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

Re: More checks when loading the driver
by Eric Pouech May 3, 2003

May 3, 2003

Uwe Bonnes wrote: > Changelog: > wine/dlls/winmm/lolvldrv.c: MMDRV_Install > Check for failing DRIVER_FindFromHDrvr this doesn't seem right to me: either the loading fails and we shouldn't get a valid handle, or it succeeds and we should be able to find it there's must be something wrong elsewhere could you please send a -debugmsg +driver,+module,+winmm to see what happens ? TIA -- Eric Pouech

2 2

winsock and virtual interfaces
by Juan Lang May 3, 2003

May 3, 2003

I'm planning to patch winsock (both ws2_32 and wsock32 varieties) so that virtual interfaces will be included in response to IP address/interface list queries. If you'd like different behavior, please speak up (though it's up to Alexandre anyway :) ) More detail: I'm patching wsock32 and winsock to use iphlpapi.dll to enumerate interfaces, get the route table, and that sort of thing. Less code duplication, and I'm trying to isolate the plaform-specific network code to fewer files. (If any of you saw my earlier call for help, the stack corruption was due to socket() mismatching.) Yet more detail: In writing iphlpapi.dll, I was curious about the code in wsock32 and winsock that tries to filter out virtual interface names (in WSAIOCTL_GetInterfaceName()). So, I tested a little bit with my Linux system, and it seems this code does not behave as desired: On my system, virtual interfaces don't show up in /proc/net/dev, so the filtering for them seems a little pointless. But if they did show up, the filtering seems counterproductive: Say you had three interfaces, lo, eth0, and eth0:0. You'd be told (by WSAIOCTL_GetInterfaceCount()) that you have three interfaces, but if you got the names of them, you'd only get two unique names. Furthermore, since the filtered name is what's used for getting the IP address, MAC address, and whatnot, in fact you'd end up with duplicate entries in responses to queries. I've attached my little test program; try it with and without virtual interfaces to see what I mean. I also tested how Win2K behaves with multiple IP addresses on an interface. In IP address tables, there's one entry per IP address, as expected. In GetAdaptersInfo, the IP address list has every configured address for the adapter, also I suppose as expected. So, in Win2K, an interface (uniquely identified by an index) can have 0..N IP addresses. I've attached some output from a test program on Win2K to illustrate. I don't emulate this behavior. Instead, I let an interface have 0 or 1 IP address, and treat virtual interfaces as unique interfaces, with their own index. This makes the code rather simpler. If anyone has a strong opinion otherwise, please speak up. (Patches coming this week. iphlpapi.dll is working, and wsock32 is using it to implement WsControl now. I want to change iphlpapi a little bit to make it more appropriate for netapi32 before I submit.) Thanks, --Juan __________________________________ Do you Yahoo!? The New Yahoo! Search - Faster. Easier. Bingo. http://search.yahoo.com GetIpAddrTable returned 3 entries: GetIpAddrTable entry for adapter index 16777219: ip 192.168.1.2 mask 255.255.255.0 broadcast 1.0.0.0 reasm size 65535 GetIpAddrTable entry for adapter index 16777219: ip 10.0.0.1 mask 255.255.255.0 broadcast 1.0.0.0 reasm size 65535 GetIpAddrTable entry for adapter index 1: ip 127.0.0.1 mask 255.0.0.0 broadcast 1.0.0.0 reasm size 65535 GetIfEntry returned 2 entries: GetIfEntry for adapter index 16777219: name: description: Wireless-G Notebook Adapter Type: ethernet physical address: 00:06:25:3d:b5:58 GetIfEntry for adapter index 1: name: description: MS TCP Loopback interface Type: loopback physical address: See 1 adapters in IP_INTERFACE_INFO Adapter 16777219's name is \DEVICE\TCPIP_{66B081D9-F4E3-4FE8-85CC-0934F6134D44} GetAdaptersInfo output: Name: {66B081D9-F4E3-4FE8-85CC-0934F6134D44} Description: Wireless-G Notebook Adapter MAC addr: 00:06:25:3d:b5:58 Index: 16777219 Type: ethernet IP addresses: 10.0.0.1/255.255.255.0 192.168.1.2/255.255.255.0 Gateways: 192.168.1.1

1 0

iexplore 5.5.
by KK singh May 3, 2003

May 3, 2003

hi, I have tried ie5 installation on wine and failed, I am trying to find workarounds for not using IE. Given a windows executable, can i detect which IE dependent calls it is making ? Or exactly what calls both direct api's and COM Interfaces are successful if IE5.5 is installed and fails otherwise. any pointers shall be helpful to me... Regards -kks PS: thanx for the earlier replies. __________________________________ Do you Yahoo!? The New Yahoo! Search - Faster. Easier. Bingo. http://search.yahoo.com

2 1

With-* test in configure (was: wine and RedHat 9)
by Sylvain Petreolle May 3, 2003

May 3, 2003

Oh, sorry. The 'with-' is a special case, as a test is done on with-* to only reject invalid shell variable names. I dont know if this should be fixed into autoconf or our configure code. Cc'ing your request to wine-devel. --- Rob Hughes <rob(a)robhughes.com> a écrit : > You have to type --with-lptn or whatever. It doesn't check it, > verified > here on Sunday's CVS source. > > I've also noticed a lot of dialogs showing up in a couple different > languages, but I suppose that's a different issue. do you have examples ? Its easy to fix as they often are untranslated messages. > > On Thu, 2003-05-01 at 11:15, Sylvain Petreolle wrote: > > --- Patrick R Griffin <griffinp(a)MIT.EDU> a écrit : > > > "./configure --with-nptl" is correct. > > > As Rob noticed, though, typing in "./configure --with-ltpn" or some > > > other typo will compile successfully - no error message - but > > > will not include the threading library corrections. > > If this works call me Joan of Arc :) > > > > [syl(a)snoop wine]$ ./configure --ltpn > > configure: error: unrecognized option: --ltpn > > Try `./configure --help' for more information. ===== Sylvain Petreolle (spetreolle at users dot sourceforge dot net) ICQ #170597259 No more War ! "What if tomorrow the War could be over ?" Morpheus, in "Reloaded". For the Law of Oil and Fire, Im an European that lives in France. For all my Brothers and friends, Im a human living on Earth. ___________________________________________________________ Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Yahoo! Mail : http://fr.mail.yahoo.com

