April 2002 - Wine-Devel - list.winehq.org

[Patch tested] TB_REPLACEBITMAP
by Pascal.Brisset＠wanadoo.fr April 29, 2002

April 29, 2002

I just tested Marco Bizzarri's implementation of TB_REPLACEBITMAP ( http://www.winehq.com/hypermail/wine-patches/2001/11/0112.html ). It looks like this patch was neither rejected nor accepted at that time. The patch applies cleanly to sources from wine-cvs-unstripped-042902. I can't tell whether it is technically adequate, but at least it significantly improves wine support for Checkpoint's Firewall-1 GUI (FWPolicy.exe), which might be of interest to many people ( see http://bugs.codeweavers.com/show_bug.cgi?id=185 ). -- Pascal

1 0

Re: Stub implementation of SNMPAPI.DLL.
by Michael Cardenas April 29, 2002

April 29, 2002

What application are you using that needs snmpapi.dll ? aol7 uses it. how did you generate this stub dll? On Mon, Apr 29, 2002 at 05:59:15PM +0200, Patrik Stridvall wrote: > *** snmpapi > > * dlls/snmpapi/Makefile.in, > dlls/snmpapi/main.c, > dlls/snmpapi/snmpapi.spec: > Stub implementation of SNMPAPI.DLL. > -- michael cardenas lead windows compatibility engineer lindows.com "Be the change you want to see in the world" -Mahatma Gandhi

1 0

dragging trillian's window causes the X server to crash
by Michael Cardenas April 29, 2002

April 29, 2002

Using wine from 04/23/2002, when drag my buddy list window in trillian the X server crashes.

3 5

warning while compiling with gcc3
by Rizsanyi Zsolt April 29, 2002

April 29, 2002

Hi! I have run into a warning, when compiling with gcc 3.1 (I dont know if it is reported or not with 2.x gccs). And I think the warning is right. I dont think that the initializer of nil is right. It initializes an zero (0) character length WCHAR array to a one character lenght array. Or I just does not understand something? Here is the code snippet from dlls/commctl32/ipaddress.c line 265: static void IPADDRESS_ClearAddress (IPADDRESS_INFO *infoPtr) { WCHAR nil[0] = { 0 }; int i; TRACE("\n"); for (i = 0; i < 4; i++) SetWindowTextW (infoPtr->Part[i].EditHwnd, nil); } Best regards Zsolt Rizsanyi

2 1

argc/argv issue - More info, hack and questions...
by Ann and Jason Edmeades April 29, 2002

April 29, 2002

Hello, I am still investigating this argv/argc issue. The sequence of events as far as I can tell: CreateProcessA(NULL, "regsvr32 comdlg32.ocx",....) Process_CREATE calls fork_and_exec( NULL, "regsvr32 comdlg32.ocx", ...) This sees there is no explicitly specified filename so inserts wine and -- into the command line (now "wine -- regsvr32 comdlg32.ocx") Eventually the new process starts with the wine -- regsvr32 comdlg32.ocx as the command line, so it needs to strip off the inserted fields, and in PROCESS_InitWine, main_exe_name is correct (path\regsvr32.exe) but the argc/argv includes these new fields. At the top it states: (Scheduler/process.c, lines 491/492): app_argv++; /* remove argv[0] (wine itself) */ app_argc--; This obviously leaves behine the -- and things go downhill from there onwards. Interestingly wine -- regsvr32 comdlg32.ocx works fine, so there must be command line processing here which doesnt occur during the CreateProcess route. Question: Should fork_and_exec insert the -- into the command line? If it should, should PROCESS_InitWine check for it, ie: app_argv++; /* remove argv[0] (wine itself) */ app_argc--; if (app_argc && strcmp(app_argv[0], "--") == 0) { app_argv++; /* remove argv[1] ('--' inserted by fork_and_exec) */ app_argc--; } However, what implications are of doing this (it works, of course). This feels like hacking around an imcompatibility within wine but I am happy to submit a patch as it definitely resolves a problem for me. Regards, Jason

1 0

Unit testing
by Jakob Eriksson April 29, 2002

April 29, 2002

Is there a tutorial or so for how to create new unit tests in Wine? If not, I'll just go ahead and use those there are for blueprints. -- regards, Jakob Eriksson The wages of sin is debugging. -- Ron Jeffries

4 6

Wineconsole using all memory
by Sylvain Petreolle April 29, 2002

April 29, 2002

Hi Eric, I had got memory problems by the past. When run, wineconsole was using the whole memory on my system. Problem was never solved: it had also disappeared whithout doing anything. (we only had put it in trace mode and enter the correct values into registry, that had solved nothing) It's already there today. So an output from wineconsole telnet gives this : (put in trace mode) Change notification: active(72) resize(80,150)fixme:bitmap:CreateCompatibleBitmap got bad width 2313600 or height 2457900, please look for reason err:ntdll:RtlpWaitForCriticalSection section 0x400fe70c "module.c: loader_section" wait timed out, retrying (60 sec) fs=0247wine: Unhandled exception, starting debugger... win-size(80,25) update(0,149) curs-geom(100,1). Done err:font:XFONT_RealizeFont plf->lfHeight = 16386, Creating a 100 pixel font and rescaling metrics fixme:bitmap:CreateCompatibleBitmap got bad width 694720 or height 2457750, please look for reason err:font:XFONT_RealizeFont plf->lfHeight = 16386, Creating a 100 pixel font and rescaling metrics Change notification: resize(132,500)fixme:bitmap:CreateCompatibleBitmap got bad width 1146288 or height 8192500, please look for reason win-size(132,51) update(0,50)err:font:XFONT_RealizeFont plf->lfHeight = 16386, Creating a 100 pixel font and rescaling metrics v-scroll(1) update(51,51)err:font:XFONT_RealizeFont plf->lfHeight = 16386, Creating a 100 pixel font and rescaling metrics v-scroll(2) update(52,52)err:font:XFONT_RealizeFont plf->lfHeight = 16386, Creating a 100 pixel font and rescaling metrics curs-pos(9,52). Done ___________________________________________________________ Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Yahoo! Mail : http://fr.mail.yahoo.com

