This mail goes to upstream as well.
[2011-03-14 18:17] Florian Ernst <florian_er...@gmx.net>
> On Tue, Jun 22, 2010 at 01:21:20PM +0200, markus schnalke wrote:
> >
> > Tree prints `argetm' in front of symlink names. See:
> > [...]
> > This bug was reported from lenny/stable, but it exists in sid/unstable
> > too.
>
> Hmm, unfortunately I cannot reproduce this misbehavior at all, neither
> in Etch, Lenny, Squeeze nor in Wheezy/Sid, using different set of
> installed and used locales. As such I now mark this bugreport
> correspondingly.
>
> Do you still experience this error? Anything you could possibly add to
> the bugreport?
Now I dug deeper in the bug and found the reasons.
The problem shows up only if you colorize the output. Then tree takes
the colors defined by $LS_COLORS. This looks like that:
no=00:fi=00:di=01;34:ln=target:pi=40;33:so=01;35:...
Each key has a color value, which may be used directly in a terminal
escape sequence, except `ln' which may also have the value ``target''.
See:
Symbolic link. If you set this to `target' instead of a numerical
value, the color is as for the file pointed to.
http://www.bigsoft.co.uk/blog/index.php/2008/04/11/configuring-ls_colors
This special case leads to the bug, reported here.
With the above $LS_COLORS, the escape sequence for a directory would
be `^[[01;34m'. For a link we get `^[[targetm', resulting in the escape
sequence `^[[t' and the text `argetm'.
In the sources start reading at tree.c:1391 parse_dir_colors(), the
``Hacked in DIR_COLORS support for linux.'' There $LS_COLORS is split
and link_flgs gets set (It could be to `target').
1433 case COL_LINK:
1434 if (c[1]) link_flgs = scopy(c[1]);
1435 break;
If link_flgs is set to `target' stat(2), instead of lstat(2), should
be used for symlinks. Or maybe we should do a further stat(2) in this
case. Maybe this function can be handled in a similar way as orphan
and missing links. Or tree would simply ignore `ln=target' and use
some default link color instead. This should be left open to upstream.
meillo
P.S.
I took tree-1.5.2 for reference.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
