apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daniel Rall <...@apache.org>
Subject Re: svn commit: r26422 - trunk/subversion/include
Date Wed, 05 Sep 2007 16:58:54 GMT
On Wed, 05 Sep 2007, Malcolm Rowe wrote:

> On Tue, Sep 04, 2007 at 03:49:35PM -0700, Eric Gillespie wrote:
> > > * subversion/include/svn_config.h
> > >   (SVN_CONFIG_DEFAULT_GLOBAL_IGNORES): Remove two redundant patterns.  (Note
> > >     that, unlike many Unix shells, fnmatch will allow metacharacters to match
> > >     leading periods).
> > 
> > Where is this documented?  We should provide a pointer from here
> > to to the fnmatch documentation.  apr_fnmatch.h doesn't seem to
> > explain it though.  If it isn't documented anywhere, we should at
> > least document this much ourselves.
> > 
> 
> Yeah, it's pretty underdocumented.  If we assume that apr_fnmatch() is
> just an extension of the POSIX fnmatch() function (APR's documentation
> doesn't say), then see e.g. here:
> 
> http://www.opengroup.org/onlinepubs/000095399/functions/fnmatch.html
> 
> "If FNM_PERIOD is set in flags, then a leading period ('.') in string
> shall match a period in pattern [... some verbiage and references that
> define a leading period and imply that it's not matched by other
> metacharacters when this flag is set ...].  If FNM_PERIOD is not set,
> then no special restrictions are placed on matching a period."
> 
> We pass no flags, so we don't get this special behaviour.  We could
> alternatively revert my change and pass in APR_FNM_PERIOD if we think
> that the current behaviour is a bug.

It'd be nice if APR's header files documented the behavior...

Mime
View raw message