Mail Archives: cygwin/2001/10/02/11:03:31
On Tuesday 2 Oct 01, Matt Landau writes:
> At 05:37 AM 10/2/2001, David Starks-Browning wrote:
> >If your non-Cygwin application (like Emacs) requires HOME be set in a
> >non-Cygwin way, then don't start emacs from your Cygwin bash shell.
>
> Let me clarify the problem.
>
> Emacs doesn't care what $HOME is set to, it cares what path it sees on the
> command line. The issue is that bash, by default, expands ~/filename to a
> posix-style path, resulting in commands like "emacs /c/users/matt/.bashrc",
> which doesn't work for any non-Cygwin program.
OK. But in fact, doesn't Emacs care about $HOME so it can find .emacs
or _emacs?
> Saying "don't run non-Cygwin programs from bash" isn't a thrilling answer;
> an interactive shell is most useful when you can use it as your shell for
> *everything*.
>
> There is a half-hearted workaround, which is to reset HOME to a win32-style
> path in .bashrc so that ~/.bashrc expands to c:/users/matt/.bashrc (since
> cygwin programs understand win32-style paths, there's little reason not to
> do this), but it breaks bash's completion. That's the real bug I was
> asking about. It's not clear to me why bash should be able to expand
> posix-style paths but not win32-style paths.
Probably because bash is a unix program, not a Windows program?
Without looking at the bash source code (bad thing to say on this
list, I know), your complaint sounds like a feature request, not a bug
report.
It's not *always* possible to make $HOME a win32 path. I would expect
ssh and rsh, for example, to interpret the colon as host:path.
> I suppose another workaround is to define a shell function that turns
> "emacs filename" into "emacs `cygpath -w filename`", converting all names
> back to win32-style paths before emacs ever sees them, but having to do
> this for every non-cygwin application I ever run from the shell is going to
> get tedious.
More tedious than not having Cygwin & bash at all? :-)
(Sorry, couldn't resist...)
Seriously though, point taken, but I don't know that it's enough to
motivate Cygwin developers (and/or Chet Ramey and/or ...) to take it
on board. (I'm not one, and I shouldn't speak for them.)
For Emacs, have you looked into cygwin-mount.el? Here is one
reference: <http://www.cygwin.com/ml/cygwin/2001-05/msg00684.html>.
Regards,
David
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -