December 2007 - Wine-Devel - list.winehq.org

Re: wine 0.9.5 IO completion port questions
by Andrey Turkin Dec. 7, 2007

Dec. 7, 2007

Marius Slyzius wrote: > > Hello all. > > > > I am trying to run ScimoreDB database which uses fibers, non-blocked > IO for sockets and files, and completion port. It did started ok > under 0.9.5 and seems it is working!!!:), so I would like to verify > few things about async IO with completion ports before recompiling it > with winelib. > You mean 0.9.50, right? > > > > 1) will write/read file/socket with completion port is true > asynchronous operation. I am asking because every time write/read > operation returns true, and never IO pending error. > Not sure it is asynchronous for files, should be asynchronous for sockets. > > > > 2) possible bug: the server sends 1500 bytes with WsaSend ( with 2 > WSABUFFER ) and the client receive only 1452 bytes and calls WsaRecv > again. Also on the server, the completion port receive event that > transferred 1452 bytes. However, the last 8 bytes are never sends nor > completion port event arrives. Is it perhaps something to do with the > VMWare tool where I run Ubuntu 7.10? > WsaSend can send partial data when operating on non-blocking socket; from your description it looks like this is the case. Is this error reproducible with the same numbers every run? Your both questions are actually not about IOCP but about asynchronous I/O support. > > Please, send me an info also to my mail: marius(a)scimore.com , because > I might miss the answer here. > > > > Thank you! > > Marius Slyzius > > Scimore UAB > > (+45) 40862122 > > > > ------------------------------------------------------------------------ > > >

3 2

replying to wine-devel?
by King InuYasha Dec. 7, 2007

Dec. 7, 2007

I have noticed some unusual behavior in the wine-devel mailing list compared to others (e.g. fedora's, etc.). Wine's wine-devel mailing list does not automatically set the reply to wine-devel(a)winehq.org. Why is that? If other mailing lists do so, how come Wine's doesn't? Personally, I find myself making a mistake sometimes and not adding in the wine-devel part because I am used to seeing it automatically set. If I remember how emails are set up, isn't there a hidden field called reply-to? AFAICT, other mailing lists either override or rewrite this field and replace it with the email address of the mailing list.

6 8

Re: [5/5] WineD3D: Use the disabled opengl extension string
by Chris Robinson Dec. 7, 2007

Dec. 7, 2007

On Thursday 06 December 2007 02:05:52 pm Stefan Dösinger wrote: > + if(wined3d_settings.disabled_extensions && > + strstr(wined3d_settings.disabled_extensions, > ThisExtn)) { You need to be careful with this. I read somewhere that using strstr to check for an opengl extension in an extension list is discourage because it can find false positives. Especially in this case, if you have, say: wined3d_settings.disabled_extensions = "GL_NV_texture_shader3 GL_NV_vertex_program2_option GL_NV_vertex_program3" Then if you check for "GL_NV_texture_shader", "GL_NV_vertex_program2", or "GL_NV_vertex_program", they'll all show up as being disabled, when they're not. Which is a double problem because "GL_NV_vertex_program1_1" would still be enabled, which requires "GL_NV_vertex_program". AFAIK, the proper way to check for OpenGL extensions in a space-delimitted list is like this: const char *ptr = strstr(wined3d_settings.disabled_extensions, ThisExtn); if(ptr) { if(ptr == wined3d_settings.disabled_extensions || isspace(ptr[-1])) { int len = strlen(ThisExtn); if(!ptr[len] || isspace(ptr[len])) { /* is disabled */ } } }

1 0

Re: d3dx8 [2/2]: add basic functions for MatrixStack
by Stefan Dösinger Dec. 6, 2007

Dec. 6, 2007

Am Mittwoch, 5. Dezember 2007 18:13:30 schrieb David.Adam(a)math.cnrs.fr: > This is my first attempt with ICOM objects. So, all the advices are > welcome. A few suggestions: +HRESULT WINAPI ID3DXMatrixStackImpl_QueryInterface(ID3DXMatrixStack *iface, REFIID riid, void **ppvObject) + ... + IClassFactory_AddRef(iface); It is better to use ID3DXMatrixStack_AddRef(iface) here. It is essentially the same, but it looks more consistent The AddRef() and Release() methods should have a TRACE which write either the new or old refcount, otherwise tracking refcounting bugs is hell.

