[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: redundant path functions

From: C. Michael Pilato <cmpilato_at_collab.net>
Date: Wed, 31 Mar 2010 08:15:14 -0400

Stefan Sperling wrote:
> I too remember being confused as to which interface I should be using.
> And I think having is_ancestor not be the reverse of is_child is a bad idea.
> We'll likely find mis-use of either interface in the code base, where
> the author didn't realise the subtle semantics of the interface.
>
> +1 on settling on one variant (I don't really care which), deprecating
> the others, and updating callers.

We had one variant before -- svn_path_is_child(). But I think people got
tired of having to implement _is_ancestor() as a pool-requiring wrapper
around that function and a new public API was born.

The is_child() variant is clearly the most powerful one, having not just the
ability to answer the question but also to return the path relative to the
ancestor (which is used in various places in our codebase). But perhaps
there's a unified API that do this for us:

   svn_*_check_ancestry(svn_boolean_t *is_ancestor,
                        const char **child_relpath, /* may be NULL */
                        const char *path1,
                        const char *path2,
                        apr_pool_t *result_pool) /* may be NULL */

The function always answers the is-ancestor question, and -- if you provide
a non-NULL child_relpath and result_pool, it will fill in the child path
relative to the ancestor for you.

Is that sane?

-- 
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Received on 2010-03-31 14:15:50 CEST

This is an archived mail posted to the Subversion Dev mailing list.