Re: Bug in mc 4.6.0-pre1 ?



Hello, Mathieu!

> 1) When I type 
> mc --help
>     or
> mc --version
> 
> It shows infos and ends with 
> ".: Nom de fichier trop long"

As I wrote in another e-mail, that's a serious issue with the wrapper 
functions.  It's another example of bad design without any security 
considerations.

If the patch to add the "-P" option in its present form were made today
and submitted to this list, it would be certainly rejected.  But todays
users expect this feature (directory change on exit from mc) to work, so
it needs to be fixed somehow.  I hope to fix it before 4.6.0 release.

Most likely mc will leave a file at some secure and local directory (we
don't want this file to be transmitted over NFS back and forth and clash
with similar files on other machines), and the script will read and
interpret it.

I think a good location would be ${TMPDIR=/tmp}/mc-$UID/mc-cd-$$ This way,
we respect $TMPDIR which is usually local or at least not shared with
other machines, we can have strict permissions on the mc-$UID directory,
and both the shell and mc know the PID of mc.  Some $RANDOM-ness can be
added if necessary.

Another solution would be to redirect some non-standard file descriptor,
e.g. run "mc 3>${TMPDIR=/tmp}/mc-$UID/mc-cd-$$", so that the name choice
would be completely in the script.  But we'll have to be careful to close
that descriptor in child processes after every fork() to prevent possible
exploits.

> 2) I use aterm with transparent background. No problem with that when
>    I start mc: the background is still transparent.
> 
> If get in trouble if I type C-o: the bgcolor become black. It's the same
> if I exit mc: the bgcolor become black too.
> 
> This is not really painy since a simple `clear` reset everything too
> normal but if I remember well the previous release of mc was ok on
> this point.

I downloaded aterm-0.4.2 and could not reproduce this problem.  You should 
give more details, in particular the version of aterm and the output of 
"mc -V" for both old and new versions of mc.

-- 
Regards,
Pavel Roskin




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]