1 0

Re: SafeDisc-1
by Dustin Navea April 28, 2002

April 28, 2002

--- Laurent Pinchart <laurent.pinchart(a)skynet.be> wrote: > I forgot a very important information: start your > program in NT mode > (--winver nt40 or above). > > Laurent Pinchart > would that include --winver win2k (not sure)? Dustin __________________________________________________ Do You Yahoo!? Yahoo! Health - your guide to health and wellness http://health.yahoo.com

2 1

argc / argv issue?
by Ann and Jason Edmeades April 28, 2002

April 28, 2002

Hello, I started debugging a relatively simple problem, which is I copied comdlg32.ocx and regsvr32.exe from a windows 2000 box to a no-windows wine installation (on Mandrake 8.2), and found that if I run 'wine regsvr32.exe comdlg32.ocx', it succeeds fine, whereas running the same command under wcmd caused failures. I debugged a little further and the problem appears to be caused by argc being wrong when wcmd (a winelib program) launches regsvr32.exe. I am continuing to investigate but I know argv/argc issues have cropped up lots and would appreciate any pointers. My debugging so far (edited, @@@@ are my extra trace points): ./wcmd -debugmsg +msvcrt,+process >~/wine/patches/dbg/wcmdproccessinfo 2>&1 WCMD starting: trace:process:process_init @@@@@ argv[0] = ./wcmd trace:process:process_init @@@@@ argc = 1 trace:process:process_init @@@@@ argv[0] = /usr/local/bin/wineconsole trace:process:process_init @@@@@ argv[1] = --use-event=52 trace:process:process_init @@@@@ argc = 2 Launching regsvr32.exe: trace:process:process_init @@@@@ argv[0] = /usr/local/bin/wine trace:process:process_init @@@@@ argv[1] = -- trace:process:process_init @@@@@ argv[2] = regsvr32 trace:process:process_init @@@@@ argv[3] = comdlg32.ocx trace:process:process_init @@@@@ argc = 4 *** Note argc is 4 (correctly) trace:msvcrt:msvcrt_init_mt_locks initializing mtlocks trace:msvcrt:msvcrt_init_io :handles (8)(12)(16) trace:msvcrt:msvcrt_init_console :Opening console handles trace:msvcrt:msvcrt_init_args got 'regsvr32 comdlg32.ocx', wide = L"regsvr32 comdlg32.ocx" argc=3 *** Note argc is 3 now, incorrectly Other debugging bits of info, originally I thought the problem was in native msvcrt so I added trace to local __getmainargs to trace out what was passed: wine: trace:msvcrt:msvcrt_init_args got 'regsvr32 comdlg32.ocx', wide = L"regsvr32 comdlg32.ocx" argc=2 *** 2 parms trace:msvcrt:__getmainargs @@@ MSVCRT Info trace:msvcrt:__getmainargs @@@ Arg 0 = regsvr32 trace:msvcrt:__getmainargs @@@ Arg 1 = comdlg32.ocx wcmd: trace:msvcrt:msvcrt_init_args got 'regsvr32 comdlg32.ocx', wide = L"regsvr32 comdlg32.ocx" argc=3 *** 3 parms trace:msvcrt:__getmainargs @@@ MSVCRT Info trace:msvcrt:__getmainargs @@@ Arg 0 = -- trace:msvcrt:__getmainargs @@@ Arg 1 = regsvr32 trace:msvcrt:__getmainargs @@@ Arg 2 = comdlg32.ocx I'll continue to investigate, but suggestions welcome. Regards, Jason jason @ the-edmeades.demon.co.uk

2 3

RE: [ros-kernel] Status of building Wine under Cygwin/Mingw
by Robert Collins April 28, 2002

April 28, 2002

> -----Original Message----- > From: Francois Gouget [mailto:fgouget(a)free.fr] > Sent: Monday, April 29, 2002 6:15 AM > To: Robert Collins > Cc: Steven_Ed4153(a)yahoo.com; wine-devel(a)winehq.com; ReactOS-Kernel > Subject: RE: [ros-kernel] Status of building Wine under Cygwin/Mingw > > > On Sun, 28 Apr 2002, Robert Collins wrote: > [...] > > > 4. add -liberty to the makefile for wrc and wmc. This is needed > > > because of getopt and optarg > > > > Cygwin has a BSD licenced getopt.c that doesn't need libiberty to > > link, only mingw (AFAIK). Anyway worth investigating? > > I usually regard getopt functions as higly unportable and > always try to avoid them. I did not check how they are used > in wrc&co, but it should not be very difficult to do entirely > without them. I believe that would be the best solution. You must end up writing the same bit of code many times. Ahh well, have fun. Rob

1 0

RE: [ros-kernel] Status of building Wine under Cygwin/Mingw
by Robert Collins April 28, 2002

April 28, 2002

> -----Original Message----- > From: Steven Edwards [mailto:Steven_Ed4153(a)yahoo.com] > Sent: Sunday, April 28, 2002 2:35 PM > 2. run ./configure --host=mingw32 --target=mingw32 > --build=mingw32 CFLAGS="-D_MINGW_ -D_WINDOWS -DWINE_NOWINSOCK" You should probably add CC="gcc -mno-cygwin" CXX="gcc -mno-cygwin" to the configure line if you don't have a mingw cross compiler installed. > 4. add -liberty to the makefile for wrc and wmc. This is > needed because of getopt and optarg Cygwin has a BSD licenced getopt.c that doesn't need libiberty to link, only mingw (AFAIK). Anyway worth investigating? Rob

3 2

CD-ROM not recognized
by Ian D. Stewart April 28, 2002

April 28, 2002

Howdy Folks, I'm trying to run Cryo's Arthur's Knights under wine. I've managed to install using setup.exe, but when I go to play the game, I get an error message telling me to insert CD #1, even though CD #1 is in the CD-ROM tray and mounted. Any clues as to why I would be getting this message, how to convince the game that the CD is actually in the CD-ROM drive, or how to research any of these issues myself? Thanx, Ian -- "How am I supposed to blend? I'm two feet taller than everybody else in here!" -- Chris Tucker Rush Hour 2