3 3

Re: msxml3: Implement cloneNode (resend)
by Alexandre Julliard Dec. 6, 2007

Dec. 6, 2007

Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> writes: > +static void test_cloneNode(void ) > +{ > + // NOTE TO SELF - r = IXMLDOMNode_selectSingleNode(node, _bstr_("br"), &node); //test Please cleanup that sort of thing before submitting. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: system.sfd: make 18 pixel strike with Japanese jvgasys glyphs
by Dmitry Timoshkov Dec. 6, 2007

Dec. 6, 2007

"Aric Stewart" <aric(a)codeweavers.com> wrote: > fonts/system.sfd | 1407 > ++++++++++++++++++++++++++++++++++++++++++++++-------- > 1 files changed, 1213 insertions(+), 194 deletions(-) For some reason widths of many glyphs are changed by this patch, that doesn't look correct. -- Dmitry.

3 3

Re: quartz: move current tests into its own test case [1/5]
by James Hawkins Dec. 6, 2007

Dec. 6, 2007

On Dec 5, 2007 10:35 PM, Lei Zhang <thestig(a)google.com> wrote: > Hi, > > I'm moving the current tests into their own function in preparation to > write more tests. > When submitting a numbered patch series, can you put the numbers [x/y] right after the module name? That makes it easier to spot the ordering of the patches. Thanks, James Hawkins

1 0

Re: secur32: Remove the SECUR32_ALLOC() macro around HeapAlloc().
by Michael Stefaniuc Dec. 5, 2007

Dec. 5, 2007

Juan Lang wrote: > On Dec 5, 2007 12:52 PM, Michael Stefaniuc <mstefani(a)redhat.de> wrote: >> Checked with Kai and Dmitry and there is nothing special about >> SECUR32_ALLOC(). There are more direct usages of HeapAlloc() >> than of the macro so instead of replacing SECUR32_ALLOC() with >> heap_alloc() just remove it altogether. > > No prob. I introduced those macros some years ago before valgrind was > usable with wine, it made debugging memory leaks (via redefinition of > the macros with special leak-tracking ones) easier. Valgrind is > certainly better. I thought that was still the reason why those $dll_alloc() and now heap_alloc() wrappers are used. I mean the main reason cause we all know that the HeapAlloc() API just sucks. bye michael -- Michael Stefaniuc Tel.: +49-711-96437-199 Sr. Network Engineer Fax.: +49-711-96437-111 -------------------------------------------------------------------- Reg. Adresse: Red Hat GmbH, Hauptstätter Strasse 58, 70178 Stuttgart Handelsregister: Amtsgericht Stuttgart HRB 153243 Geschäftsführer: Brendan Lane, Charlie Peters, Michael Cunningham, Werner Knoblich

1 0

A small question regarding cpuid
by Vijay Kiran Kamuju Dec. 5, 2007

Dec. 5, 2007

Hi, I was going through dlls/kernel32/cpu.c and stumbled upon the functions do_cpuid and have_cpuid. It seems the assembly code for both the Intel and AMD 64 bit processors. Is it possible to remove the ifdefs or add the x86_64 ifdef. And one more question does this change help 64 bit windows support? Thanks, VJ

1 0

Re: wined3d:context.c Check for a context is added
by Stefan Dösinger Dec. 5, 2007

Dec. 5, 2007

Am Mittwoch, 5. Dezember 2007 17:25:55 schrieb Anatoly Lyutin: > We need to check context before than refer it. > > Changelog: > Check for a context is added Does that happen to fix any game? The idea was that the activeContext pointer never points to NULL, so a check like that only workaround a bug somewhere else

1 0

Re: wined3d:context.c Check for a context is added
by Roderick Colenbrander Dec. 5, 2007

