[PATCH] Support remote access to (some) "hgrc" parameters

Thomas Arendsen Hein thomas at intevation.de
Wed Jan 9 02:01:54 CST 2008


* Glenn Ammons <ammons at us.ibm.com> [20080108 21:15]:
> "Thomas Arendsen Hein" <thomas at intevation.de> writes:
> > * Glenn Ammons <ammons at us.ibm.com> [20080103 10:32]:
> >> To answer your second question, there is no relation between
> >> parameters in [public.foo] and those in [foo].  They are separate
> >> sections.  The first one is accessible remotely.
> >
> > This sounds unfortunate as the public values are kind of useless
> > then. I'd rather like to see e.g. a [public] section (or use the
> > same name as the extension which provides this) and list whole
> > sections or single parameters tp publish here.
> 
> I'm not sure that I understand the objection.  What do you mean by
> "kind of useless"?
> 
> One nice property of the "public." convention is that extensions that
> don't expect public parameters stop working if their parameters are
> moved to a public section.  For example, if the "bank" extension looks
> for its password parameter here:

That's the problem, I don't think every extension should have to
know to additionally look at the public.foo section. You can't
prevent user errors by that behaviour, because if the user puts a
password in a public section, it will be accessible, even if the
code doesn't use it.

> Another nice property of "public." is that whether a section is public
> or not is obvious from the name of the section.

Hmm, maybe. My fear is that you might run into too many confusing
sections and therefore put a value in the wrong one.

Thomas

-- 
thomas at intevation.de - http://intevation.de/~thomas/ - OpenPGP key: 0x5816791A
Intevation GmbH, Osnabrueck - Register: Amtsgericht Osnabrueck, HR B 18998
Geschaeftsfuehrer: Frank Koormann, Bernhard Reiter, Dr. Jan-Oliver Wagner


More information about the Mercurial-devel mailing list