On Mon, 2010-03-29, Greg Stein wrote: > On Mon, Mar 29, 2010 at 13:10, Julian Foad <[email protected]> wrote: > > The attached patch is an attempt to close the gap between "transmit text > > deltas" and commit post-processing, by passing the temporary new text > > base file paths around. It succeeds in that, at least it looks right > > and passes the test suite. > > except for that printf() :-P > > > The part of this patch that I haven't finished is with back-compat of > > svn_wc__process_committed_internal(), and what is the difference between > > its 'queue' parameter and its checksum/recurse/no_unlock/etc. > > parameters, being values which are alternatively available in the queue. > > > > svn_wc__process_committed_internal() is called by > > svn_wc_process_committed_queue2() which passes the 'queue' param, and > > also by the deprecated svn_wc_process_committed4() with QUEUE=NULL. I > > had been assuming that if QUEUE==NULL then all the parameters that are > > available in the queue (checksum for one) are not available, but that's > > not how the back-compat wrapper wants to work. I'll need to fix that. > > I think the right thing to do is to re-write the wrapper > > (svn_wc_process_committed4()) to construct a new queue with one item and > > pass that, and stop having the other parameters (checksum etc.) passed > > as separate parameters. I'll look at that tomorrow. I may already have > > committed changes that break that back-compat; I'll check.
Well, at least I checked that I don't seem to have made it worse by anything I've done recently. Still digging in history. - Julian