3 8

Re: Small DSound fix...
by Christian Costa May 3, 2003

May 3, 2003

Lionel Ulmer wrote: >Hi all, > >This makes games which do NOT enumerate the sound devices and go through the >ClassFactory work. > >It has also some miscellaneous tracing fixes. > >Changelog: > - misc. tracing fixes > - handle IID_IDirectSound(8) as the default sound driver > A ce que je vois ton patch remplace le mien (qui n'a pas ete accepte d'ailleurs)... P'tite curiosite.. :-) : Mais qu'est ce que ces foutues IIDs viennent faire a la place des GUID des devices ? C'est une bidouille ou bien DirectSoundCreate les acceptent vraiment ? Christian

4 5

Volunteers needed - windows API db script ready
by Dave Miller May 2, 2003

May 2, 2003

The script which dumps dll imports and exports now seems to parse everything correctly and organize it nicely in a text file. We need the script run against various windows versions. The output can be sent directly to me, but please verify I do not already have the data for that particular version of windows. I don't need 10 copies of win2k SP3 for example. :) The script is available at http://home.ptd.net/~compsol. Please make sure you get createdb.pl. There is another script, createwinedb.pl which will ONLY parse the wine spec files. To run this you will need winedump, perl, and access to your windows directory from within Linux. The cleaner the windows install the better. We don't want to clutter the database with info from dlls not provided with windows. Usage is as follows: createdb.pl -d -p <path to windows dir> -o <winver (ex. win98) This will save the imports and exports and also create winver.ref, where winver is specified on the command line with -o. The .ref file is the one which will contain all the info we need. This will be a large file, so I recommend compressing with gzip if you are going to email it to me. It may be large enough to bounce back into your mailbox if you do not. :)

5 4

Re: A little Xlib adventure
by Mike Hearn May 2, 2003

May 2, 2003

Any problems with this patch Alexandre? It's missed 3 rounds of commits now. thanks -mike On Tue, 2003-04-29 at 13:59, Mike Hearn wrote: > Got to start somewhere. This patch adds support for the netWM ping > protocol, which lets window managers like metacity present a nice "This > app is not responding, kill it?" window when the program freezes. This > unfortunately happens all too often with IE in particular, and while us > developers can see (and understand) critical section timeout messages, > people running from the menu or desktop icons cannot. > > I haven't bothered sending a message to the window (WM_NULL) because if > the program stops responding, it'll stop processing X messages as well. > If you want to test it out, just put an infinite loop in a test app then > try and close it. > > Tested with metacity from gnome2.2 - may work with latest KWin, if it > doesn't then file a bug in KDE bugzilla. > > ChangeLog: > Support _NET_WM_PING protocol so the WM can detect freezes

1 0

Re: Stub for SetSystemTimeAdjustment
by Dimitrie O. Paun May 2, 2003

May 2, 2003

On Fri, 2 May 2003, Lionel Ulmer wrote: > Try the attached patch, it will work MUCH better :-) Please add a FIXME with a stub message, so we don't forget about it... -- Dimi.

1 0

wine debugger and Agent by Forte Inc
by schuetzen May 2, 2003

May 2, 2003

running Forte Agent in Wine. works GREAT! except that almost every time I send a message or reply from the s/w, the Wine debugger runs and I have to shut the software down and then restart it. this is aggravating when I am running a dozen elists which are moderated. the message in the end of the debug script is Unhandled exception: divide by zero in 32 bit code (0X0043e44), In 32-bit mode. 0X0043e44 (agent.exe.(a)ENUMERATEACTIVEWNDS$QP6HWND_P (wrapping line here) 9TENUMINFO+0xa9b4 in C:\Program Files\Agent\agent.exe): div1 0x4(%ebx),%eax ===== Agent was downloaded from the forte.inc website and then installed into Linux (thisis a new and fresh install on a new hard drive) with Suse starting up Wine and installing Wine and then installing Agent. all automatically. really need help with this -- powered by Linux, Suse 8.2 Pro + Agent under Wine emulator Charles L Hamilton, Houston, TX chasm @ texas. net www.schuetzen.net

1 0

Re: Avoid non-portable signal(SIGCHLD, SIG_IGN)
by Alexandre Julliard May 2, 2003

May 2, 2003

Joerg Wunsch <j(a)uriah.heep.sax.de> writes: > The Posix way is to use sigaction(), and set the SA_NOCLDWAIT flag > bit. This variant is supported under FreeBSD (don't know for the > other BSDs). It has the added bonus that unavailability of the > automatic zombie reaping feature can be detected at compile-time (the > macro is not defined then). The fact that the macro exists doesn't mean it works. Most Linux kernels don't support SA_NOCLDWAIT, even though the symbol may be defined in the headers. I don't have the Posix standard, but at least according to the single Unix spec using SIG_IGN is just as standard as SA_NOCLDWAIT, and a quick grep of the FreeBSD sources seems to indicate that it should work. Did I miss something? -- Alexandre Julliard julliard(a)winehq.com

3 12