wine-devel

wine-devel@winehq.org

2 participants
84469 discussions

(no subject)
by Mike McCormack 22 Aug '01

22 Aug '01

Hi Eric, How about returning a function pointer instead of an id? When the file handle is registered with the wine-server, we would also send function pointers for the read and write operations. This way we could have seperate implementations of ReadFile or WriteFile for different types of handles... Mike > - extend the get_handle_fd to also return an id corresponding > to the type fd being used. ------------------------------------------ mailto:Mike_McCormack@start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

1 0

Re: documentation make fix
by Mike McCormack 21 Aug '01

21 Aug '01

Yeah, that looks better, and seems to work too. thanks, Mike Original message from: Francois Gouget <fgouget(a)free.fr> > >On Tue, 21 Aug 2001, Mike McCormack wrote: > >> >> Not that anybody uses the documentation anyway... :-) >> >> ChangeLog: >> * recurse all directories when making docs. >> Don't fail if C_SRCS is empty. > > Maybe the attached patch would be a better way to do this: > - use the regular make trick for recursing > - test whether we have targets before entering the for loop > > Note that I did not really test it. > > > Well, I have done a little testing now. And c2man is not happy if you >don't create documentation/man3w beforehand... > So I added an mkdir too, same for doc-html. > > >-- >Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ > "Utilisateur" (nom commun) : > Mot utilis� par les informaticiens en lieu et place d'"idiot". ------------------------------------------ mailto:Mike_McCormack@start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

1 0

Re: reaktivate
by Malte Starostik 21 Aug '01

21 Aug '01

Am Montag, 20. August 2001 08:59 schrieb Per Wigren: > I get this when I'm trying to load an Active-X component in konqueror: > > konqueror: /usr/lib/libreaktivate: could not load library > '/usr/lib/libreaktivate.so' as Winelib application: > /usr/X11R6/lib/libXv.so.1: undefined symbol: XGetErrorDatabaseText > > Any ideas what to do? Using cvs from today of both wine and reaktivate. Never seen that one before, but it's IMHO rather a wine problem. Has anyone seen this before? Or, did you by any chance update some but not all X libraries? -- Malte Starostik PGP: 1024D/D2F3C787 [C138 2121 FAF3 410A 1C2A 27CD 5431 7745 D2F3 C787]

1 0

Re: ReadFile and other issues
by Alexandre Julliard 21 Aug '01

21 Aug '01

eric pouech <eric.pouech(a)wanadoo.fr> writes: > no, reading of input_records would be sone as it is today, with > the specific console request 'read_console_input' OK, we agree then. > the only thing we'd have to have is another request to get the fd from > a console input handle (get_handle_fd would no longer work), and we > still need a (background) mechanism to convert the unix xterm characters > into INPUT_RECORD (a get_console_fd should do). Anyway, those fd are > even rather static (they only change on Alloc|Free Console and at > process startup if we inherit them from parent), so we can savely keep > them around, process wide I'm not sure we should be converting xterm input into INPUT_RECORD. It may be better to have two modes, a simple console that returns an fd to ReadFile and no input records (and then ReadConsoleInput could generate them on the fly if necessary), and a complex console that supports all the features but that doesn't have an associated fd at all. IMO the current xterm-based console support is a ugly hack; we should do a proper Windows console using a normal Wine window with a message loop to retrieve input events. And usage of stdin/stdout should be limited to programs that don't need special console features, and should only support ReadFile/WriteFile. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: ReadFile and other issues
by Alexandre Julliard 21 Aug '01

