Hi Michael,

Le jeudi 20 août 2015, 20:01:55 Michael Albinus a écrit :
> Pierre Téchoueyres <pierre.techouey...@free.fr> writes:
> >     Hello everybody,
> 
> Hi Pierre,
> 
> > I've made some tests with stat and sed and it seems that stat support the
> > null character only with --printf switch. sed seems to catch null with
> > \o0.
> > 
> > For the fear of shell's handling of null char : the first problem is the
> > stat program. Linux and BSD have differents options (and I don't know how
> > it work on Solaris nor in HP*UX nor ...)
> > So maybe the solution is to first check for the version of stat and use an
> > alist of options for each ones (maybe a custom option) but that's a bigger
> > work.
> 
> Well, this is too much hassle I fear. We DON'T know which shells, and
> which broken implementations of them, are used remotely. So I rather
> oppose such a patch.
> 
OK I understand.
But actually stat isn't normalized between unices and we rely on -c switch 
which doesn't exist on BSD. That why I thinked of an custom option and a 
dispatch depending on the version check. Anyway.

> What do you think about my other proposal, using `tramp-end-of-heredoc'
> as marker?
> 
This should work, but I guess the debugging output will be ... interresting 
...

I'll try to work on it.

Best regards Pierre


_______________________________________________
Tramp-devel mailing list
Tramp-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/tramp-devel

Reply via email to