4 4

Re: CD-ROM not recognized
by Ian D. Stewart April 28, 2002

April 28, 2002

On 27 Apr 2002 13:52:30 -0700, Dustin Navea wrote: > --- "Ian D. Stewart" <idstewart(a)compuvative.com> > wrote: > > Howdy Folks, > > > > I'm trying to run Cryo's Arthur's Knights under > > wine. I've managed to > > install using setup.exe, but when I go to play the > > game, I get an error > > message telling me to insert CD #1, even though CD > > #1 is in the CD-ROM > > tray and mounted. > > > > Any clues as to why I would be getting this message, > > how to convince the > > game that the CD is actually in the CD-ROM drive, or > > how to research any > > of these issues myself? > > > > > > Thanx, > > Ian > > > > > > -- > > "How am I supposed to blend? I'm two feet taller > > than everybody else in here!" > > -- Chris Tucker > > Rush Hour 2 > > > > > > hello, i noticed your problem and thought I would take > a gander at it. It could only be 1 thing and there > are 3 ways to try to solve it. The problem is that > the game more than likely uses copy protection to > prevent software pirates (warez kiddies) from making > illegal copies of it. Under the cd drive's section > ([Drive D] or whatever) remove the line that says > "Label" = that will allow wine to read the name of the > CD...some games use the name of the cd as their only > copy protection (Frogger 2), and if that is the case > with your cd, then that would fix it. Try that. No joy. The 2nd way > would be if they used some sort of on-cd copy > protection (safedisc or securerom), to check out > TransGaming's version of wine (dubbed WineX). It has > many games supported (has the ability to read > safedisc/securerom-enabled cd's) and would be able to > read it. If that doesn't work, you can always call > for a vote for them to support it and put the > safedisc/securerom code into their program. I'll take a look into this. Thanx. > The 3rd > and least recommended way is to get what is called a > crack. It allows you to use the game without the cd > being in the drive. But remember though that the > government of the US and many other countries > considers it illegal and you could go to jail for > using one. Now this provides an interesting peek into the minds of our legislatures. As I understand it, under the DMCA, developing and/or distributing technology *capable of* circumventing copy protection is illegal. However, use of said technology is not... Thanx again for the suggestions. Ian -- "How am I supposed to blend? I'm two feet taller than everybody else in here!" -- Chris Tucker Rush Hour 2

1 0

Wine[X] and Grand Prix 3 (Microprose)
by Rene Rebe April 28, 2002

April 28, 2002

Hi all. I just turned my last 10 hours into configuring Wine to run Formula One Grand Prix 3 from Microprose. After several tries using the latest Wine and WineX (making many mistakes with the registry files ...) I got it to run using the last-night WineX cvs soruce ;-) I only have trhee graphic problems: a) the menus (where you configure several game options/details) do not appear in the right colors. Like a 8bit image displayed with the wrong color-pallete. b) in the race with very perfect weather (0% rain risk) - where the sky should be completely blue - the sky doesn't get painted at all -> resulting in "tearing" textures, because this area doesn't get refresh. (when some clouds show up the graphic is perfect) and c) Direct3D is not recognized at all -> resulting that the game runs in software-rendered mode. For a) I guess it migth be a wrong 16bit mode (or even an 8 bit mode) - my XFree-4.2.0 is running in 16bit mode normally. 24bit XFree results in the same mess and 8bit X results in a slighty worser images. About b) I guess some bitblt or clear function is nor working properly. I also tested the PerferctGraphics and DoubleBuffer option of Wine yielding the same results ... For c) OpenGL on my Matrox G450 is configured and runs fine - I have no clue whatever Wine[X] needs to provide Direct3D ... ? I can provide whatever trace output is needed - and would also offer to write some code if someone points me to the place in the souce where something is missing (I know C and C++ ...). I could also provide trace output to get (generic) Wine run the game. (It seems to be a problem that more DirectDraw is missing?). The rest of the game runs just fine, including Sound and Joystick support. Configuration: - Linux 2.4.19-pre7 - glibc-2.2.6 - gcc-2.95.3 - XFree-4.2.0 - AMD Athlon 1700XP; Matrox MGA 450 (OpenSouce driver) [- ROCK Linux (http://www.rocklinux.org) k33p h4ck1n6 René -- René Rebe (Registered Linux user: #248718 <http://counter.li.org>) eMail: rene.rebe(a)gmx.net rene(a)rocklinux.org Homepage: http://drocklinux.dyndns.org/rene/ Anyone sending unwanted advertising e-mail to this address will be charged $25 for network traffic and computing time. By extracting my address from this message or its header, you agree to these terms.

