Re: ad-hoc multi-hop editing of remote files

[Chris Withers]

On 20/10/2011 12:57, Michael Albinus wrote:
> Look at the syntax. It's horrible to parse, expecially when you want to
> complete (like in ido-mode). Do you know exactly whether a substring is
> meant as method, user name or host name? I don't.
>
> One proposal was to require always a method name in case of multihops. I
> don't know whether this will make things better.

Yeah, I think being explicit would be exactly what's needed.

> > Yes please, but proper ad-hoc, allowing specification of the transit
> > user, remote host and remote user all in one path. If there's anything
> > I can do to make that happen more quickly, please let me know...
>
> I have no real idea how to achieve this, sorry. If you come with a
> working solution you're welcome!

/user|address@hidden:/

> Why not to go the other direction, and make ido aware of proxy hosts?

Not sure what you mean by that, can you give an example?

> > > > http://cvs.savannah.gnu.org/viewvc/tramp/ChangeLog?root=tramp&view=markup
> > > > the correct change log to be looking at? I hope not, since it doesn't
> > > > mention either proxies arriving or multi: going away...
> > >
> > > For the Lisp code, you shall use
> > > http://cvs.savannah.gnu.org/viewvc/tramp/lisp/ChangeLog?root=tramp&view=markup
> >
> > Okay, what's the other ChangeLog?
>
> There are 3 ChangeLogs. "viewvc/tramp/ChangeLog" is for overall
> makefiles and configuration scripts. "viewvc/tramp/lisp/ChangeLog" is
> for the real Lisp code. And  "viewvc/tramp/texi/ChangeLog" is for the
> documentation.

Thanks for the explanation :-)

cheers,

Chris

PS: How come still CVS as opposed to SVN or, better yet, Git?

--
Simplistix - Content Management, Batch Processing & Python Consulting
            - http://www.simplistix.co.uk