart, INT iend, INT delta, HRGN hrgn) { HDC dc; HFONT old_font = 0; - LPWSTR start, cp; + LPWSTR current_position, cp; INT fw; - LINEDEF *current_def; - LINEDEF **previous_next; + LINEDEF *current_line; + LINEDEF *previous_line; + LINEDEF *start_line; + INT line_index = 0, nstart_line = 0, nstart_index = 0; + INT line_count = es->line_count; + INT orig_net_length; + RECT rc; - current_def = es->first_line_def; - do { - LINEDEF *next_def = current_def->next; - HeapFree(GetProcessHeap(), 0, current_def); - current_def = next_def; - } while (current_def); - es->line_count = 0; - es->text_width = 0; + if (istart == iend && delta == 0) + return; dc = GetDC(wnd->hwndSelf); if (es->font) old_font = SelectObject(dc, es->font); + previous_line = NULL; + current_line = es->first_line_def; + + /* Find starting line. istart must lie inside an existing line or + * at the end of buffer */ + do { + if (istart < current_line->index + current_line->length || + current_line->ending == END_0) + break; + + previous_line = current_line; + current_line = current_line->next; + line_index++; + } while (current_line); + + if (!current_line) /* Error occurred start is not inside previous buffer */ + { + FIXME(" modification occurred outside buffer\n"); + return; + } + + /* Remember start of modifications in order to calculate update region */ + nstart_line = line_index; + nstart_index = current_line->index; + + /* We must start to reformat from the previous line since the modifications + * may have caused the line to wrap upwards. */ + if (!(es->style & ES_AUTOHSCROLL) && line_index > 0) + { + line_index--; + current_line = previous_line; + } + start_line = current_line; + fw = es->format_rect.right - es->format_rect.left; - start = es->text; - previous_next = &es->first_line_def; + current_position = es->text + current_line->index; do { - current_def = HeapAlloc(GetProcessHeap(), 0, sizeof(LINEDEF)); - current_def->next = NULL; - cp = start; + if (current_line != start_line) + { + if (!current_line || current_line->index + delta > current_position - es->text) + { + /* The buffer has been expanded, create a new line and + insert it into the link list */ + LINEDEF *new_line = HeapAlloc(GetProcessHeap(), 0, sizeof(LINEDEF)); + new_line->next = previous_line->next; + previous_line->next = new_line; + current_line = new_line; + es->line_count++; + } + else if (current_line->index + delta < current_position - es->text) + { + /* The previous line merged with this line so we delete this extra entry */ + previous_line->next = current_line->next; + HeapFree(GetProcessHeap(), 0, current_line); + current_line = previous_line->next; + es->line_count--; + continue; + } + else /* current_line->index + delta == current_position */ + { + if (current_position - es->text > iend) + break; /* We reached end of line modifications */ + /* else recalulate this line */ + } + } + + current_line->index = current_position - es->text; + orig_net_length = current_line->net_length; + + /* Find end of line */ + cp = current_position; while (*cp) { if ((*cp == '\r') && (*(cp + 1) == '\n')) break; cp++; } + + /* Mark type of line termination */ if (!(*cp)) { - current_def->ending = END_0; - current_def->net_length = strlenW(start); - } else if ((cp > start) && (*(cp - 1) == '\r')) { - current_def->ending = END_SOFT; - current_def->net_length = cp - start - 1; + current_line->ending = END_0; + current_line->net_length = strlenW(current_position); + } else if ((cp > current_position) && (*(cp - 1) == '\r')) { + current_line->ending = END_SOFT; + current_line->net_length = cp - current_position - 1; } else { - current_def->ending = END_HARD; - current_def->net_length = cp - start; + current_line->ending = END_HARD; + current_line->net_length = cp - current_position; } - current_def->width = (INT)LOWORD(GetTabbedTextExtentW(dc, - start, current_def->net_length, + + /* Calculate line width */ + current_line->width = (INT)LOWORD(GetTabbedTextExtentW(dc, + current_position, current_line->net_length, es->tabs_count, es->tabs)); + /* FIXME: check here for lines that are too wide even in AUTOHSCROLL (> 32767 ???) */ - if ((!(es->style & ES_AUTOHSCROLL)) && (current_def->width > fw)) { + if ((!