On Fri, 2003-02-07 at 13:30, Karl Fogel wrote:
> Was it Brane who pointed out earlier that this function is really a
> generic looper to match a path across fnmatch patterns? Not saying it
> should be moved right now; just making a mental note that if we find
> another situation where we're matching a filename against a list of
> patterns, probably we should genericize this and move it to svn_path.h
> and libsvn_subr/path.c...
We already have a very similar function, svn_cstring_match_glob_list, a
function I introduced when fiddling with svn_config. The only
difference was that match_glob_list was taking a comma-separated string
of glob patterns instead of an apr_array_header_t. I decided that
match_glob_list should be more general anyway, so I modified it to take
an apr_arrary_header_t and nuked svn_wc_is_ignored in its favor.
If people (and in particular Eric) think it would be better to have
svn_wc_is_ignored as a one-line function for clarity, then I apologize
and can add it back in on request.
(It's a question of phenomenology, really. From a reductionist
perspective, svn_wc_is_ignored is the same as
svn_cstring_match_glob_list, and having one pass its arguments to the
other is a gratuitous layer of abstraction. But from a holistic
perspective, you might argue that it's not; it performs the logical
function of checking a string against an ignore list which merely
happens to be of the form accepted by match_glob_list. Normally I like
to think I am holistic programmer, but for some reason the holistic
argument isn't sitting well with me in this case, so I chose the
reductionist perspective. </philosophy>)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Feb 8 00:06:58 2003