October 2010 - Wine-Devel - list.winehq.org

Google Summer of Code Mentors?
by Jeremy White Oct. 14, 2010

Oct. 14, 2010

Did we have an 'official' leader for GSOC this past year? If not, could I ask someone (*cough* Kai *cough*) to nominate themselves? We need to coordinate getting some information to the SFC so that we can collect the mentor stipend from Google. Cheers, Jeremy

3 2

Re: ws2_32: Add support and tests for WSARecvMsg and IP_PKTINFO.
by Mike Kaplinskiy Oct. 14, 2010

Oct. 14, 2010

On Tue, Oct 12, 2010 at 8:01 PM, Erich Hoover <ehoover(a)mines.edu> wrote: > Real Name: > Erich Hoover > Description: > While searching for something else I discovered a bug relevant to the > interface-bound UDP broadcast patches I've been working on. Apparently, > IP_PKTINFO does work on Windows when used with the special WSARecvMsg > function (Bug #19493). The attached patch addresses this issue by > implementing WSARecvMsg through WS2_recvfrom and by converting the Unix > IP_PKTINFO response to the Windows equivalent. > ChangeLog: > ws2_32: Add support and tests for WSARecvMsg and IP_PKTINFO. > > > > Sorry I must have missed the RFC when reading my mail. These comments are purely mine, and may not have any value for getting your patch in :). +static inline WSACMSGHDR *create_control_message(int level, int type, WSACMSGHDR *current, ULONG *maxsize, void *data, int len) This doesn't seem like a "create", more like a "fill" or "write". + int newsize = (int)*maxsize; + + /* Make sure there is at least enough room for this entry */ + newsize -= sizeof(WSACMSGHDR) + CMSG_ALIGN(len); + if (newsize < 0) + return 0; + *maxsize = (ULONG)newsize; Just declare it as a ULONG/size_t so you don't have to cast. +static inline int convert_control_headers(struct msghdr *hdr, ULONG *maxsize) You shouldn't mix using memory as unix-style CMSGs and windows-style CMSGs. Just allocate something to store the unix messages, read them in and convert them into the user-passed buffer. This seems similar to what we do for address conversion. + memset(cmsg_win, 0x00, sizeof(WSACMSGHDR)); /* Don't use garbage data if no headers are found */ I think in general that is discouraged at wine (don't quote me on that). You should just initialize the values manually. + if (wsa->control && !convert_control_headers(&hdr, wsa->controllen)) + { + /* Insufficient room for control headers */ + errno = EINVAL; + return -1; + } This should probably be in an #ifdef of some sort. + msg->dwFlags |= WINE_MSG_HASCTRL; + memcpy(buftmp, msg->lpBuffers, msg->dwBufferCount * sizeof(WSABUF)); + buftmp[msg->dwBufferCount] = msg->Control; + ret = WS2_recvfrom( s, buftmp, msg->dwBufferCount, lpNumberOfBytesRecvd, + &msg->dwFlags, msg->name, &msg->namelen, + lpOverlapped, lpCompletionRoutine ); You shouldn't add internal flags like that. Just rewrite WS2_recvfrom to allow returning message headers instead of hacking around it. +static int (WINAPI *pWSARecvMsg)(SOCKET,LPWSAMSG,LPDWORD,LPWSAOVERLAPPED,LPWSAOVERLAPPED_COMPLETION_ROUTINE) = 0; +GUID WSARecvMsg_GUID = WSAID_WSARECVMSG; For one reason or another the convention is to get it during your test function (see AcceptEx & ConnectEx). (This does have some merit as SIO_GET_EXTENSION_FUNCTION_POINTER can vary with the socket that's passed) + s1addr.sin_family = AF_INET; + s1addr.sin_port = htons(9375); + s2addr.sin_family = AF_INET; + s2addr.sin_port = htons(9375); You should probably bind to 0 and then use getsockname to get the actual port you bound to. + s1=socket(AF_INET, SOCK_DGRAM, 0); + ok(s1!=INVALID_SOCKET, "socket() failed error: %d\n", WSAGetLastError()); Spaces around = & != would make me happy (and below). Also you don't error out correctly if socket creation fails. diff --git a/include/mswsock.h b/include/mswsock.h index 322ab20..5749e59 100644 --- a/include/mswsock.h +++ b/include/mswsock.h This should be a separate patch. diff --git a/include/ws2ipdef.h b/include/ws2ipdef.h index 11b3689..ec0b85a 100644 --- a/include/ws2ipdef.h +++ b/include/ws2ipdef.h Same here. Mike.

