Re: [PATCH] possible fix for #714, finally ;-)

Hello.

On Sat 2002-09-21 at 17:01:11 -0400, Garrett Rooney wrote:
> On Saturday, September 21, 2002, at 04:50 PM, Benjamin Pflugmann wrote:
>
> >Sorry... forgot to include an example of the new output:
> >
> >nora:~> svn revert -R /usr/local/src/phi/test/one/bla
> >subversion/libsvn_wc/adm_ops.c:972: (apr_err=200007, src_err=0)
> >svn: Trying to use an unsupported feature
> >svn: revert: error using subcommand for
> >`/usr/local/src/phi/test/one/bla'
> >subversion/libsvn_wc/adm_ops.c:1009: (apr_err=200007, src_err=0)
> >svn: try instead: svn update '/usr/local/src/phi/test/one/bla'
> >Exit 1
>
> why does it say 'error using subcommand'... why not 'error reverting'

Because it is "revert: error using subcommand" and I considered
"revert: error reverting" kind of redundant. But as I said, English is
not my strength, so I may be way off here.

> or something like that? or 'cannot revert nonexistant directory
> `blah'' or something else that would give more of a clue what's going
> on.

The problem is, that revert_error() has a quite tight template:

  svn: revert: SOME_MESSAGE for `/usr/local/src/phi/test/one/bla'

So "cannot revert nonexistant directory" does not fit, because the
extra "for". Is it acceptable to not use revert_error()? Then this
would be less of a hassle.

Besides, I would interpret "cannot revert nonexistant directory" such
that the directory is not in the repository at all?!

> 'error using subcommand' just seems cryptic to me.

Agreed. That's why I asked for suggestion. ;-)

Bye,

        Benjamin.