[issue166] Add per-user hgignore

Thomas Arendsen Hein thomas at intevation.de
Thu Mar 16 08:18:07 UTC 2006


* Bryan O'Sullivan <bos at serpentine.com> [20060316 06:10]:
> On Wed, 2006-03-15 at 16:38 -0500, Colin McMillen wrote:
> 
> > [ignore]
> > glob  = BLAH foobah
> > glob2 = FOO
> > re    = .*LA$
> > bad   = ignored
> 
> I'm sorry, but this syntax is just horrible.

In the tracker I suggested glob and glob.something (i.e. separating
with a dot).

> Why not do something like this instead?
> 
>         [ui]
>         ignore = /path/to/ignore/file
>         ignore.other = /path/to/other/file

Because then you have to write:
ignore.1 = glob:BAH
ignore.2 = glob:foobah
ignore.a = glob:FOO
ignore.b = re:.*LA$

> This uses the same naming syntax, and probably machinery, as hooks (i.e.
> the files are read in lexicographical order of key names).  It can also
> use the existing hgignore parsing machinery.  And it's not deeply
> unnatural and weird :-)

The same naming syntax as hooks is good, and this is what I
suggested in http://www.selenic.com/mercurial/bts/issue166
But the hooks are not written like this:
[ui]
hook.changegroup.foo = jkdfhdjk
hook.commit.bar = vjhbdbn

lexicographical order of key names isn't important here, as we only
do ignore and never unignore/include.

Thomas

-- 
Email: thomas at intevation.de
http://intevation.de/~thomas/



More information about the Mercurial mailing list