On Sat, 2009-11-07, Daniel Näslund wrote:
> On Fri, Nov 06, 2009 at 01:57:51PM -0800, Daniel Rall wrote:
> > On Fri, Nov 6, 2009 at 10:39 AM, Daniel Näslund <daniel_at_longitudo.com> wrote:
>
> New log:
>
> [[[
> Remove use of deprecated svn_path_join().
>
> * subversion/libsvn_client/mergeinfo.c
> (svn_client__adjust_mergeinfo_source_paths): Replace svn_path_join()
> with svn_uri_join(). An svn_mergeinfo_t key is the "base" parameter.
> (svn_client__get_wc_mergeinfo): Replace svn_path_join() with
> svn_relpath_join(). The "base" parameter is created with a relative
> path from svn_dirent_dirname().
> (svn_client_mergeinfo_get_merged): Replace svn_path_join() with
> svn_uri_join(). target_repos_root is the "base" parameter.
>
> Patch by: Daniel Näslund <daniel_at_longitudo.com>
> Suggested by: dlr
> ]]]
>
> > > Some background: In subversion/include/svn_mergeinfo.h svn_mergeinfo_t
> > > is just a typedef for a apr_hash_t. The keys are described like this:
> > >
> > > * (c) @c svn_mergeinfo_t, called "mergeinfo". A hash mapping merge
> > > * source paths (@c const char *, starting with slashes) to
> > > * non-empty rangelist arrays. A @c NULL hash is used to represent
> > > * no mergeinfo and an empty hash is used to represent empty
> > > * mergeinfo.
> > >
> > > I'm assuming that means a dirent.
> >
> > It's a path fragment. In this case, the path is relative to the
> > repository root. While you may be able to use it as a local file
> > system path in some situations, it's really a URL fragment.
It is not _really_ a URL fragment, it is a path fragment that
_represents_ a URL fragment.
> After some reading I agree. But, they keys in svn_mergeinfo_t is not
> uri_encoded! Or atleast they don't say anything about beeing
> uri-encoded and when svn_client_mergeinfo_get_merged wants a hash of
> full urls, the keys are being uri-encoded before the join! And in
> logs_for_mergeinfo_rangelist() the following lines says that keys are
> required to be repository-absolute but does not uri-encode the string:
>
> [[[
> /* FILTER_LOG_ENTRY_BATON_T->TARGET_MERGEINFO_CATALOG's keys are required
> to be repository-absolute. */
>
> if (apr_hash_count(target_mergeinfo_catalog))
> {
> apr_hash_index_t *hi;
> svn_mergeinfo_catalog_t rekeyed_catalog = apr_hash_make(scratch_pool);
>
> for (hi = apr_hash_first(scratch_pool, target_mergeinfo_catalog);
> hi;
> hi = apr_hash_next(hi))
> {
> const char *path = svn_apr_hash_index_key(hi);
>
> if (!svn_dirent_is_absolute(path))
> apr_hash_set(rekeyed_catalog,
> svn_dirent_join("/", path, scratch_pool),
> APR_HASH_KEY_STRING,
> svn_apr_hash_index_val(hi));
> }
> target_mergeinfo_catalog = rekeyed_catalog;
> ]]]
>
> What I'm saying is that if repository-absolute means that it's an uri
> (generalization of url) then it should be uri-encoded, right?
No. When we call it a "repository-relative path" or a
"repository-absolute path" (both meaning the same thing) we mean it is a
plain path with no URI encoding.
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2416127
Received on 2009-11-10 13:37:35 CET