21 Aug '01

eric pouech <eric.pouech(a)wanadoo.fr> writes: > to implement properly this (mainly the ability to forward ReadFile calls > to ReadConsole for console handles), I'd like to do the following: > - extend the get_handle_fd to also return an id corresponding to the > type fd being used. As of today, handle / fd relationship mainly > assumes that we can map a Windows operation on a handle onto a similar > operation on a unix fd. The ReadFile/ReadConsoleInput doesn't fit in > this category. I'm not sure I understand you right; are you suggesting that get_handle_fd should return an fd from which you can read input records in the case of a console? It would seem preferable to have get_handle_fd simply fail with a special error code saying that the handle is a console, and then ReadFile can fall back to using ReadConsoleInput which doesn't require an fd at all. > - this type of feature would help in other areas too. For example, we > never check that Comm functions really get a Comm handle... Well, we check by the fact that we then do a serial ioctl on the fd, which will fail if it isn't serial. I'm not sure we need more checks. -- Alexandre Julliard julliard(a)winehq.com

2 1

Archive of overlapped serial I/O example code, experiences
by Dan Kegel 21 Aug '01

21 Aug '01

For what it's worth, here is some example test code I wrote for win32 demoing overlapped serial I/O: http://www.alumni.caltech.edu/~dank/serio/serio.zip It might be of some use to people testing wine. I haven't tried it yet there. Also, http://www.alumni.caltech.edu/~dank/overlap.htm chronicles the problems many people have had with serial I/O under win95. The messages there get into interesting details of how people use overlapped serial I/O, and how serial I/O really behaves under various flavors of windows. Again, might be of some use, who knows. - Dan -- "I have seen the future, and it licks itself clean." -- Bucky Katt

1 0

ReadFile and other issues
by eric pouech 21 Aug '01

21 Aug '01

while we're at talking about the ReadFile implementation with overlapped data, there's (an unrelated) point about ReadFile I'd like to discuss I've started looking at implementing proper console input in Wine (mainly the line mode editing). Apart from providing the functionnality, it would also allow to remove this unneeded code from the debugger the Win32 console input API is made of two layers 1/ the low level : each input event is stored into an INPUT_RECORD struct 2/ the high level which uses the low level events to make strings of them 1/ two functions are used: ReadConsoleInput(A|W) 2/ three functions are used: ReadFile and ReadConsole(A|W) the current implementation in Wine, makes the handle to the console input a file handle to the fd gotten from the console hence, the current implementation is broken in several ways - ReadFile is implemented as a raw read on the input fd, whereas it should be a higher level of function (forwarded from ReadFile to ReadConsole) - wait operations on handle are wrong: currently the handle is signaled as soon as a char is read (at the unix level) from the stdin fd, whereas it should be done at the level of the INPUT_RECORD queue to implement properly this (mainly the ability to forward ReadFile calls to ReadConsole for console handles), I'd like to do the following: - extend the get_handle_fd to also return an id corresponding to the type fd being used. As of today, handle / fd relationship mainly assumes that we can map a Windows operation on a handle onto a similar operation on a unix fd. The ReadFile/ReadConsoleInput doesn't fit in this category. - this type of feature would help in other areas too. For example, we never check that Comm functions really get a Comm handle... would such a mechanism be acceptable ? if so, should we do for every handle in the server, or just the ones we might convert into unix fd (file, consoles, pipes, sockets, "serials"...) A+ -- --------------- Eric Pouech (http://perso.wanadoo.fr/eric.pouech/) "The future will be better tomorrow", Vice President Dan Quayle

1 0

Re: icons uncoloured
by Gerard Patel 21 Aug '01

21 Aug '01

At 01:39 PM 21/08/2001 GMT, you wrote: >I've just done a cvs update and recompiled and now the icons come up in >gray, if at all. > >Is this a known issue? I suspect it could be solved by this one I just sent to wine-patches Gerard Index: graphics/x11drv/bitmap.c =================================================================== RCS file: /home/wine/wine/graphics/x11drv/bitmap.c,v retrieving revision 1.32 diff -u -r1.32 bitmap.c --- graphics/x11drv/bitmap.c 2001/08/20 19:19:18 1.32 +++ graphics/x11drv/bitmap.c 2001/08/21 14:17:07 @@ -73,7 +73,7 @@ if (hbitmap == dc->hBitmap) return hbitmap; /* nothing to do */ if (!(bmp = GDI_GetObjPtr( hbitmap, BITMAP_MAGIC ))) return 0; - if (bmp->header.dwCount) + if (bmp->header.dwCount && (hbitmap != GetStockObject(DEFAULT_BITMAP))) { WARN( "Bitmap already selected in another DC\n" ); GDI_ReleaseObj( hbitmap ); ~ ~

2 1

icons uncoloured
by Bill Medland 21 Aug '01

21 Aug '01

I've just done a cvs update and recompiled and now the icons come up in gray, if at all. Is this a known issue? Bill

1 0

Re: documentation make fix
by Francois Gouget 21 Aug '01

21 Aug '01

On Tue, 21 Aug 2001, Mike McCormack wrote: > > Not that anybody uses the documentation anyway... :-) > > ChangeLog: > * recurse all directories when making docs. > Don't fail if C_SRCS is empty. Maybe the attached patch would be a better way to do this: - use the regular make trick for recursing - test whether we have targets before entering the for loop Note that I did not really test it. Well, I have done a little testing now. And c2man is not happy if you don't create documentation/man3w beforehand... So I added an mkdir too, same for doc-html. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ "Utilisateur" (nom commun) : Mot utilisé par les informaticiens en lieu et place d'"idiot".

1 0

Re: fix for Lawson's CLOCAL problem
by Mike McCormack 21 Aug '01

21 Aug '01

Hi Alexandre, >Of course, I have no problem with that. The create_file request takes >attributes, so create_serial can do the same (it would be better if >create_serial were done inside create_file, but that's another >story...) agreed. >> Another question: do you think it would be possible to implement a >> "write memory" message from the wineserver to the client? .. >But that would require multiple copies of the data (and many context >switches for large reads); that wouldn't be very efficient. I think >the current approach is much better. What is the problem with doing >CancelIo? IMO it would be messy to implement CancelIo with the current architecture. If it were all in the server (including actual read and write calls) it might be simpler... but probably not efficient, like you say. i think most of the data transferred from a serial port is pretty small though... but if the scheme was expanded to handle overlapped socket operations that wouldn't be true. Additionally, i though it might be useful to have a way to write small chunks of data to the client's memory. For example, when implementing WaitForNamedPipe and ConnectNamedPipe, i need to write a result code into an overlapped structure... it's only 4 bytes, but i need to create a new type of APC to deal with it. If i had a method to write directly into client memory, it would be simpler. Mike ------------------------------------------ mailto:Mike_McCormack@start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

1 0

Re: fix for Lawson's CLOCAL problem
by Alexandre Julliard 21 Aug '01

21 Aug '01

Mike McCormack <mike_mccormack(a)start.com.au> writes: > Hmmm. i'm not exactly sure about this. i don't think the server ever > blocked on serial port reads or opens... i think the reason Lawson had > problems was that the client thread was blocking in the overlapped > read code. (FILE_AsyncReadService in files/file.c) I don't know if this was Lawson problem, but open() can potentially block on a serial port, and we cannot have that in the server; so your change to set O_NONBLOCK is good, but it shouldn't depend on FILE_FLAG_OVERLAPPED. > If we use non-blocking fds unconditionally, a non overlapped ReadFile > may result in a busy loop in the client thread. Sure, you need to reset the O_NONBLOCK flag to do blocking reads. > Additionally, overlapped read (and write) requests should fail unless > the handle is opened with FILE_FLAG_OVERLAPPED. There are programs > that depend on that behaviour (as reported by Eric Kohl, 3 Feb 2001 on > this list). This is another reason to propagate CreateFile's > attributes parameter into the wineserver. Of course, I have no problem with that. The create_file request takes attributes, so create_serial can do the same (it would be better if create_serial were done inside create_file, but that's another story...) > Another question: do you think it would be possible to implement a > "write memory" message from the wineserver to the client? This message > would be carried over wait_fd and would be received by the client > asynchronously in wait_reply(). The message would contain a length, > address and data to be written. If this is possible, i could move > async i/o back into the server, keep it efficient and perhaps > implement CancelIo properly... But that would require multiple copies of the data (and many context switches for large reads); that wouldn't be very efficient. I think the current approach is much better. What is the problem with doing CancelIo? -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: winsock stuff
by Daniel Walker 21 Aug '01

