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

Re: svn commit: r1089058 - in /subversion/trunk/subversion: libsvn_wc/wc_db.c tests/libsvn_wc/op-depth-test.c

From: Philip Martin <philip.martin_at_wandisco.com>
Date: Tue, 05 Apr 2011 16:41:15 +0100

rhuijben_at_apache.org writes:

> This change also slightly changes the behavior of delete and revert on
> children of a copy operation as it allows you to revert the separate commands
> that will be performed on commit.

I suppose revert removes the higher op-depth child leaving the
not-present child at the parent's op-depth. But that not-present still
counts as deleted and will generate a delete on commit. It's not really
new behaviour, with the old model it was possible to delete the
mixed-rev child (creating a new base-deleted layer) and get the same
behaviour. What is different is that in the old model one could revert
the delete and go back to the mixed-rev child being present, in the new
model it is not possible to revert.

When the mixed-rev child is deleted (either as not-present in the new
model or base-deleted in the old) one runs into the problem of
copy_tests.py 89, mixed_rev_copy_del, that the client doesn't know
whether or not to send the delete.

> On top of that this commit will finally allow the commit processing to use
> depth handling.

Not sure what you mean by "depth handling" but commit is still going to
have to detect the not-present child at the same op-depth as the parent
and generate a delete if there is no higher op-depth for the child.

Received on 2011-04-05 17:41:57 CEST

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.