Re: svn commit: r1614080 - in /subversion/branches/authzperf: BRANCH-README subversion/libsvn_repos/authz.c subversion/libsvn_subr/config.c subversion/libsvn_subr/config_impl.h

Branko Čibej <brane_at_wandisco.com> writes:

> Let's please have the rules documented in the wiki, then we can comment
> on them.
>
>> * Select the path rule(s) that cover the largest section of the path.
>> This is what we do today and there can be only one such match
>> w/o the use of wildcards.
>>
>> * If there is more than one such rule, apply the one declared last.
>> This seems to be the easiest rule to follow and understand.
>
> Well, I disagree, I'd expect an exact match to override a wildcard match.

That seems reasonable but raises the expectation that there is a more
general order for wildcards. Given a path:

  /A/B/C/D

is either of the patterns:

  /A/B/C/*
  /*/B/C/D

an override for the pattern:

  /*/*/*/*

?

Does either of the patterns /A/*/*/D and /*/B/C/* override the other?

If we choose and release some simple wildcard behaviour how easy will it
be to implement more sophisticated behaviour in future releases?

-- 
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*