21 Aug '01

This seems to work.. Daniel Walker --- server/sock.c 2001/05/14 20:09:39 1.18 +++ server/sock.c 2001/08/21 01:31:04 @@ -113,9 +113,14 @@ { struct sock *sock = (struct sock *)obj; unsigned int emask; + char rBuf; assert( sock->obj.ops == &sock_ops ); if (debug_level) fprintf(stderr, "socket %d select event: %x\n", sock->obj.fd, event); + + if (event == POLLIN && !recvfrom(sock->obj.fd, &rBuf, sizeof(rBuf), MSG_PEEK, NULL, NULL)) + event = POLLHUP; + if (sock->state & WS_FD_CONNECT) { /* connecting */

1 0

Re: fix for Lawson's CLOCAL problem
by Mike McCormack 21 Aug '01

21 Aug '01

Hi Alexandre, Hmmm. i'm not exactly sure about this. i don't think the server ever blocked on serial port reads or opens... i think the reason Lawson had problems was that the client thread was blocking in the overlapped read code. (FILE_AsyncReadService in files/file.c) If we use non-blocking fds unconditionally, a non overlapped ReadFile may result in a busy loop in the client thread. Additionally, overlapped read (and write) requests should fail unless the handle is opened with FILE_FLAG_OVERLAPPED. There are programs that depend on that behaviour (as reported by Eric Kohl, 3 Feb 2001 on this list). This is another reason to propagate CreateFile's attributes parameter into the wineserver. Another question: do you think it would be possible to implement a "write memory" message from the wineserver to the client? This message would be carried over wait_fd and would be received by the client asynchronously in wait_reply(). The message would contain a length, address and data to be written. If this is possible, i could move async i/o back into the server, keep it efficient and perhaps implement CancelIo properly... Mike > The open on the server side must always be non-blocking, we cannot > have the server block on anything. If there are cases where we need to > block on open, it will have to be done on the client side. ------------------------------------------ mailto:Mike_McCormack@start.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

1 0

Re: winsock stuff
by Alexandre Julliard 21 Aug '01

21 Aug '01

Daniel Walker <diwalker(a)earthlink.net> writes: > How is it unreliable? Why is MSG_PEEK more reliable? I'm not opposed to > using MSG_PEEK , just curios.. FIONREAD is unreliable because you cannot distinguish between an EOF condition and an empty receive buffer. recvmsg(MSG_PEEK) is supposed to return 0 on EOF, and EWOULDBLOCK on empty buffer, so you can determine the real status. > Another issue is why doesn't poll() give > us a POLLHUP? That would be ideal, get poll() to work the way we need it > too. As I understand it poll() returns POLLHUP only when both directions of the socket have been closed; but if you shutdown() only the read or write side, you don't get POLLHUP so that you can continue using poll() for the other direction. This has apparently been changed between kernels 2.2 and 2.4; in 2.2 POLLHUP is returned as soon as the read side of the socket has been closed. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: winsock stuff
by Alexandre Julliard 21 Aug '01

21 Aug '01

Daniel Walker <diwalker(a)earthlink.net> writes: > To improve on this patch we could only do the FIONREAD if the sock is > in a non-blocking state. That's the only time that doing the FIONREAD > would be an advantage. Also we wouldn't need to worry about screwing up > blocking sockets. Even further, we could only do the FIONREAD if socket > is non-blocking and looking for an FD_CLOSE event. Using FIONREAD is simply not an option, it cannot be made reliable. A possibility to detect EOF could be a recvmsg with MSG_PEEK flag, though you should probably verify first that it behaves as expected. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: winsock stuff
by Alexandre Julliard 20 Aug '01

