On 10/14/07, Michael Albinus <[EMAIL PROTECTED]> wrote: [snip] > I cannot reproduce the problem. Locally, I can use gdb for remote > buffers without problems, and I see the source buffer. > > Do you have some special settings one should care? I simple scenario I > could follow?
I have set up a few default transfer methods (external) for a few boxes as an alist. But I don't think that should affect this. > > > I looked into this and it seems to me that the problem is that the > > function 'tramp-handle-file-exists-p' returns nil. > > This calls the macro 'with-file-property' which should set the > > property file-exists-p to t after fetching testing this on the remote > > machine with 'test -e'. But it seems to just go ahead and set this > > value without checking remotely (using tramp-set-file-property). > > I am using edebug to step through the macro - however I don't quite > > understand macros yet. > > Hmm. That's kind of caching Tramp uses internally. Hard to debug, I confess. > The idea is that tramp-get-file-property is called first, which shall > return the cached value. Only in case it returns 'undef, the body of > with-file-property is evaluated, and the result is put into the cache via > tramp-set-file-property. > [snip] I don't have access to my setup right now (its at work) - but as far as I can remember - the body (that tries to fetch the correct property remotely) doesn't get evaluated. I shall take a look tomorrow and send you a trace. On a side note - I would like to get involved in tramp development - are there any small tasks / bugfixes that you can think of - that would help me get a good grip on the source code ? Regards AM _______________________________________________ Tramp-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/tramp-devel
