On Sun, 2004-05-02 at 09:13, C. Michael Pilato wrote:
> Doh! I was thinking of svn_fs_parse_id(). Sorry about that. But I
> suppose we keep that public for completeness' sake.
Apologies for the confusion. svn_fs_parse_id() is the problematic one
(because it creates an ID from a string, with no FS object to mediate).
It appears to be used only in changes-test.c and in the FS modules
themselves.
(Branko's comment was almost certainly meant to apply to parsing IDs, so
that they can be stored in the working copy, but as he pointed out, his
comment isn't necessarily all that relevant since we can add the APIs we
need at such a time as we start doing that.)
We can't nuke svn_fs_parse_id from the API in 1.x, but we can certainly
worry less about its quality of implementation if we think nobody is
going to use it. So, for now, I will plan to make it so that
svn_fs_parse_id() is deprecated and works only with the base FS
library. If anyone asks for that functionality (I will ask clkao if he
uses it), we will add an svn_fs_parse_id2() which takes an FS object and
works with any FS type.
It also sounds like consensus centers around a libsvn_fs_subr library.
Max appears to be working on that.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon May 3 18:03:49 2004