1 1

Status of building Wine under Cygwin/Mingw
by Steven Edwards April 28, 2002

April 28, 2002

Note: send any spelling/grammer errors to /dev/null. I've compiled a little list of things that are still lacking on the port of wine to mingw/msys/cygwin. Unless I mention cygwin as the target it can be assumed for the purpose of this doc I am speaking about a mingw target. About 5 or 6 months ago James Tablor and Casper Hornstroup ported wine to Mingw for ReactOS. ReactOS is very close to having basic Windowing support so I have set out to compleate the work they have started by bringing the wine/Rewind trees up to speed with our changes. The current wine build system still has a few things Missing before we can build without needing the reactos source tree/makefiles. Note: I have ordered the list by importance and the steps needed to build. 1. Make configure changes to build wine support under mingw/cygwin. Currently configure Has issues detecting how it should build the wine files as either dlls,so or .a I have submitted a patch for this but it was rejected because it was just a work-around And did not fix the problem. You will need to still run ./configure under cygwin as dlltool/wrap do not detect properly under msys with configure. With the current port in the ReactOS source tree we use our own Makefiles/resources and I would like to get away from this. 2. run ./configure --host=mingw32 --target=mingw32 --build=mingw32 CFLAGS="-D_MINGW_ -D_WINDOWS -DWINE_NOWINSOCK" 3. Remove winedump from the tools/Makefile as Mingw/Cygwin and others have support for dumping PE executable information and it is loaded with Unixisms. 4. add -liberty to the makefile for wrc and wmc. This is needed because of getopt and optarg 5. add a configure check to #define snprintf _snprintf for wrc when building with mingw. 6. Remove server, library and tsx11 for mingw. If you want to build wineserver under cygwin we will need to adapt get/set thread context from GDB's win32 support. 7. 90% of the source from the dlls compiles with no problems. The issue is when you want to link the objects as a dll and compile in the resources. winebuild now has support to build the *.def files needed for dlltool so there will need to be changes on the way that the spec/def files are used when building on a Win32 host. If one tries to build wine currently under Mingw/Cygwin it will puke when linking to the needed imports because winebuild will try to import from lib*.so rather then from dllname.a I've also been having troble linking with resources compiled with wrc under Mingw/Cygwin. 7. We do not want to build advapi32, Kernel32, ntdll, user32, gdi, msvcrt or crt.dll 8. These are the dlls we know that we can build and need: avicap32, avifile32, comctl32, commdlg, crypt32, ddraw, dinput, dplay, dplayx dsound(needs imports from winmm), imagehlp, lzexpand, mapi32, ole32, oleaut32 olecli, oledlg, olepro32, olesvr, richedit, rpcrt4, serialui, setupapi, shdocvw shell32, shfolder, shlwapi, tapi32, twain, url, urlmon, version, winmm, winspool wintrust. 9. Debugging will be differnt under Windows/ReactOS due to the wine debugger not working. Under ReactOS currently we have all of the FIXMEs and TRACES mapped to our NTDLL as DbgPrnt and others. I'm not 100% sure as to how we are doing this yet. 10. DLL seperation and Wineserver dependace. I wont even really be able to start looking at this untill the 10 issues above are worked out. I know that when I tested programs/dlls from our first wine port under ReactOS there were issues due to a lack of internal wineserver functions. 11. The chain of dependance for the Sound/Graphics for Winmm/DirectX and others will need interfaces developed for the HAL (Codeweavers/Transgamming?) Thanks Steven "Every revolution was once a thought in one man's mind" - Ralph Waldo Emerson

1 0

Portable instalation
by Marcelo Welter April 27, 2002

April 27, 2002

Greatings how i can make a portable instalation of wine (like a tgz or rpm) to send to final users i need a small and compact method to this (to donkey users..hehehehe) anybody have a step-by-step to install wine ? Thanks Marcelo

2 1

dlls/msvcrt/file.c broken on FreeBSD 4.5
by Gerald Pfeifer April 27, 2002

April 27, 2002

The following patch to dlls/msvcrt/file.c revision 1.23 date: 2002/04/26 19:05:16; author: julliard; state: Exp; lines: +4 -0 Better support for configure detection of missing types, added check for ssize_t. Removed a couple of no longer used portability functions. Various portability fixes in port.h. breaks FreeBSD 4.5: /usr/bin/gcc -c -I. -I. -I../../include -I../../include -g -O2 -Wall -mpreferred-stack-boundary=2 -fPIC -D__WINE__ -DUSE_MSVCRT_PREFIX -D_MT -D_REENTRANT -I/usr/X11R6/include -o file.o file.c In file included from file.c:43: ../../include/msvcrt/sys/stat.h:34: syntax error before `.' ../../include/msvcrt/sys/stat.h:35: syntax error before `.' ../../include/msvcrt/sys/stat.h:36: syntax error before `.' ../../include/msvcrt/sys/stat.h:48: syntax error before `.' ../../include/msvcrt/sys/stat.h:49: syntax error before `.' ../../include/msvcrt/sys/stat.h:50: syntax error before `.' file.c: In function `_fstat': file.c:693: structure has no member named `st_atimespec' file.c:695: structure has no member named `st_mtimespec' file.c:695: structure has no member named `st_ctimespec' file.c: In function `_stat': file.c:1181: structure has no member named `st_atimespec' file.c:1183: structure has no member named `st_mtimespec' file.c:1183: structure has no member named `st_ctimespec' file.c:1185: structure has no member named `st_atimespec' file.c:1185: structure has no member named `st_mtimespec' file.c:1185: structure has no member named `st_ctimespec' file.c: In function `_wstat': file.c:1242: structure has no member named `st_atimespec' file.c:1244: structure has no member named `st_mtimespec' file.c:1244: structure has no member named `st_ctimespec' file.c:1246: structure has no member named `st_atimespec' file.c:1246: structure has no member named `st_mtimespec' file.c:1246: structure has no member named `st_ctimespec' gmake[2]: *** [file.o] Error 1 gmake[2]: Leaving directory `/.amd_mnt/vexpert/files8/test/wine/dlls/msvcrt' gmake[1]: *** [msvcrt/msvcrt.dll.so] Error 2 Gerald

1 0

./tests/sock.c: No such file or directory
by Marcelo Welter April 27, 2002

April 27, 2002

helo, ./configure make depend ./tests/sock.c: No such file or directory and now...? Marcelo

1 0

Correction for WWN 121
by Speeddymon April 26, 2002

April 26, 2002

Hello, just had an update for WWN 121... Near the bottom it says that Alexandre rejected a patch for the ldconfig problem (the one i submitted) that puts $libdir/wine into ld.so.conf because users need to set $WINEDLLPATH... the corrections are this: users need to set $LD_LIBRARY PATH and The latest wineinstall patch I submitted removes the $libdir/wine addition and instead always runs ldconfig (just in case of modules being renamed) if anyone wants to append that to 121 or add it into the 122 __________________________________________________ Do You Yahoo!? Yahoo! Games - play chess, backgammon, pool and more http://games.yahoo.com/

2 1

Re: Correction for WWN 121
by vinn April 26, 2002

April 26, 2002

> Hello, just had an update for WWN 121... Near the > bottom it says that Alexandre rejected a patch for the Dustin - I'll make the change and get it into 121. Sorry about that, I summarized that thread about halfway into it. -brian.

