Re: minor bug in network file edit.
- Bram Moolenaar wrote:
>not to people who use rcp, scp or rsync :)
> Michael Beattie wrote:
> > Wichert (debian maintainer of vim) suggested I mail you about this minor
> > bug with network file editing:
> > omnic@relativity:~> vim scp://aeon/var/www/index.html
> > from vim:
> > :!scp aeon:var/www/index.html /tmp/vAMUDV4L
> > scp: var/www/index.html: No such file or directory
> > 1 returned
> > Error detected while processing function NetRead->NetGetFile:
> > line 5:
> > Can't open file /tmp/vAMUDV4L
> > so:
> > omnic@relativity:~> vim scp://aeon//var/www/index.html
> > from vim:
> > :!scp aeon:/var/www/index.html /tmp/vAW7yok9
> > <html>....
> > works fine.
> > not sure how you'd go about this one... mandating absolute paths?
> Well, at least you can get the file you want this way.
> When using scp://, the path you give is relative to your login directory.
> It's probably the same for rcp://. I think this is OK, because it makes it
> easy to access your own files.
> Accessing files with an absolute directory is used for ftp:// and http://.
> Perhaps it's confusing that scp:// and rcp:// are different?
> Would it be sufficient to explain this in the docs, or do we really need toi think an explanation would be sufficient. nobody's complained about
> change how it works?
the rcp behaviour and it's the same. anyway, i thought everything after
the : was protocol dependent anyway (at least everything after the 3rd / is).
just because ftp and http uris look the same, doesn't mean all uris have to.
i'm sure we'll get used to the scp://host/~/path but it's a mistake to think
that rfc2396 requires it (i.e. just because there is such a thing as an
absolute uri, that does not mean that what appears after the 3rd / in such
an absolute uri must be an absolute unix path to a file on a disk somewhere -
that is a misinterpretation of the rfc). i think this usage will be less
intuitive to scp users than the current behaviour. ymmv.