20 Aug '01

Daniel Walker <diwalker(a)earthlink.net> writes: > I noticed that WSAEnumNetworkEvents() was returning all events that had > occurred and not just the events that had been selected with > WSAEventsSelect(). Win98 will trigger the event object for all events > but WSAEnumNetworkEvents() (in win98) would return zero if the event > hadn't been selected. I was trying to gain compatibility with win98. I'd suggest doing the masking in WSAEnumNetworkEvents then. > An FD_READ event with a zero buffer would mean that the other side > closed the connection so it would be equivalent to a POLLHUP from poll() > .. I didn't find an occasion when poll() actually returned POLLHUP when > the socket closed, on Linux anyway. The application that I was testing > got confused when recv() returned zero. Even though recv() can legally > return zero, I think non-blocking applications are going to be looking > for FD_CLOSE events as opposed to recv() returning zero. But FIONREAD can return 0 bytes even when the socket isn't closed (for instance if another thread got the data first). You can only determine that it was really closed when you do the actual read. So I don't really see how we can avoid having recv() return zero; well maybe we can make it return EWOULDBLOCK and tell the server the connection is closed, but I'm not sure it's any better. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: winsock stuff
by Alexandre Julliard 20 Aug '01

20 Aug '01

Daniel Walker <diwalker(a)earthlink.net> writes: > -changed get_socket_event() so it only returns events that > the application is looking for. > -Changed sock_poll_event() to interpret a POLLIN event with > zero bytes waiting to be read as a POLLHUP. This doesn't look really right to me. Could you please explain a bit more what it is supposed to fix? -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: Refcounting in SelectObject
by Alexandre Julliard 20 Aug '01

20 Aug '01

Uwe Bonnes <bon(a)elektron.ikp.physik.tu-darmstadt.de> writes: > Appended code tries to clean up better by: > - Remembering an unsuccessfull DeleteObject by adding 0x8000 to the Refcount > - Call DeleteObject in SelectObject when the RefCount of the released handle > is 0x8000 > - Decrementing the RefCount when a hBitmap is still selected into a DC when > DeleteDC is called for that DC > - Call DeleteObject in DeleteDC when RefCount is 0x8000 > > For some applications I tested, I always saw the delayed DeleteObject done > after the refused DeleteObject in the first place. Is this approach better? Yes, looks good. Though we can probably safely use the dwCount field in GDIOBJHDR since it is not used for anything useful right now, instead of adding new fields. I've changed that and applied the patch. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: fix for Lawson's CLOCAL problem
by Alexandre Julliard 20 Aug '01

20 Aug '01

Mike McCormack <mike_mccormack(a)start.com.au> writes: > Open serial ports with O_NONBLOCK if FILE_FLAG_OVERLAPPED is > specified. The open on the server side must always be non-blocking, we cannot have the server block on anything. If there are cases where we need to block on open, it will have to be done on the client side. -- Alexandre Julliard julliard(a)winehq.com