1 0

Name change
by Dustin Navea April 26, 2002

April 26, 2002

Hello, just noticed that my name in WWN121 was posted as Speeddymon, I forgot that I had set it to that so many many years ago... This is my real name and this is what I will be posting under from now on... :) __________________________________________________ Do You Yahoo!? Yahoo! Games - play chess, backgammon, pool and more http://games.yahoo.com/

1 0

Re: Patch acceptance notification?
by Francois Gouget April 26, 2002

April 26, 2002

On Fri, 26 Apr 2002, Speeddymon wrote: > If we submit a patch, and it is accepted (and > commited) it would be nice to know if any of my > patches will be in the next snapshot... If it is committed to CVS, it will be in the next Wine snapshot. This is automatic and systematic. The only thing that could 'prevent it' is if a huge bug was found between the commit and the snapshot release and the patch ended up being reverted, but that is pretty rare and cannot be forseen. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ May your Tongue stick to the Roof of your Mouth with the Force of a Thousand Caramels.

2 1

winsock2 unit test update
by Martin Wilck April 26, 2002

April 26, 2002

A patch for the winsock 2 unit test posted yesterday. It now has an overlapped client. If you look at the code, it does pretty nasty things to wine, and still works :-) BUGS: probably doesn't compile under Windows, I'm using some gcc extensions. You just get too used to them ... If somebody was able to port these to Windows and test the outcome I would be really happy. Martin diff -ruNX ignore CVS/wine/dlls/winsock/tests/sock.c TMP/wine/dlls/winsock/tests/sock.c --- CVS/wine/dlls/winsock/tests/sock.c Fri Apr 26 20:58:26 2002 +++ TMP/wine/dlls/winsock/tests/sock.c Fri Apr 26 20:56:53 2002 @@ -26,7 +26,6 @@ #include <winsock2.h> #define MAX_CLIENTS 4 /* Max number of clients */ -#define NUM_TESTS 2 /* Number of tests performed */ #define FIRST_CHAR 'A' /* First character in transferred pattern */ #define BIND_SLEEP 10 /* seconds to wait between attempts to bind() */ #define BIND_TRIES 6 /* Number of bind() attempts */ @@ -86,6 +85,7 @@ test_params *general; DWORD sock_flags; int buflen; + int timeout; } client_params; /* This type combines all information for setting up a test scenario */ @@ -104,6 +104,7 @@ SOCKET s; struct sockaddr_in addr; sock_info sock[MAX_CLIENTS]; + int id; } server_memory; /* Thread local storage for client */ @@ -113,6 +114,7 @@ struct sockaddr_in addr; char *send_buf; char *recv_buf; + int id; } client_memory; /**************** Static variables ***************/ @@ -237,6 +239,7 @@ server_memory *mem = (LPVOID) LocalAlloc ( LPTR, sizeof (server_memory)); TlsSetValue ( tls, mem ); + mem->id = GetCurrentThreadId (); mem->s = WSASocketA ( AF_INET, gen->sock_type, gen->sock_prot, NULL, 0, par->sock_flags ); ok ( mem->s != INVALID_SOCKET, "Server: WSASocket failed" ); @@ -267,7 +270,7 @@ if ( mem->sock[i].s != INVALID_SOCKET ) closesocket ( mem->sock[i].s ); } - ok ( closesocket ( mem->s ) == 0, "closesocket failed" ); + ok ( closesocket ( mem->s ) == 0, "server (%x): closesocket failed", mem->id ); LocalFree ( (HANDLE) mem ); ExitThread ( GetCurrentThreadId () ); } @@ -283,6 +286,7 @@ WaitForSingleObject ( server_ready, INFINITE ); + mem->id = GetCurrentThreadId (); mem->s = WSASocketA ( AF_INET, gen->sock_type, gen->sock_prot, NULL, 0, par->sock_flags ); @@ -310,6 +314,39 @@ ExitThread ( GetCurrentThreadId () ); } +static SOCKET do_async_connect ( client_memory *mem ) +{ + HANDLE event; + WSANETWORKEVENTS wsa_events; + int tmp, err; + + event = WSACreateEvent (); + + WSAEventSelect ( mem->s, event, FD_CONNECT ); + + tmp = connect ( mem->s, (struct sockaddr*) &mem->addr, sizeof ( mem->addr ) ); + if ( tmp != 0 && ( err = WSAGetLastError () ) != WSAEWOULDBLOCK ) + ok ( 0, "do_async_connect (%x): connect error: %d", mem->id, err ); + + tmp = WaitForSingleObject ( event, INFINITE ); + ok ( tmp == WAIT_OBJECT_0, + "do_async_connect (%x): wait for connect event failed: %d", mem->id, tmp ); + err = WSAEnumNetworkEvents ( mem->s, event, &wsa_events ); + wsa_ok ( err, 0 ==, "do_async_connect (%x): WSAEnumNetworkEvents error: %d\n" ); + + err = wsa_events.iErrorCode[ FD_CONNECT_BIT ]; + ok ( err == 0, "do_async_connect (%x): connect error: %d", mem->id, err ); + + WSAEventSelect ( mem->s, (HANDLE) 0, 0 ); + WSACloseEvent ( event ); + + if ( err ) + return INVALID_SOCKET; + + trace ( "do_async_connect (%x) connected\n", mem->id ); + return mem->s; +} + /**************** Servers ***************/ /* @@ -323,11 +360,10 @@ id = GetCurrentThreadId(); char *p; - trace ( "simple_server (%x) starting\n", id ); - set_so_opentype ( FALSE ); /* non-overlapped */ server_start ( par ); mem = TlsGetValue ( tls ); + id = mem->id; wsa_ok ( set_blocking ( mem->s, TRUE ), 0 ==, "simple_server (%x): failed to set blocking mode: %d"); wsa_ok ( listen ( mem->s, SOMAXCONN ), 0 ==, "simple_server (%x): listen failed: %d"); @@ -352,7 +388,8 @@ ok ( n_recvd == n_expected, "simple_server (%x): received less data then expected: %d of %d", id, n_recvd, n_expected ); p = test_buffer ( mem->sock[0].buf, gen->chunk_size, gen->n_chunks ); - ok ( p == NULL, "simple_server (%x): test pattern error: %d", p - mem->sock[0].buf); + ok ( p == NULL, "simple_server (%x): test pattern error at %d; %c", + id, p - mem->sock[0].buf, (p ? *p : '*')); /* Echo data back */ n_sent = do_synchronous_send ( mem->sock[0].s, mem->sock[0].buf, n_expected, par->buflen ); @@ -381,16 +418,15 @@ int n_sent, n_recvd, n_expected = gen->n_chunks * gen->chunk_size, id; char *p; - id = GetCurrentThreadId(); - trace ( "simple_client (%x): starting\n", id ); /* wait here because we want to call set_so_opentype before creating a socket */ WaitForSingleObject ( server_ready, INFINITE ); - trace ( "simple_client (%x): server ready\n", id ); check_so_opentype (); set_so_opentype ( FALSE ); /* non-overlapped */ client_start ( par ); mem = TlsGetValue ( tls ); + id = mem->id; + trace ( "simple_client (%x) started\n", id ); /* Connect */ wsa_ok ( connect ( mem->s, (struct sockaddr*) &mem->addr, sizeof ( mem->addr ) ), @@ -414,7 +450,8 @@ /* check data */ p = test_buffer ( mem->recv_buf, gen->chunk_size, gen->n_chunks ); - ok ( p == NULL, "simple_client (%x): test pattern error: %d", id, p - mem->recv_buf); + ok ( p == NULL, "simple_client (%x): test pattern error at %d: %c", + id, p - mem->recv_buf, (p ? *p : '*')); /* cleanup */ read_zero_bytes ( mem->s ); @@ -429,37 +466,21 @@ { test_params *gen = par->general; client_memory *mem; - int id = GetCurrentThreadId(), n_expected = gen->n_chunks * gen->chunk_size, - tmp, err, n; - HANDLE event; + int n_expected = gen->n_chunks * gen->chunk_size, id, err, n; + HANDLE event = 0; WSANETWORKEVENTS wsa_events; char *send_last, *recv_last, *send_p, *recv_p; long mask = FD_READ | FD_WRITE | FD_CLOSE; - trace ( "event_client (%x): starting\n", id ); client_start ( par ); - trace ( "event_client (%x): server ready\n", id ); - mem = TlsGetValue ( tls ); + id = mem->id; + trace ( "event_client (%x) started\n", id ); - /* Prepare event notification for connect, makes socket nonblocking */ - event = WSACreateEvent (); - WSAEventSelect ( mem->s, event, FD_CONNECT ); - tmp = connect ( mem->s, (struct sockaddr*) &mem->addr, sizeof ( mem->addr ) ); - if ( tmp != 0 && ( err = WSAGetLastError () ) != WSAEWOULDBLOCK ) - ok ( 0, "event_client (%x): connect error: %d", id, err ); - - tmp = WaitForSingleObject ( event, INFINITE ); - ok ( tmp == WAIT_OBJECT_0, "event_client (%x): wait for connect event failed: %d", id, tmp ); - err = WSAEnumNetworkEvents ( mem->s, event, &wsa_events ); - wsa_ok ( err, 0 ==, "event_client (%x): WSAEnumNetworkEvents error: %d\n" ); - - err = wsa_events.iErrorCode[ FD_CONNECT_BIT ]; - ok ( err == 0, "event_client (%x): connect error: %d", id, err ); - if ( err ) goto out; - - trace ( "event_client (%x) connected\n", id ); + if ( do_async_connect ( mem ) == INVALID_SOCKET ) + goto out; + event = WSACreateEvent (); WSAEventSelect ( mem->s, event, mask ); recv_p = mem->recv_buf; @@ -470,7 +491,7 @@ while ( TRUE ) { err = WaitForSingleObject ( event, INFINITE ); - ok ( err == WAIT_OBJECT_0, "event_client (%x): wait failed", id, tmp ); + ok ( err == WAIT_OBJECT_0, "event_client (%x): wait failed: %d", id, err ); err = WSAEnumNetworkEvents ( mem->s, event, &wsa_events ); wsa_ok ( err, 0 ==, "event_client (%x): WSAEnumNetworkEvents error: %d\n" ); @@ -535,20 +556,183 @@ } ok ( send_p == send_last, - "simple_client (%x): sent less data then expected: %d of %d", + "event_client (%x): sent less data then expected: %d of %d", id, send_p - mem->send_buf, n_expected ); ok ( recv_p == recv_last, - "simple_client (%x): received less data then expected: %d of %d", + "event_client (%x): received less data then expected: %d of %d", id, recv_p - mem->recv_buf, n_expected ); recv_p = test_buffer ( mem->recv_buf, gen->chunk_size, gen->n_chunks ); - ok ( recv_p == NULL, "event_client (%x): test pattern error: %d", id, recv_p - mem->recv_buf); + ok ( recv_p == NULL, "event_client (%x): test pattern error at %d: %c", + id, recv_p - mem->recv_buf, (recv_p ? *recv_p : '*')); out: - WSACloseEvent ( event ); + if ( event ) WSACloseEvent ( event ); trace ( "event_client (%x) exiting\n", id ); client_stop (); } +/* + * ReadFile_client: A client using overlapped IO with ReadFile() / WriteFile() + */ +static void WINAPI ReadFile_client ( client_params *par ) +{ + test_params *gen = par->general; + client_memory *mem; + int id, n_expected = gen->n_chunks * gen->chunk_size, n_requests; + int tmp, err, i, sent_total = 0, recvd_total = 0; + char *send_last, *recv_last, *send_p, *recv_p; + LPHANDLE event = NULL, send_event, recv_event; + LPOVERLAPPED send_over, recv_over; + LPDWORD n_sent, n_recvd; + + client_start ( par ); + mem = TlsGetValue ( tls ); + id = mem->id; + trace ( "ReadFile_client (%x): started\n", id ); + + if ( do_async_connect ( mem ) == INVALID_SOCKET ) + goto out; + + if ( n_expected % par->buflen == 0 ) + n_requests = n_expected / par->buflen; + else + n_requests = n_expected / par->buflen + 1; + + event = (LPVOID) LocalAlloc ( LPTR, ( 1 + 2 * n_requests ) * sizeof ( HANDLE ) ); + send_over = (LPVOID) LocalAlloc ( LPTR, 2 * n_requests * sizeof ( OVERLAPPED ) ); + n_sent = (LPVOID) LocalAlloc ( LPTR, 2 * n_requests * sizeof ( DWORD ) ); + + send_event = event + 1; + recv_event = send_event + n_requests; + recv_over = send_over + n_requests; + n_recvd = n_sent + n_requests; + + event[0] = WSACreateEvent (); + WSAEventSelect ( mem->s, event[0], FD_CLOSE ); + + for ( i = 0; i < 2 * n_requests; i++ ) + event[ 1 + i ] = CreateEventA ( NULL, TRUE, FALSE, NULL ); + + /* Brutal: queue all requests right away */ + send_last = mem->send_buf + n_expected; + recv_last = mem->recv_buf + n_expected; + for ( i = 0, send_p = mem->send_buf, recv_p = mem->recv_buf; + i < n_requests; + i++, send_p += par->buflen, recv_p += par->buflen ) + { + tmp = min ( par->buflen, send_last - send_p ); + send_over[i].hEvent = send_event[i]; + err = WriteFile ( (HANDLE) mem->s, send_p, tmp, &n_sent[i], &send_over[i] ); + if ( !err ) tmp = GetLastError (); + ok ( err == TRUE || tmp == ERROR_IO_PENDING, + "ReadFile_client (%x): WriteFile error: %d", id, tmp ); + + tmp = min ( par->buflen, recv_last - recv_p ); + recv_over[i].hEvent = recv_event[i]; + err = ReadFile ( (HANDLE) mem->s, recv_p, tmp, &n_recvd[i], &recv_over[i] ); + if ( !err ) tmp = GetLastError (); + ok ( err == TRUE || tmp == ERROR_IO_PENDING, + "ReadFile_client (%x): ReadFile error: %d", id, tmp ); + } + trace ( "ReadFile_client (%x): all requests queued\n", id ); + shutdown ( mem->s, SD_SEND ); + + /* Wait for the FD_CLOSE event */ + err = WaitForSingleObject ( event[0], par->timeout ); + + ok ( err == WAIT_OBJECT_0 || err == WAIT_TIMEOUT, + "ReadFile_client (%x): error waiting for FD_CLOSE: %d", id, err ); + + if ( err == WAIT_OBJECT_0 ) + trace ( "ReadFile_client (%x): close event\n", id ); + else if ( err == WAIT_TIMEOUT ) + trace ( "ReadFile_client (%x): timeout\n", id ); + + /* Wait for read results only if TIMEOUT received - indicates not all data could be read */ + for ( i = 0, recv_p = mem->recv_buf; i < n_requests; i++, recv_p += par->buflen ) + { + err = GetOverlappedResult ( (HANDLE) mem->s, &send_over[i], &n_sent[i], FALSE ); + ok ( err == TRUE, "ReadFile_client (%x): write request %d error %d", + id, i, GetLastError () ); + sent_total += n_sent[i]; + + err = GetOverlappedResult ( (HANDLE) mem->s, &recv_over[i], &n_recvd[i], err == WAIT_TIMEOUT ); + ok ( err == TRUE, "ReadFile_client (%x): read request %d error %d", + id, i, GetLastError () ); + + /* If a read did not get all requested bytes, rearrange data */ + if ( mem->recv_buf + recvd_total < recv_p ) + memmove ( mem->recv_buf + recvd_total, recv_p, n_recvd[i] ); + recvd_total += n_recvd[i]; + } + + /* Try to read remaining data, if any */ + while ( recvd_total < n_expected ) + { + trace ( "ReadFile_client (%x): %d bytes missing, retrying ...\n", id, n_expected - recvd_total ); + recv_p = mem->recv_buf + recvd_total; + ResetEvent ( recv_over[0].hEvent ); + err = ReadFile ( (HANDLE) mem->s, recv_p, recv_last - recv_p, &n_recvd[0], &recv_over[0] ); + if ( err == FALSE ) + { + err = GetLastError (); + ok ( err == ERROR_IO_PENDING, "ReadFile_client (%x): ReadFile(2) error: %d", id, err ); + if ( err != ERROR_IO_PENDING ) break; + err = GetOverlappedResult ( (HANDLE) mem->s, &recv_over[0], &n_recvd[0], TRUE ); + ok ( err == TRUE, "ReadFile_client (%x): read request error %d", + id, GetLastError () ); + if ( err == FALSE ) break; + if ( n_recvd[0] == 0 ) + { + trace ( "ReadFile_client (%x): empty receive (overlapped)\n", id ); + break; + } + recvd_total += n_recvd[0]; + } + else + { + if ( n_recvd[0] == 0 ) + { + trace ( "ReadFile_client (%x): empty receive (immediate)\n", id ); + break; + } + recvd_total += n_recvd[0]; + } + } + + /* Wait for the FD_CLOSE event again - it should be signalled now */ + err = WaitForSingleObject ( event[0], par->timeout ); + + ok ( err == WAIT_OBJECT_0, + "ReadFile_client (%x): error waiting for FD_CLOSE: %d", id, err ); + if ( err == WAIT_OBJECT_0 ) + trace ( "ReadFile_client (%x): close event (2)\n" ); + + /* Check IO results */ + ok ( sent_total == n_expected, + "ReadFile_client (%x): sent less data then expected: %d of %d", + id, sent_total, n_expected ); + + ok ( recvd_total == n_expected, + "ReadFile_client (%x): received less data then expected: %d of %d", + id, recvd_total, n_expected ); + + recv_p = test_buffer ( mem->recv_buf, gen->chunk_size, gen->n_chunks ); + ok ( recv_p == NULL, + "ReadFile_client (%x): test pattern error at %d: %c", + id, recv_p - mem->recv_buf, (recv_p ? *recv_p : '*')); + + for (i = 0; i < 1 + 2 * n_requests; i++) + WSACloseEvent ( event[i] ); + + LocalFree ( (HANDLE)event ); + LocalFree ( (HANDLE)send_over ); + LocalFree ( (HANDLE)n_sent ); +out: + trace ( "ReadFile_client (%x) exiting\n", id ); + client_stop (); +} + /**************** Main program utility functions ***************/ static void Init (void) @@ -627,7 +811,7 @@ inet_addr: "127.0.0.1", \ inet_port: 9374 -static test_setup tests [NUM_TESTS] = +static test_setup tests [] = { /* Test 0: synchronous client and server */ { @@ -662,6 +846,25 @@ clt_params: { buflen: 128 } + }, + /* Test 2: ReadFile() overlapped client, synchronous server */ + { + general: { + STD_STREAM_SOCKET, + chunk_size: 2084, + n_chunks: 16, + n_clients: 2 + }, + srv: simple_server, + srv_params: { + buflen: 64 + }, + clt: ReadFile_client, + clt_params: { + sock_flags: WSA_FLAG_OVERLAPPED, + buflen: 128, + timeout: 1000 + } } }; @@ -670,9 +873,10 @@ START_TEST( sock ) { int i; - Init(); + int n_tests = sizeof (tests) / sizeof (test_setup); - for (i = 0; i < NUM_TESTS; i++) + Init(); + for (i = 0; i < n_tests; i++) { trace ( " **** STARTING TEST %d **** \n", i ); do_test ( &tests[i] );

