Hello, I found a bug in Latest wine releases. I found it in the wine-20040914,wine-20040813 rpms for the RedHat Distributions.I had tested it in RedHat-9.0,7.3 systems. The bug is that, when we try to open(file->open) from any windows applications, the application hangs. For eg., even in Notepad, when I click File-> Open, wine shows the following error. --------- wine: Unhandled exception (thread 000b), starting debugger... WineDbg starting on pid 0xa Unhandled exception: divide by zero in 32-bit code (0x403a5ff5). In 32 bit mode. 0x403a5ff5: idivl 0xffffffc4(%ebp),%eax Wine-dbg> -------------- But the problem does not exist for older wine releases. Eg: wine-20030508. I guess, the reason for the bug is with the config file. Regards, Saravanan ______________________________________ Scanned and protected by Email scanner
On 28 Oct 2004 11:04:56 +0530, you wrote:> Hello, > > I found a bug in Latest wine releases. I found it in the > wine-20040914,wine-20040813 rpms for the RedHat Distributions.I had > tested it in RedHat-9.0,7.3 systems. > The bug is that, when we try to open(file->open) from any windows > applications, the application hangs. For eg., even in Notepad, when I > click File-> Open, wine shows the following error. > --------- > wine: Unhandled exception (thread 000b), starting debugger... > WineDbg starting on pid 0xa > Unhandled exception: divide by zero in 32-bit code (0x403a5ff5). > In 32 bit mode. > 0x403a5ff5: idivl 0xffffffc4(%ebp),%eax > Wine-dbg> > -------------- > But the problem does not exist for older wine releases. Eg: > wine-20030508. > I guess, the reason for the bug is with the config file.No, it is a known bug that exists already far too long. With your help I like to create a fix for it. The problem lies in determine what exactly the file system is on one of your simulated drives, this is supposedly a FAT file system (FAT12, FAT16, FAT32). First determine what drive this is. If you are not sure, run wine like: WINEDEBUG=+file wine ... and make it crash. There should be a line with CreateFileW just before the crash, the name of the device in the form "\\\\C:\\". Then assuming it is drive C:, and your ~/.wine/dosdrives/c: points to the mount point of a mounted file system, lets say /dev/hda1. Then make a copy of the superblock this way: dd if=/dev/hdax of=superb.bin count=4 (/dev/hdax is the device you found in the previous step) and send me the file superb.bin Rein. -- Rein Klazes rklazes@xs4all.nl
Hello,> > I guess, the reason for the bug is with the config file. > > No, it is a known bug that exists already far too long. With your help >I like to create a fix for it.But, how it happens that In a same system old wines works but the new ones does not. In the case of Latest Wine, when I remove all sym-links from the dosdevices directory, except the "c", then it works.> The problem lies in determine what exactly the file system is on oneof> your simulated drives, this is supposedly a FAT file system (FAT12, > FAT16, FAT32). > > First determine what drive this is. If you are not sure, run winelike:> > WINEDEBUG=+file wine ... > > and make it crash. There should be a line with CreateFileW just before > the crash, the name of the device in the form "\\\\C:\\". > > Then assuming it is drive C:, and your ~/.wine/dosdrives/c: points to > the mount point of a mounted file system, lets say /dev/hda1.I got the following before the crash. (My C drive is pointing to ../drive_c) ---- trace:file:CreateFileW returning 0x44 trace:file:CreateFileW L"c:\\windows\\win.ini" GENERIC_READ FILE_SHARE_READ OPEN_EXISTING attributes 0x80 trace:file:RtlDosPathNameToNtPathName_U (L"c:\\windows\\win.ini",0x4074cdcc,(nil),(nil)) trace:file:RtlGetFullPathName_U (L"c:\\windows\\win.ini" 520 0x4074cb48 (nil)) trace:file:wine_nt_to_unix_file_name L"\\??\\C:\\windows\\win.ini" -> "/root/.wine/dosdevices/c:/windows/win.ini" trace:file:CreateFileW returning 0x44 ----> Then make a copy of the superblock this way: > > dd if=/dev/hdax of=superb.bin count=4 > > (/dev/hdax is the device you found in the previous step)Here what should i give as argument to `if=`. Regards, Saravanan ______________________________________ Scanned and protected by Email scanner
On 28 Oct 2004 16:17:50 +0530, you wrote:> Hello, > > > I guess, the reason for the bug is with the config file. > > > > No, it is a known bug that exists already far too long. With your help > >I like to create a fix for it. > > But, how it happens that In a same system old wines works but the new > ones does not.That is called a regression, happens all the time. Could be caused by a new function that was not there before, but crashes on your disk.> > In the case of Latest Wine, when I remove all sym-links from the > dosdevices directory, except the "c", then it works.Then c: is not the problem.> > The problem lies in determine what exactly the file system is on one > of > > your simulated drives, this is supposedly a FAT file system (FAT12, > > FAT16, FAT32). > > > > First determine what drive this is. If you are not sure, run wine > like: > > > > WINEDEBUG=+file wine ... > > > > and make it crash. There should be a line with CreateFileW just before > > the crash, the name of the device in the form "\\\\C:\\". > > > > Then assuming it is drive C:, and your ~/.wine/dosdrives/c: points to > > the mount point of a mounted file system, lets say /dev/hda1. > > I got the following before the crash. (My C drive is pointing to > ../drive_c) > ---- > trace:file:CreateFileW returning 0x44 > trace:file:CreateFileW L"c:\\windows\\win.ini" GENERIC_READ > FILE_SHARE_READ OPEN_EXISTING attributes 0x80 > trace:file:RtlDosPathNameToNtPathName_U > (L"c:\\windows\\win.ini",0x4074cdcc,(nil),(nil)) > trace:file:RtlGetFullPathName_U (L"c:\\windows\\win.ini" 520 0x4074cb48 > (nil)) > trace:file:wine_nt_to_unix_file_name L"\\??\\C:\\windows\\win.ini" -> > "/root/.wine/dosdevices/c:/windows/win.ini" > trace:file:CreateFileW returning 0x44 > ----You need the last CreateFile before the crash ("starting debugger") and is must look like CreateFileW(L"\\\\x:\\> > Here what should i give as argument to `if=`.You will need to find that CreateFileW first. Rein. -- Rein Klazes rklazes@xs4all.nl
On 29 Oct 2004 10:29:15 +0530, you wrote:> Earlier the application crashed, when i had a symlink to / in the > Dosdevices directory.I have made a patch that should prevent wine crashing on this. If you compile from source you can try it out: http://cvs.winehq.org/patch.py?id=14356 if not, wait to the next wine release. Rein. -- Rein Klazes rklazes@xs4all.nl
On Fri, 29 Oct 2004 12:40:22 -0700, you wrote:> Thanks for your quick response. I had wine-20040813-1.src.rpm with me. > But i have a doubt, whether i have to compile it for each Redhat > distributions; Can't I use a single WINE RPM for different RedHat versions > ?.I do not know anything about Red Hat or RPM's, but I think you should be careful. If the binary packages are for different versions, that is probably for a reason. CC 'd to the wine users list. Please keep the discussion there. Rein. -- Rein Klazes rklazes@xs4all.nl