2 2

Ongoing Debian package maintenance.
by Ben Klein Oct. 14, 2010

Oct. 14, 2010

Hello everyone, My health has not improved at all since my last call for help with the Debian package management. I really need someone to take over from me, or at least some help to devise a new set of (semi-)automated build scripts. Any volunteers? Kind regards, Ben

5 11

Removal of very very old versions from bugzilla?
by Scott Ritchie Oct. 13, 2010

Oct. 13, 2010

Wise idea or no? If not is there at least a way to hide them from the report bug page? I'm talking about the 2001-era version numbers, before 0.9 even. -Scott Ritchie

2 1

Re: [PATCH] jscript/tests: add tests for ActiveScriptSite_OnScriptError and IActiveScriptError. [try 3]
by Reece Dunn Oct. 13, 2010

Oct. 13, 2010

On 9 October 2010 15:51, Reece Dunn <msclrhd(a)googlemail.com> wrote: > try 1 -- initial series (1-3) > try 2 -- revised test infrastructure based on Jacek's comments > try 3 -- fleshed out the tests to complete the test coverage for > IActiveScriptError Any reason this has not been committed yet? Thanks, - Reece

1 0

Re: [5/5] msi/tests: Don't create a nested restore point on recent versions of MSI.
by testbot＠testbot.winehq.org Oct. 13, 2010

Oct. 13, 2010

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=6155 Your paranoid android. === WXPPROSP3 (32 bit install) === Timeout === W2K3R2SESP2 (32 bit install) === Timeout

1 0

Re: [4/5] mscoree: Use the new search methods for GetCORVersion/GetCORSystemDirectory. (try 2)
by Alexandre Julliard Oct. 13, 2010

Oct. 13, 2010

"Vincent Povirk" <vincent(a)codeweavers.com> writes: > This shouldn't introduce a failure if Mono is missing. I'm afraid it does: ../../../tools/runtest -q -P wine -M fusion.dll -T ../../.. -p fusion_test.exe.so fusion.c && touch fusion.ok wine: Install the Windows version of Mono to run .NET executables fusion.c:193: Test failed: Expected S_OK, got 80131700 fusion.c:194: Test failed: Expected "C:\windows\assembly\NativeImages_el32.dll_32", got "nochange" make: *** [fusion.ok] Error 2 -- Alexandre Julliard julliard(a)winehq.org

1 0

RFC: Fixing imagelist control tests for 16bpp
by Austin Lund Oct. 13, 2010

Oct. 13, 2010

It seems that the imagelist control on XP has some strange device dependence even though the tests use device independent bitmaps. Attached is a patch to try and correct the test for what is broken. Is this the right approach?

1 0

Re: [transl] Deal with WS_EX_LAYOUTRTL for RTL languages
by Yaron Shahrabani Oct. 12, 2010

Oct. 12, 2010

On Tue, Oct 12, 2010 at 11:08 AM, Paul Vriens <paul.vriens.wine(a)gmail.com>wrote: > Hi, > > This patch makes sure that a missing WS_EX_LAYOUTRTL for EXSTYLE is flagged > for RTL languages and that it's ignored if present. > > It's not ideal yet as the potential errors shown on the webpage do not > fully explain the issue/error. More changes are needed to get that into > place but this patch at least makes sure that the real errors are catched > and don't get lost in the multitude of errors we have now (for Hebrew for > example). > > Changelog > Deal with WS_EX_LAYOUTRTL for RTL languages > Hey Paul, I think these should apply to several other languages: Divehi Syriac Urdu (There are several other languages but I took the following out of the following list: http://source.winehq.org/transl/) Another language which I'm not sure about is Pujabi since it has 3 popular scripts: Gurmukhi and Devanagari are RTL written while Shahmukhi is RTL, Wikipedia in written in the RTL script while Ubuntu is translated using the LTR scripts (can't determine which though). Its all about location, Pakistanies are writing in RTL whille all others write with the LTR scripts, It all depends on who you ask :) Your patch to the translation check is working great, thanks alot! Yaron Shahrabani <Hebrew translator> > > -- > Cheers, > > Paul. > > > >

1 0

Re: [4/5] mscoree: Use the new search methods for GetCORVersion/GetCORSystemDirectory.
by Vincent Povirk Oct. 12, 2010

Oct. 12, 2010

Hang on, this one will also introduce a failure. On Tue, Oct 12, 2010 at 12:51 PM, Vincent Povirk <vincent(a)codeweavers.com> wrote: > > > > > >

1 0