Dec. 5, 2007

> We need to check context before than refer it. > > Changelog: > Check for a context is added > > -- > Best regards > Anatoly Lyutin. > Hi, I'm not sure if your patch is correct. The block of code in front of the if-statement you added should have found or else created a context. If afterwards there is no context it sounds like something went wrong. Returning NULL isn't correct then. When we just return we keep using the old context (if one was around) which might work for your program but it isn't guaranteed to work at all. So what is the real problem? Regards, Roderick -- Der GMX SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! Ideal für Modem und ISDN: http://www.gmx.net/de/go/smartsurfer

2 1

Re: dlls/icmp/icmp_main.c -- attempt at fixing regression in r1.30 (RESEND)
by Michael Stefaniuc Dec. 5, 2007

Dec. 5, 2007

Gerald Pfeifer wrote: > On Wed, 5 Dec 2007, Alexandre Julliard wrote: >> Looks ok, please submit the patch with a proper Changelog entry. > > Thanks for looking into this! Here we go! :-) > ChangeLog: > Correctly handle underflow in IcmpSendEcho(); bug introduced in rev 1.30. Still using cvs? I guess nobody else can do anything with rev 1.30. bye michael > Index: dlls/icmp/icmp_main.c > =================================================================== > RCS file: /home/wine/wine/dlls/icmp/icmp_main.c,v > retrieving revision 1.30 > diff -u -3 -p -r1.30 icmp_main.c > --- dlls/icmp/icmp_main.c 6 Nov 2006 11:51:07 -0000 1.30 > +++ dlls/icmp/icmp_main.c 22 Nov 2007 21:28:59 -0000 > @@ -433,8 +433,9 @@ DWORD WINAPI IcmpSendEcho( > * Decrease the timeout so that we don't enter an endless loop even > * if we get flooded with ICMP packets that are not for us. > */ > - Timeout -= (recv_time - send_time); > - if (Timeout < 0) Timeout = 0; > + DWORD t = (recv_time - send_time); > + if (Timeout > t) Timeout -= t; > + else Timeout = 0; > continue; > } else { > /* This is a reply to our packet */ > > -- Michael Stefaniuc Tel.: +49-711-96437-199 Consulting Communications Engineer Fax.: +49-711-96437-111 -------------------------------------------------------------------- Reg. Adresse: Red Hat GmbH, Hauptstätter Strasse 58, 70178 Stuttgart Handelsregister: Amtsgericht Stuttgart HRB 153243 Geschäftsführer: Brendan Lane, Charlie Peters, Michael Cunningham, Werner Knoblich

1 0

Re: [3/5] shell32: fix buffer sizes
by Alexandre Julliard Dec. 5, 2007

Dec. 5, 2007

Vitaly Lipatov <lav(a)etersoft.ru> writes: > lstrcpynW needs size in TCHARS > RegEnumKey needs size in TCHARS It would be better to change the HCR_GetDefaultVerbW function to take a len in WCHARs since it's returning a string in all cases. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: dlls/icmp/icmp_main.c -- attempt at fixing regression in r1.30 (RESEND)
by Alexandre Julliard Dec. 5, 2007

Dec. 5, 2007

Gerald Pfeifer <gerald(a)pfeifer.com> writes: > Since Alexandre has asked me to take care of overflow in some other > context, I figured I could ping this underflow patch in return. ;-) Looks ok, please submit the patch with a proper Changelog entry. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: dlls/icmp/icmp_main.c -- attempt at fixing regression in r1.30 (RESEND)
by Steven Edwards Dec. 5, 2007

Dec. 5, 2007

Hello, On Dec 4, 2007 7:34 PM, Gerald Pfeifer <gerald(a)pfeifer.com> wrote: > Since Alexandre has asked me to take care of overflow in some other > context, I figured I could ping this underflow patch in return. ;-) > > Gerald > > ---------- Forwarded message ---------- > From: Gerald Pfeifer <gerald(a)pfeifer.com> > To: <wine-patches(a)winehq.org>, Damjan Jovanovic <damjan.jov(a)gmail.com> > Date: Thu, 22 Nov 2007 22:32:08 +0100 (CET) > Subject: dlls/icmp/icmp_main.c -- attempt at fixing regression in r1.30 > > I believe revision 1.30 of dlls/icmp/icmp_main.c > > date: 2006-11-06 11:51:07 +0000; author: julliard; state: Exp; lines: +10 -15; > Damjan Jovanovic <damjan.jov(a)gmail.com> > icmp: Changed select to poll. Could we just revert this patch? I think poll is not a valid api on win32 so this breaks the windows port. I don't really see the need in this change here in the first place. Thanks -- Steven Edwards "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo

1 0

dib engine
by Piotr Maceluch Dec. 4, 2007

Dec. 4, 2007

Hi, what's the status of DIB engine (or where can one read about it)? Are the up-to-date sources still available? What kind of help would be needed? Thanks Peter Maceluch -- -----------=== F O R P S I ===----------- Profesjonalne uslugi, rewelacyjne ceny Sprawdz nas - nowa oferta !! Twoja strona w sieci - od 2 zl http://www.forpsi.pl -----------------------------------------

2 1

Re: dlls/wined3d/state.c -- fix error handling
by Stefan Dösinger Dec. 4, 2007

Dec. 4, 2007

Am Dienstag, 4. Dezember 2007 01:34:55 schrieb Gerald Pfeifer: > This looks like a mistaken consistency check. The code where mapped_stage > is set looks as follows: > > DWORD mapped_stage = stateblock->wineD3DDevice->texUnitMap[texUnit]; > > and at other places we do check for -1, so this looks like another bug > in error handling. In fact the texUnitMap entries are never set to any value < 0 other than -1, so it doesn't make much difference, but for the sake of consistency we should check for -1 here as well

1 0

Re: ntdll: detach modules based on inter-module dependencies (try 2)
by Alexandre Julliard Dec. 4, 2007

Dec. 4, 2007

"Lei Zhang" <thestig(a)google.com> writes: > This patch changes how process_detach() works when called with > bForceDetach set to TRUE. A module would get detached when that module > has no dependencies. After a module gets detached, process_detach() > scans through all the modules and update their dependencies. I'm > assuming there's no circular dependencies between modules. You can't assume that, some modules do have circular dependencies. My suspicion is that this bug should be fixed in winmm, but either way you'll need to show some test cases. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: [AppDB] Change licence to AGPL
by Trent Waddington Dec. 3, 2007

Dec. 3, 2007

interesting. Do you have a list of contributors and their signoff on this license change? Or is it just you two? On Nov 24, 2007 12:36 PM, Alexander Nicolaysen Sørnes <alex(a)thehandofagony.com> wrote: > Change licence to AGPL > > > > Alexander N. Sørnes > > > >

6 7

wine 0.9.5 IO completion port questions
by Marius Slyzius Dec. 3, 2007

Dec. 3, 2007

Hello all. I am trying to run ScimoreDB database which uses fibers, non-blocked IO for sockets and files, and completion port. It did started ok under 0.9.5 and seems it is working!!!:), so I would like to verify few things about async IO with completion ports before recompiling it with winelib. 1) will write/read file/socket with completion port is true asynchronous operation. I am asking because every time write/read operation returns true, and never IO pending error. 2) possible bug: the server sends 1500 bytes with WsaSend ( with 2 WSABUFFER ) and the client receive only 1452 bytes and calls WsaRecv again. Also on the server, the completion port receive event that transferred 1452 bytes. However, the last 8 bytes are never sends nor completion port event arrives. Is it perhaps something to do with the VMWare tool where I run Ubuntu 7.10? Please, send me an info also to my mail: marius(a)scimore.com , because I might miss the answer here. Thank you! Marius Slyzius Scimore UAB (+45) 40862122

1 0

re: kernel32:add ReadFileScatter
by Dan Kegel Dec. 3, 2007

Dec. 3, 2007

Hey Louis, Focht is using a pseudonym, and I believe it's Wine policy to not accept patches without real names. His patches should be used as guides only, and should not be reposted to wine-patches, until such time as he decloaks. If he doesn't want to decloak, I think we can't use his patches. I rather wish he would avoid attach patches to bugs, as it tempts people into submitting them as you have done, which is a waste of effort given Wine's policy. But I greatly value his ability to describe where and how Wine is wrong. - Dan

