Re: URL syntax (was: Re: Tramp, "[]" in file names, and file-expand-wildcards)

[Stefan Monnier]

> "Stefan Monnier" <monnier+gnu/address@hidden> writes:
> > But I don't think that the current way substitute-in-file-name
> > checks for file-name-handlers is enough since the url file-name-handler
> > will simply not be called on file names like "/foo/http://blabla";.
> > 
> > The substitute-in-file-name command has special support to recognize
> > "/foo/a:/bar" under W32 and MS-DOS, but we need to make it more generic.
> 
> Is there any reason not to have substitute-in-file-name _always_ look
> for appropriate file-name handers?  I don't think it's called all that
> often, so speed probably isn't a real concern here.

I'm not sure I understand.  `substitute-in-file-name' does always
look for file-name-handlers.  But URL's file-name handler doesn't
match "/foo/http://blabla";.
Of course URL could add a file-name-handler that matches
"/[a-z]+:/" (i.e. without anchoring with \`) specifically for
`substitute-in-file-name'.


        Stefan