1 0

Improved FD_CLOSE scheme for sockets
by Martin Wilck April 26, 2002

April 26, 2002

Just a small patch to the server sock.c - it tries to implement exactly the strategies that I described in my previous FD_CLOSE posting. Seems to work pretty well with my test cases. It makes a difference to my previous implementation only in rare cases, but I am sure it will be better at preventing server busy-loops in these cases. Martin diff -ruNX ignore CVS/wine/server/sock.c TMP/wine/server/sock.c --- CVS/wine/server/sock.c Fri Apr 26 20:58:26 2002 +++ TMP/wine/server/sock.c Fri Apr 26 20:36:44 2002 @@ -188,8 +188,6 @@ /* We need this to delay FD_CLOSE events until all pending overlapped requests are processed */ if ( !events || async_active ) return; - if (events & FD_CLOSE) sock->hmask |= FD_CLOSE; - if (sock->event) { if (debug_level) fprintf(stderr, "signalling events %x ptr %p\n", events, sock->event ); @@ -284,7 +282,7 @@ { /* incoming data */ sock->pmask |= FD_READ; - sock->hmask |= FD_READ; + sock->hmask |= (FD_READ|FD_CLOSE); sock->errors[FD_READ_BIT] = 0; if (debug_level) fprintf(stderr, "socket %d is readable\n", sock->obj.fd ); @@ -330,6 +328,7 @@ if ( event & ( POLLERR|POLLHUP ) ) sock->state &= ~(FD_WINE_CONNECTED|FD_WRITE); sock->pmask |= FD_CLOSE; + sock->hmask |= FD_CLOSE; if (debug_level) fprintf(stderr, "socket %d aborted by error %d, event: %x - removing from select loop\n", sock->obj.fd, sock->errors[FD_CLOSE_BIT], event); @@ -390,7 +389,7 @@ if (mask & FD_WRITE || (sock->flags & WSA_FLAG_OVERLAPPED && IS_READY (sock->write_q))) ev |= POLLOUT; /* We use POLLIN with 0 bytes recv() as FD_CLOSE indication. */ - if (sock->mask & ~sock->hmask & FD_CLOSE && !(sock->hmask & FD_READ) ) + if (sock->mask & ~sock->hmask & FD_CLOSE) ev |= POLLIN; return ev; @@ -447,6 +446,7 @@ { case ASYNC_TYPE_READ: q = &sock->read_q; + sock->hmask &= ~FD_CLOSE; break; case ASYNC_TYPE_WRITE: q = &sock->write_q; @@ -478,7 +478,7 @@ async_insert ( q, async ); } } - else if ( async ) destroy_async ( async ); + else if ( async ) destroy_async ( async ); else set_error ( STATUS_INVALID_PARAMETER ); pollev = sock_reselect ( sock ); @@ -788,6 +788,8 @@ sock->pmask &= ~req->mask; /* is this safe? */ sock->hmask &= ~req->mask; + if ( req->mask & FD_READ ) + sock->hmask &= ~FD_CLOSE; sock->state |= req->sstate; sock->state &= ~req->cstate;

1 0

Re: wine/ dlls/winsock/socket.c dlls/winsock/ws2_3 ...
by Martin Wilck April 26, 2002

April 26, 2002

Alexandre, -static int FILE_GetUnixHandleType( HANDLE handle, DWORD access, enum fd_type *type, int *flags ) +static int FILE_GetUnixHandleType( HANDLE handle, DWORD access, enum fd_type *type, int *flags_ptr ) { - int ret, fd = -1; + int ret, flags, fd = -1; - ret = wine_server_handle_to_fd( handle, access, &fd, type, flags ); + ret = wine_server_handle_to_fd( handle, access, &fd, type, &flags ); + if (flags_ptr) *flags_ptr = flags; This is of course the most robust solution - but because FILE_GetUnixHandleType is internal, you might as well require all functions calling it to pass a valid flags pointer, and let FILE_GetUnixHandle call wine_server_handle_to_fd directly, at a minimum performance gain. Martin -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck(a)Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

2 2