(es->style & ES_AUTOHSCROLL)) && (current_line->width > fw)) { INT next = 0; INT prev; do { prev = next; - next = EDIT_CallWordBreakProc(es, start - es->text, - prev + 1, current_def->net_length, WB_RIGHT); - current_def->width = (INT)LOWORD(GetTabbedTextExtentW(dc, - start, next, es->tabs_count, es->tabs)); - } while (current_def->width <= fw); - if (!prev) { + next = EDIT_CallWordBreakProc(es, current_position - es->text, + prev + 1, current_line->net_length, WB_RIGHT); + current_line->width = (INT)LOWORD(GetTabbedTextExtentW(dc, + current_position, next, es->tabs_count, es->tabs)); + } while (current_line->width <= fw); + if (!prev) { /* Didn't find a line break so force a break */ next = 0; do { prev = next; next++; - current_def->width = (INT)LOWORD(GetTabbedTextExtentW(dc, - start, next, es->tabs_count, es->tabs)); - } while (current_def->width <= fw); + current_line->width = (INT)LOWORD(GetTabbedTextExtentW(dc, + current_position, next, es->tabs_count, es->tabs)); + } while (current_line->width <= fw); if (!prev) prev = 1; } - current_def->net_length = prev; - current_def->ending = END_WRAP; - current_def->width = (INT)LOWORD(GetTabbedTextExtentW(dc, start, - current_def->net_length, es->tabs_count, es->tabs)); + + /* If the first line we are calculating, wrapped before istart, we must + * adjust istart in order for this to be reflected in the update region. */ + if (current_line->index == nstart_index && istart > current_line->index + prev) + istart = current_line->index + prev; + /* else if we are updating the previous line before the first line we + * are re-caulculating and it expanded */ + else if (current_line == start_line && + current_line->index != nstart_index && orig_net_length < prev) + { + /* Line expanded due to an upwards line wrap so we must partially include + * previous line in update region */ + nstart_line = line_index; + nstart_index = current_line->index; + istart = current_line->index + orig_net_length; + } + + current_line->net_length = prev; + current_line->ending = END_WRAP; + current_line->width = (INT)LOWORD(GetTabbedTextExtentW(dc, current_position, + current_line->net_length, es->tabs_count, es->tabs)); } - switch (current_def->ending) { + + + /* Adjust length to include line termination */ + switch (current_line->ending) { case END_SOFT: - current_def->length = current_def->net_length + 3; + current_line->length = current_line->net_length + 3; break; case END_HARD: - current_def->length = current_def->net_length + 2; + current_line->length = current_line->net_length + 2; break; case END_WRAP: case END_0: - current_def->length = current_def->net_length; + current_line->length = current_line->net_length; break; } - es->text_width = max(es->text_width, current_def->width); - start += current_def->length; - *previous_next = current_def; - previous_next = &current_def->next; - es->line_count++; - } while (current_def->ending != END_0); + es->text_width = max(es->text_width, current_line->width); + current_position += current_line->length; + previous_line = current_line; + current_line = current_line->next; + line_index++; + } while (previous_line->ending != END_0); + + /* Finish adjusting line index's by delta or remove hanging lines */ + if (previous_line->ending == END_0) + { + LINEDEF *pnext = NULL; + + previous_line->next = NULL; + while (current_line) + { + pnext = current_line->next; + HeapFree(GetProcessHeap(), 0, current_line); + current_line = pnext; + es->line_count--; + } + } + else + { + while (current_line) + { + current_line->index += delta; + current_line = current_line->next; + } + } + + /* Calculate rest of modification rectangle */ + if (hrgn) + { + HRGN tmphrgn; + /* + * We calculate two rectangles. One for the first line which may have + * an indent with respect to the format rect. The other is a format-width + * rectangle that spans the rest of the lines that changed or moved. + */ + rc.top = es->format_rect.top + nstart_line * es->line_height - + (es->y_offset * es->line_height); /* Adjust for vertical scrollbar */ + rc.bottom = rc.top + es->line_height; + rc.left = es->format_rect.left + (INT)LOWORD(GetTabbedTextExtentW(dc, + es->text + nstart_index, istart - nstart_index, + es->tabs_count, es->tabs)) - es->x_offset; /* Adjust for horz scroll */ + rc.right = es->format_rect.right; + SetRectRgn(hrgn, rc.left, rc.top, rc.right, rc.bottom); + + rc.top = rc.bottom; + rc.left = es->format_rect.left; + rc.right = es->format_rect.right; + /* + * If lines were added or removed we must re-paint the remainder of the + * lines since the remaining lines were either shifted up or down. + */ + if (line_count < es->line_count) /* We added lines */ + rc.bottom = es->line_count * es->line_height; + else if (line_count > es->line_count) /* We removed lines */ + rc.bottom = line_count * es->line_height; + else + rc.bottom = line_index * es->line_height; + rc.bottom -= (es->y_offset * es->line_height); /* Adjust for vertical scrollbar */ + tmphrgn = CreateRectRgn(rc.left, rc.top, rc.right, rc.bottom); + CombineRgn(hrgn, hrgn, tmphrgn, RGN_OR); + DeleteObject(tmphrgn); + } + if (es->font) SelectObject(dc, old_font); + ReleaseDC(wnd->hwndSelf, dc); } @@ -2140,7 +2292,7 @@ es->format_rect.bottom = es->format_rect.top + es->line_height; if ((es->style & ES_MULTILINE) && !(es->style & ES_AUTOHSCROLL)) - EDIT_BuildLineDefs_ML(wnd, es); + EDIT_BuildLineDefs_ML(wnd, es, 0, strlenW(es->text), 0, (HRGN)0); } @@ -2836,6 +2988,7 @@ UINT e; UINT i; LPWSTR p; + HRGN hrgn = 0; TRACE("%s, can_undo %d, send_update %d\n", debugstr_w(lpsz_replace), can_undo, send_update); @@ -2916,9 +3069,14 @@ CharLowerBuffW(p, strl); s += strl; } - /* FIXME: really inefficient */ if (es->style & ES_MULTILINE) - EDIT_BuildLineDefs_ML(wnd, es); + { + INT s = min(es->selection_start, es->selection_end); + + hrgn = CreateRectRgn(0, 0, 0, 0); + EDIT_BuildLineDefs_ML(wnd, es, s, s + strl, + strl - (es->selection_end - es->selection_start), hrgn); + } else EDIT_CalcLineWidth_SL(wnd, es); @@ -2930,7 +3088,12 @@ /* force scroll info update */ EDIT_UpdateScrollInfo(wnd, es); - /* FIXME: really inefficient */ + if (hrgn) + { + EDIT_UpdateTextRegion(wnd, hrgn, TRUE); + DeleteObject(hrgn); + } + else EDIT_UpdateText(wnd, NULL, TRUE); if(es->flags & EF_UPDATE) @@ -3134,7 +3297,7 @@ EDIT_EM_EmptyUndoBuffer(es); es->flags &= ~EF_MODIFIED; es->flags &= ~EF_UPDATE; - EDIT_BuildLineDefs_ML(wnd, es); + EDIT_BuildLineDefs_ML(wnd, es, 0, strlenW(es->text), 0, (HRGN)0); EDIT_UpdateText(wnd, NULL, TRUE); EDIT_EM_ScrollCaret(wnd, es); /* force scroll info update */ @@ -3200,7 +3363,7 @@ EDIT_EM_EmptyUndoBuffer(es); es->flags &= ~EF_MODIFIED; es->flags &= ~EF_UPDATE; - EDIT_BuildLineDefs_ML(wnd, es); + EDIT_BuildLineDefs_ML(wnd, es, 0, strlenW(es->text), 0, (HRGN)0); EDIT_UpdateText(wnd, NULL, TRUE); EDIT_EM_ScrollCaret(wnd, es); /* force scroll info update */ @@ -3403,7 +3566,7 @@ es->word_break_proc16 = NULL; if ((es->style & ES_MULTILINE) && !(es->style & ES_AUTOHSCROLL)) { - EDIT_BuildLineDefs_ML(wnd, es); + EDIT_BuildLineDefs_ML(wnd, es, 0, strlenW(es->text), 0, (HRGN)0); EDIT_UpdateText(wnd, NULL, TRUE); } } @@ -3422,7 +3585,7 @@ es->word_break_proc = NULL; es->word_break_proc16 = wbp; if ((es->style & ES_MULTILINE) && !(es->style & ES_AUTOHSCROLL)) { - EDIT_BuildLineDefs_ML(wnd, es); + EDIT_BuildLineDefs_ML(wnd, es, 0, strlenW(es->text), 0, (HRGN)0); EDIT_UpdateText(wnd, NULL, TRUE); } } @@ -3683,6 +3846,8 @@ */ static void EDIT_WM_Destroy(WND *wnd, EDITSTATE *es) { + LINEDEF *pc, *pp; + if (es->hloc32W) { while (LocalUnlock(es->hloc32W)) ; LocalFree(es->hloc32W); @@ -3695,6 +3860,15 @@ while (LOCAL_Unlock(wnd->hInstance, es->hloc16)) ; LOCAL_Free(wnd->hInstance, es->hloc16); } + + pc = es->first_line_def; + while (pc) + { + pp = pc->next; + HeapFree(GetProcessHeap(), 0, pc); + pc = pp; + } + HeapFree(GetProcessHeap(), 0, es); *(EDITSTATE **)wnd->wExtra = NULL; } @@ -4447,7 +4621,7 @@ EDIT_SetRectNP(wnd, es, &r); if (es->style & ES_MULTILINE) - EDIT_BuildLineDefs_ML(wnd, es); + EDIT_BuildLineDefs_ML(wnd, es, 0, strlenW(es->text), 0, (HRGN)0); else EDIT_CalcLineWidth_SL(wnd, es); @@ -4678,6 +4852,21 @@ if (dy) EDIT_EM_LineScroll(wnd, es, 0, dy); return 0; +} + +/********************************************************************* + * + * EDIT_UpdateText + * + */ +static void EDIT_UpdateTextRegion(WND *wnd, HRGN hrgn, BOOL bErase) +{ + EDITSTATE *es = *(EDITSTATE **)((wnd)->wExtra); + + if (es->flags & EF_UPDATE) + EDIT_NOTIFY_PARENT(es, EN_UPDATE, "EN_UPDATE"); + + InvalidateRgn(wnd->hwndSelf, hrgn, bErase); }