4 4

handle unimplemented function stubs
by nessi＠nessi-online.de Dec. 3, 2007

Dec. 3, 2007

Hello all, I'm a Windows software developer and I'm trying to execute an application under the wine environment. Now I have a problem introduced with wine 0.9.49. I use the function RegOverridePredefKey from the advapi32.dll. To be compatible with older windows versions, I use the following code: LONG WINAPI (*RegOverridePredefKey)(HKEY hKey, HKEY hNewHKey); HMODULE hMod=GetModuleHandle("Advapi32"); if (hMod) (FARPROC)RegOverridePredefKey=GetProcAddress(hMod,"RegOverridePredefKey"); if (RegOverridePredefKey) RegOverridePredefKey(HKEY_CLASSES_ROOT, hKey); In wine 0.9.49 a stub entry for this function was created, but the implementation is missing. The result is, that the GetProcAddress function is successful, but the RegOverridePredefKey call fails. I can catch the EXCEPTION_WINE_STUB exception (see code below), but the application will nevertheless closed. if (RegOverridePredefKey) { long err = ERROR_CALL_NOT_IMPLEMENTED; try { // test for C-based structured exceptions err=RegOverridePredefKey(HKEY_CLASSES_ROOT, hKey); } #define EXCEPTION_WINE_STUB 0x80000100 __except(GetExceptionCode()==EXCEPTION_WINE_STUB ? EXCEPTION_EXECUTE_HANDLER : EXCEPTION_CONTINUE_SEARCH) { RegOverridePredefKey=NULL; } } What can I do ? I know, that I can find a workaround for this case, but this in not my problem. How can I handle, in principle, such exceptions? What is the reason for creating a function stub without an implementation ??? Thanks Nessi

2 1

Re: rpcrt4: Handle all non-continuable exceptions in CStdStubBuffer_Invoke.
by Alexandre Julliard Dec. 3, 2007

Dec. 3, 2007

Robert Shearman <rob(a)codeweavers.com> writes: > --- > dlls/rpcrt4/cstub.c | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > I have a test for this in my tree, which passes on Windows, but > doesn't pass in Wine yet until we have some more proxy/stub code that > I can use to call an object's methods (ideally IClassFactory). This doesn't work: ../../../tools/runtest -q -P wine -M rpcrt4.dll -T ../../.. -p rpcrt4_test.exe.so cstub.c && touch cstub.ok cstub.c:874: Test failed: ret c0000005 make[2]: *** [cstub.ok] Error 1 -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: Remove four useless checks in dlls/gdi32/enhmetafile.c (RESEND)
by Alexandre Julliard Dec. 3, 2007

Dec. 3, 2007

Gerald Pfeifer <gerald(a)pfeifer.com> writes: > One of the really uncontested ones, I think. ;-) > > ---------- Forwarded message ---------- > From: Gerald Pfeifer <gerald(a)pfeifer.com> > To: <wine-patches(a)winehq.org> > Date: Thu, 1 Nov 2007 16:10:03 +0100 (CET) > Subject: Remove four useless checks in dlls/gdi32/enhmetafile.c > > The members of EMRCREATEDIBPATTERNBRUSHPT are of type DWORD, so > comparing them for >= 0 is a noop which always evaluates to true. > > Gerald > > ChangeLog: > Remove four unnecessary comparisions of DWORD variables for >= 0. Validating the record is not unnecessary, it just needs to be done correctly by checking for wraparounds etc. -- Alexandre Julliard julliard(a)winehq.org

2 6

kernel32:add ReadFileScatter
by luis.busquets＠ilidium.com Dec. 3, 2007

Dec. 3, 2007

Just to add more clarity to the IPR of computer programs and if it is legal to decompile and implement a program again. In Europe, software is protected by Council Directive 91/250/EEC of 14 May 1991 on the legal protection of computer programs (http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:31991L0250:EN:H…) There are a number of things that a software vendor cannot forbid. In particular article 6 speaks about decompilation. It can be done only if the information necessary to achieve interoperability has not previously been readily available (a.1.b) to achieve programs interoperable with the first one (a.1) but not to be used for the development, production or marketing of a computer program substantially similar in its expression (a. 6.2.c). It could be argued that wine is not replicating Windows as it is just the top API. Nevertherless, to avoid any legal problems, it is better to refrain from adding any patches obtained by decompilation.

1 0