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

svn commit: r21797 - in trunk: . subversion/tests/cmdline subversion/tests/cmdline/svntest

From: David Glasser <glasser_at_mit.edu>
Date: 2006-10-06 16:39:19 CEST

On 10/6/06, Garrett Rooney <rooneg@electricjellyfish.net> wrote:
> On 10/6/06, David Glasser <glasser@mit.edu> wrote:
> > It looks like this doesn't pass on the Debian buildbot or my Debian
> > machine; it works on OSX. Playing around by hand seems to show that
> > the editor script is being called, and it is properly editing the
> > file, but svn is not noticing. Any ideas?
> Timestamps? Are we just not noticing that the file was edited?

Yup, that's it: svn_cl__edit_externally assumes nothing was edited if
the mtime and size hasn't changed. It's not clear to me that using
this heuristic wins you that much hear, where it's one relatively
small file (as opposed to for "svn st" or whatever).


* This is a bug and should be fixed by removing all the mtime and size
checks from svn_cl__edit_externally and just using an equality check.
Sounds easy, except for dealing with LF conversions etc.

* This is a bug and should be fixed by adding svn_sleep_for_timestamps
to svn_cl__edit_externally. This does work, but I don't really think
we want every single command that asks for a log message to take an
extra second. (This is probably never an issue for log messages
anyway since you really ought to always change the size; I can
certainly imagine propedits that don't change the size and that you
can do very quickly.)

* This isn't enough of a bug to be worth fixing and I should just add
a sleep to the mock editor in the test suite (and a comment to



David Glasser | glasser_at_mit.edu | http://www.davidglasser.net/
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Oct 6 16:39:43 2006

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.