1 0

[CHECKER] Stanford enhanced gcc found 741 errors in wine source
by Dan Kegel 27 Mar '01

27 Mar '01

... just kidding. Has anyone else been watching the success of the Stanford MC team in finding linux kernel bugs using an enhanced version of gcc? See http://www.lwn.net/2001/0322/kernel.php3 and scroll down to 'Global Kernel Analysis'. They've done a fabulous job. This and the Kernel Janitor project promise to help keep the linux kernel stable even in those funny corner cases that don't get exercised enough to make the bugs easy to track down. Would Wine be another possible application of the MC work? - Dan

1 0

Nasty Wine/kernel bug, anyone?
by Jeremy White 27 Mar '01

27 Mar '01

I'm able to consistently reproduce a kernel bug whilst running Wine (both 0305 and todays CVS) and the version of Outlook Express that comes with Win98. That is, if I use a pretty stock Wine config, and launch msimn.exe, we get the splash screen, and then a process stuck in 'D' (uninterruptible sleep). At this point, AFAIK, the only solution is to wait for the IO to timeout (I haven't been patient enough to see if it does), or to reboot. Because the process is already gone, I can't strace or gdb it. Anyone a) Want to take this on? b) Have advice on how I can generate enough debug info that we can feed it to the Kernel guys to get it fixed? And yes, this happens both on a 2.2 system and a 2.4 system. I saw at least other user report on this (Steve Fox reported this with Lotus Notes). My best guess is that this is a bug in the vfat driver, but that's just a hunch. Jer

2 7

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

wine-devel