On 8 Jun 2010, at 11:41, Eric Kow wrote:
On Tue, Jun 08, 2010 at 05:27:48 +0300, Dan Pascu wrote:
Keep in mind that I do not know if the error occurred when there was
something to pull or it was a dummy pull that did nothing because
there were no remote changes. So we do not know if the error was
correlated with a certain patch, or just the fact that the script
running pull had no terminal.
Could we try to narrow this down by only running dummy pulls,
for example, setting it up so that it's always pulling from an
unchanged repo?
We could, but that is a production environment I will not mess with
for such tests.
Dan: do you already have the infrastructure in place to
automatically
log the Darcs errors?
You can say so, but it's something cron does by default. If there is
some output from the commands it is emailed to the administrator.
OK, so I think what I vaguely had in mind was maybe some sort of
procmail recipe that dumped all of the darcs messages into a single
box that can be scrutinised...
If you're asking if you can use this environment for testing, them I'm
afraid not. It's a production environment running private company
projects, so it cannot be used for testing nor can I provide access to
it.
However I get the email from cron, so I can report if I see anything
unusual, but I cannot automatically redirect all that email to a
public address everyone can read.
--
Dan
_______________________________________________
darcs-users mailing list
darcs-users@darcs.net
http://lists.osuosl.org/mailman/listinfo/darcs-users