1 0

-clocal and the wineserver
by lawson_whitney＠juno.com 19 Aug '01

19 Aug '01

Good day! Any time clocal falls out of the modem's termios (say I use cu -l on it) the wineserver hangs trying to open it, or at least the app hangs, and I haven't found a way to kill it without killing the wineserver. Both 16 and 32 bit versions of juno are affected. I haven't really investigated it very thouroughly, but I wasted a lot of time restoring backup versions of the app and wine. Should I look into mickey-mousing the open to avoid the problem, or - is it really a wineserver problem which I am probably not even marginally competent to muck with? I'd be glad to make traces or winedbg output, but it seems the problem should be easy to duplicate. Hmmm, setting clocal back on from another terminal unhangs the app. That is neater than killing the wineserver, I guess. terminal.exe from wfw 3.11 behaves similarly, with an extra twist: it hangs on the Win16Mutex until somebody sets clocal (well, I guess carrier detect would work, too). [whit@giftie user0001]$ wine terminal.exe err:ntdll:RtlpWaitForCriticalSection section 0x400fad6c "syslevel.c: Win16Mutex" wait timed out, retrying (60 sec) fs=008f fixme:pthread_kill_other_threads_np Console: Making console complex (creating an xterm)... fixme:pthread_kill_other_threads_np fixme:pthread_kill_other_threads_np fixme:pthread_kill_other_threads_np fixme:pthread_kill_other_threads_np Warning: locale not supported by Xlib, locale set to C Warning: X locale modifiers not supported, using default xterm: unable to open font "vga", trying "fixed".... Failed to open input method passed master doesn't match slave name passed master doesn't match slave name [whit@giftie user0001]$ Regards, Lawson ---oof--- ________________________________________________________________ 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: PATCH: olepictures black & white
by Alexandre Julliard 19 Aug '01

19 Aug '01

Marcus Meissner <marcus(a)jet.franken.de> writes: > Somehow the loaded JPEGs are now black & white in the IPicture implementation. > > Exchanging CreateCompatibleDC(0) by CreateDCA("DISPLAY",0,0,0); makes > them colored again. > > But I am not sure whether the problem itself was a bug on my side, or > on the WINE side... Maybe a bit of both... I think the patch below should fix the GDI part; but is there any reason you don't use GetDC(0) instead of CreateCompatibleDC? It would avoid creating a new DC every time. Index: objects/dib.c =================================================================== RCS file: /opt/cvs-commit/wine/objects/dib.c,v retrieving revision 1.60 diff -u -r1.60 dib.c --- objects/dib.c 2001/08/17 00:07:42 1.60 +++ objects/dib.c 2001/08/19 19:27:42 @@ -875,7 +875,9 @@ /* Now create the bitmap */ - if (fColor) handle = CreateCompatibleBitmap( hdc, width, height ); + if (fColor) + handle = CreateBitmap( width, height, GetDeviceCaps( hdc, PLANES ), + GetDeviceCaps( hdc, BITSPIXEL ), NULL ); else handle = CreateBitmap( width, height, 1, 1, NULL ); if (!handle) return 0; -- Alexandre Julliard julliard(a)winehq.com

1 0

PATCH: oleauto header problem
by Nikolas Zimmermann 19 Aug '01

19 Aug '01

Hi there, Reaktivate doesn't compile anymore. We need this patch. Bye Bye Niko -- Nikolas Zimmermann wildfox(a)kde.org

1 0

Re: WM_GETTEXTLENGTH change in winproc.c
by Gerard Patel 19 Aug '01

19 Aug '01

At 02:35 PM 18/08/2001 -0700, you wrote: >For standard controls like buttons it should never be a problem since >there is both an ASCII and Unicode win proc, so in a pure ASCII app we >should never do the conversion. I think the problem is that the >dialog code does a CreateWindowW, so the button is Unicode even though >the app does everything in ASCII. I guess we'll have to maintain both >A and W win procs inside each window, not only inside the class. How about creating Ascii windows when the app calls DialogBox..A variants ? Gerard

1 0

Re: WM_GETTEXTLENGTH change in winproc.c
by Alexandre Julliard 18 Aug '01

18 Aug '01

Gerard Patel <gerard.patel(a)nerim.net> writes: > How about creating Ascii windows when the app calls DialogBox..A variants ? That could be a workaround, but since dialog resources are Unicode it's better to avoid converting back and forth. Besides, Windows maintains two different win procs for windows of standard classes so we should do it this way too. I'll try to come up with something. -- Alexandre Julliard julliard(a)winehq.com

1 0

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