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

wnaspi and DVD Burner

by Stefan Leichter

Hello, i tried to make some progress with Ahead Nero, but now i'm stuck. I found the NeroSDK-1.08.zip and inside the comand line tool NeroAPITest.exe. The source code of NeroAPITest.exe is in the archive too. This makes finding the problems easier. Executing the comand 'wine NeroAPITest.exe --listdrives' the open function in SCSI_OpenDevice fails first with ENOMEDIUM. Putting a disk into the drive the command fails with errno 30 (Read-only). Opening the drive read only works, but next is a write on the file descriptor that fails because it was opened readonly ... The scenario above fails only for my PATA DVD burner. My SCSI DVD reader and my SCSI CD reader are working well with the command sequence listed above. My OS is Debian etch (32bit) the burner is a Plextor PX-716A. Is the reason for this the difference between IDE and SCSI in the kernel? Is it a fault of the DVD burner device (firmware)? What might be a workaround for the problem? Thanks for your thought Stefan

Oct. 17, 2007

Re: Wine Stats Malfunctioning

by Alexandre Julliard

luis.busquets(a)ilidium.com writes: > Web page: > > http://www.winehq.org/site/winapi_stats > > on wine statistics does not work. Yes, winapi_extract failed to parse a file in the last release. I've uploaded a fixed version. -- Alexandre Julliard julliard(a)winehq.org

Oct. 17, 2007

Re: [PATCH] wininet: Substitute strchrW with memchrW in InternetCrackUrlW.txt

by Nigel Liang

On 10/17/07, Frank Richter <frank.richter(a)gmail.com> wrote: > On 17.10.2007 06:49, Nigel Liang wrote: > > Hi, > > > > strchrW assumes a NULL-terminated string. May crash if terminating character is > > not found. memchrW is better because you can specify the maximum number of > > bytes to search. > > On the flipside, memchrW() doesn't stop at a NUL... Ops! Please ignore this patch, the bounds of the memchr are way off... Curiously, this did not crash on any of the tests I run even though it is always going past the string boundary. On the other hand, using strchrW would reliably crash after roughly 10 minutes of activity...

Oct. 17, 2007

Re: Possible WIne regression with Warcraft

by Michael Lothian

> Please fix your hosts file while your at it. It might not be this > specific issue, but it's going to cause other problems for certain. What would you recommend?

Oct. 17, 2007

Re: [PATCH] wininet: Substitute strchrW with memchrW in InternetCrackUrlW.txt

by Frank Richter

On 17.10.2007 06:49, Nigel Liang wrote: > Hi, > > strchrW assumes a NULL-terminated string. May crash if terminating character is > not found. memchrW is better because you can specify the maximum number of > bytes to search. On the flipside, memchrW() doesn't stop at a NUL... -f.r.

Oct. 17, 2007

Wine Stats Malfunctioning

by luis.busquets＠ilidium.com

Web page: http://www.winehq.org/site/winapi_stats on wine statistics does not work.

Oct. 17, 2007

Re: widl [1/2]: Improve file cleanup when errors occur

by Alexandre Julliard

Dan Hipschman <dsh(a)linux.ucla.edu> writes: > @@ -446,10 +441,18 @@ static void rm_tempfile(void) > abort_import(); > if(temp_name) > unlink(temp_name); > - if (header_name) > - unlink(header_name); > - if (client_name) > - unlink(client_name); > - if (server_name) > - unlink(server_name); > + if (nerrors) { > + if (do_header) > + unlink(header_name); > + if (do_client) > + unlink(client_name); > + if (do_server) > + unlink(server_name); > + if (do_idfile) > + unlink(idfile_name); > + if (do_proxies) > + unlink(proxy_name); > + if (do_typelib) > + unlink(typelib_name); > + } This will break cleanup on signals. I think it's better to always remove the files instead of relying on the error count being set correctly. -- Alexandre Julliard julliard(a)winehq.org

Oct. 17, 2007

Re: "make test" failure in gdi32/bitmap.c

by Dmitry Timoshkov

"Francois Gouget" <fgouget(a)free.fr> wrote: >> err:x11drv:X11DRV_CreateBitmap Trying to make bitmap with planes=1, bpp=24 >> err:bitmap:DIB_GetBitmapInfo (39): unknown/wrong size for header >> err:bitmap:DIB_GetBitmapInfo (11): unknown/wrong size for header >> err:bitmap:DIB_GetBitmapInfo (13): unknown/wrong size for header >> err:bitmap:DIB_GetBitmapInfo (16): unknown/wrong size for header > > I don't know anything about these however. The test intentionally passes incorrect data to various APIs which leads to the above ERRs. Perhaps it's time to turn those ERRs to WARNs since the parts of the code they belong to are mature enough these days. -- Dmitry.

Oct. 17, 2007

Re: "make test" failure in gdi32/bitmap.c

by Francois Gouget

On Tue, 16 Oct 2007, Dan Kegel wrote: > "make test" fails for me with: > > bitmap.c:1113: Test failed: GetDIBits with 4 bpp DIB selected in DC: > Invalid DIB bits > bitmap.c:1113: Test failed: GetDIBits with 8 bpp DIB selected in DC: > Invalid DIB bits I'm working on the above two. From my testing they only happen when X is in 16bpp mode (and probably 15bpp too). If I switch to 24bpp then all is fine. The reason is that dlls/winex11.drv/dib.c uses two types of color maps: * the first one contains RGBQUADs so in it white is (0xff, 0xff, 0xff, 0) * the second one maps an index to the corresponding X color (an int). So in 15bpp white would be 0x7fff. The bug happens because the code casts the int* color map to the RGBQUAD* color map in many places so that it interprets 0x7fff as an RGBQUAD and thus returns the wrong color index. > err:x11drv:X11DRV_CreateBitmap Trying to make bitmap with planes=1, bpp=24 > err:bitmap:DIB_GetBitmapInfo (39): unknown/wrong size for header > err:bitmap:DIB_GetBitmapInfo (11): unknown/wrong size for header > err:bitmap:DIB_GetBitmapInfo (13): unknown/wrong size for header > err:bitmap:DIB_GetBitmapInfo (16): unknown/wrong size for header I don't know anything about these however. -- Francois Gouget <fgouget(a)free.fr> http://fgouget.free.fr/ tcA thgirypoC muinelliM latigiD eht detaloiv tsuj evah uoY

Oct. 17, 2007

Re: [PATCH] wininet: Conformance test for InternetCrackUrlA.txt

by Dmitry Timoshkov

"Nigel Liang" <ncliang(a)gmail.com> wrote: > + ok( GetLastError() == ERROR_INVALID_PARAMETER, "Last error wrong: %p\n", > + GetLastError()); GetLastError() returns not a pointer. -- Dmitry.

Oct. 17, 2007