[PATCH 0 of 2] The ui.ignore config option, one more time
Danek Duvall
danek.duvall at oracle.com
Thu Sep 19 17:39:07 UTC 2013
Isaac Jurado wrote:
> Replying Thomas Arendsen Hein:
> > * Laurens Holst <laurens.nospam at grauw.nl> [20130918 12:07]:
> >> Op 12-09-13 00:29, Isaac Jurado schreef:
> >>> This is another attempt to change the behaviour of ui.ignore.*
> >>> configuration option to a more useful one.
> >>>
> >>> The idea is to leave absolute paths as they are, but interpret
> >>> relative paths from the repository root instead of the current
> >>> working directory. Specially since the latter can be emulated with
> >>> something like:
> >>>
> >>> [ui]
> >>> ignore = $PWD/custom-ignore
> >>
> >> IMO it should be relative to the configuration file in which it is
> >> specified.
> >
> > Instinctively I would have assumed that they are relative to the repo
> > root, so something like this in ~/.hgrc would be possible:
> >
> > [ui]
> > ignore = .hg/localignore
> >
> >> (See previous discussions.)
> >
> > But reading this previous discussions I am no in favour of reading it
> > relative to the config file to match the behaviour of %include, even
> > if this differs from the behaviour of [paths].
>
> I hope it is obvious that the current behaviour is not really practical,
> specially since it can be achieved with the PWD environment variable
> (however, I'm not sure if that works on Windows too).
>
> And given that it is not documented *at all*, changing its semantics
> does not seem so grave. I wonder how many users are relying on the
> current behaviour for relative paths.
>
> Having said that, I can't decide whether it is better to be relative to
> the current configuration file or to the repository root. I've never
> used %include myself so for me it feel more natural basing relative
> paths on the repository root, and also much easier to change as it can
> be seen in the patch.
>
> Any more opinions?
Two different config tokens, one for each behavior? Or a prefix like we
have for file patterns ("re:", "glob:", etc)?
Danek
More information about the Mercurial-devel
mailing list