This commit is contained in:
Joey Hess 2018-06-04 12:12:56 -04:00
parent ac06c34d05
commit 95f7295b67
No known key found for this signature in database
GPG key ID: DB12DB0FF05F8F38
2 changed files with 24 additions and 0 deletions

View file

@ -163,6 +163,8 @@ addCopies = addLimit . limitCopies
limitCopies :: MkLimit Annex limitCopies :: MkLimit Annex
limitCopies want = case splitc ':' want of limitCopies want = case splitc ':' want of
-- Note that in case of a group having the same name as a trust
-- level, it's parsed as a trust level, not as a group.
[v, n] -> case parsetrustspec v of [v, n] -> case parsetrustspec v of
Just checker -> go n $ checktrust checker Just checker -> go n $ checktrust checker
Nothing -> go n $ checkgroup v Nothing -> go n $ checkgroup v

View file

@ -0,0 +1,22 @@
[[!comment format=mdwn
username="joey"
subject="""comment 1"""
date="2018-06-04T16:06:57Z"
content="""
Thanks for pointing out this ambiguity, which I don't remember having
ever considered.
The code for this does look for the name of a trust level first,
and only a group name if it's not a trust level. So, adding a group
with a colliding name will never change the behavior of such a command
or preferred content expression.
The only problem then would be that you couldn't match
on a group by that name. But if you run into that problem,
you can simply rename your group.
So I don't think that git-annex needs a sanity check, really.
(I have added a comment in the code to make clear that the order of
parsing this does matter.)
"""]]