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

Re: CVS update: subversion/subversion/libsvn_wc adm_ops.c

From: Greg Stein <gstein_at_lyra.org>
Date: 2001-03-10 01:45:08 CET

On Fri, Mar 09, 2001 at 02:35:20PM -0000, sussman@tigris.org wrote:
> +/* Publically declared, so libsvn_client can pass it off to the RA
> + layer with svn_wc_bump_target(). */
> +struct svn_wc_bump_baton

If WC returns a "struct svn_wc_bump_baton *", then you don't need to make it
public. Code won't be able to deref the pointer, unless it includes a WC
private header that defines the structure.

> +/* This is the "new" callback that the RA layer uses to bump each
> + committed target, one-at-a-time. It's a function of type
> + svn_ra_close_commit_func_t.

Hmm. Isn't it more correct to say that the client library passes this
function to RA for use as a close-commit-func? The wording above seems to
indicate that RA directly calls this function.

[ I believe the theory is that no RA module ever calls a WC function; that
  appears to be true with ra_dav and ra_local ]

> + Eventually, the "track" editor's close_edit() routine needs to call
> + this too, for those times when it's specifically comingled with the
> + XML-output editor. (svn_wc_close_commit and its helpers will then
> + be deprecated!) */

Why does the track editor need to do anything? From the comment above, it
sounds like there is a special case in the track editor for when it is
combined with the XML editor? That's not cool...


Greg Stein, http://www.lyra.org/
Received on Sat Oct 21 14:36:25 2006

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