[Bug 2356] HD reported as device when mountpoint matches logical link in dosdevices

Wine Bugs wine-bugs at winehq.org
Mon Jul 12 17:02:14 CDT 2004


http://bugs.winehq.org/show_bug.cgi?id=2356





------- Additional Comments From klaus.piemont at gmx.de  2004-12-07 17:02 -------
> Your program doesn't know anything about a Linux file system. Perhaps it did an
> API call to see what the filesystem is and follows the result (VFAT, NTFS,
> Network, CDROM etc. ).

That is the funny thing. I have no idea why this call was made in the first
place. Let me get this a little bit clearer.
What you do in DVDlab is that you select the mpeg files from some explorer like
window into a filelist. This part works. In the filelist you see the filename,
the path (including the driveletter) and some attributes like frame size. This
means that at that point DVDlab must have had opened the file already once to
get that internal information.
In a second operation you select a file from the filelist and insert it into the
project. And here it fails. The area where normaly thumbnails of the movie are
shown remains simply black. I can't think of any possible reason why the drive
itself should be accessed at this point, since every information needed is
already in the filelist shown. A simple open file for read should do the job.
Maybe it is not even intended by the programmer, but he uses a library which
does this internaly. Don't know.

> I was thinking it was the RW permissions of /dev/hdb5 .

Maybe or maybe not. If I where Linux I would never let any program accessing a
mounted partition directly. But I'm rather a newbie in the Linux world and does
not know much about the internals.

> Not possible, it is a windows program that has no knowledge about "logical" wine
> HD's. The program "knows" that it is a Windows volume, and wine has to emulate
that.

That is exactly what I was trying to state. And for that reason the resolving of
the drive letter L: directly to the device /dev/hdb5 makes no sense. Wine should
return a simulated windows volume which is actually mapped through the wine
layer to the directory structure. 

> As far as I understand what you want is a way to make the emulated "L:" behave
> the same way whether dosdevices/l: points to a mount point or not. 

Here we are. That was my intention. I had this kind of control up to wine
20040404 where I could specify "HD" in the config file and everything was clear.
Therefor I came up with the idea to depend the behavior on the filesystem in
fstab. A harddisk filesystem means to ignore that dosdevices/l: points to a
mount point and the path is handled the same way as any other path. Maybe one
could add a override possibility in the config file to reactivate the current
behaviour for a drive lette, even if I could not think about an app which would
require this.

> Did you happen to notice what the CreateFile() returns and what the LastError()
> was in those cases?

I'm sorry, but I have not. I was happy enough to get to the source of my problem
by scanning through the code. I never had really worked with the windows api and
this information was rather meaningless for me. Since programming in Java for
several years now I found even reading C a bit of an adventure :D

-- 
Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.



More information about the wine-bugs mailing list