Re: problems with tramp 2.0.12 and 2.0.24 on xemacs

[Kai Großjohann]

Bruce Lowekamp <address@hidden> writes:

> I applied the patch to the 2.0.12 version that comes with the xemacs
> package, since there aren't any other issues with that package (that
> I've run into).  It doesn't seem to solve the problem, though.  (I
> thought that the "truename is "//" seemed a bit strange, so I did try
> to set result to "/" if it was "//", but that didn't help either, so I
> took it out.)
>
> To get a simpler error, I put a java file in /tmp, set
> tramp-debug-buffer to t and visited the java file in /tmp.  Eventually
> it exits on (in the message log):
>
> tramp: file attributes with perl: /[ssh/grid.cs.wm.edu]/..
> tramp: Processing step `..'
> tramp: True name of `/[ssh/grid.cs.wm.edu]/../' is `//'
> tramp: Finding true name for `/[ssh/grid.cs.wm.edu]/'
> tramp: True name of `/[ssh/grid.cs.wm.edu]/' is `//'
> tramp: Finding true name for `/[ssh/grid.cs.wm.edu]/../'
> tramp: Check /..
> File mode specification error: (error "Lisp nesting exceeds
> `max-lisp-eval-depth'")
>
> Following is a bit of what appears in the debug log starting from
> right after the succesful file transfer.  Any further help on this
> would be appreciated.  I'm happy to provide further debug info if you
> can tell me what I need to run or change to generate it.  I don't know
> how to force xemacs to give me a backtrace to see what the actual call
> chain was that caused this when it exits on eval-depth.  I think I set
> the "backtrace on error" option, but it didn't seem to have an effect.

I'm now guessing that the problem comes from jde-root-dir-p which has
a special case for Ange-FTP but not for Tramp.

My guess is that it would work to add another kludge in that function
for Tramp, but I think that the function could be improved so that it
does not need (so many) kludges.

Let me talk to the JDE folks.

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)