On Mon, 2010-09-13 at 10:53 +0100, Philip Martin wrote:
> Senthil Kumaran S <senthil_at_collab.net> writes:
> > In the previous versions of subversion ie., 1.5.x or 1.4.x the keyword
> > will not expand, but this is not the case with 1.6.x in which it gets
> > expanded for the first update. This is what the issue #3471 is about
> > and the committed patch tests that.
> I'm still confused. The test creates an unexpanded keyword and then
> runs update and the keyword gets expanded. Are you saying that update
> should not expand the keyword, because that's the way 1.5 behaved?
> I think it's right that update expands the keyword, and so when the
> timestamp changes that is also correct.
> What is a bug is that a *second* update causes the file's timestamp to
> get changed a second time. A second update doesn't change the file
> contents and so the timestamp should not change. But the test doesn't
> run a second update. If the test did run a second update it could use
> do_sleep_for_timestamps rather than sleep.
In 1.5.x & 1.6.x, the keyword is expanded for the first update and
timestamp is changed for further updates. Here We are just locking the
file with keywords property set and we are not committing the file.Then
running "svn update" should expand the keyword & change the timestamp?
But in 1.4.x , the keyword will not expand and timestamp won't change
for first update.
Thanks & Regards,
Received on 2010-09-13 13:02:22 CEST