Just to set things straight ...
this isn't a talk about nitpicking , it's a talk about avoiding regressions
it's imho easier to rollback or go through the diffs of suspect changeset
than traverse several big changesets having non related diffs inside and
split them after pushed
divide and conquer before the push, it's that simple
if you have a better ideas how to avoid regressions, then I'd like to
know them
for the below, using hg split and mq the rollbacks aren't that big deal
the big deal is to identify the code and split changesets, if you want
to keep parts of them ...
hmm?
L
On 22.2.2012 15:12, Jens Elkner wrote:
On Wed, Feb 22, 2012 at 08:45:32AM +0100, Trond Norbye wrote:
On Tue, Feb 21, 2012 at 6:49 PM, Jens Elkner
<[1]jel+openg...@cs.uni-magdeburg.de> wrote:
Ehmm, I think, there's not really something to split/doesn't make sense.
So you think a changeset that claims to improve logging should contain
reformatting of code completely unrelated to the logging change belongs in
the same changset and it doesn't make sense to split that into two different
changesets?
Not really (but doesn't hurt) and yes.
What I think is, it is just a huge waste of time to rollback all changes
and painfully re-merge stuff, just because one discovered, that a
autosave function or shortcut just did some MINOR reformatting. That's
IMHO really unproductive and I actually have no time for such nitpicking.
If you think, inappropriate formatting slipped through, just hit Ndd or
whatever your favorite editor provides and you'r done. No need to waste
anyones time ...
Regards,
jel.
_______________________________________________
opengrok-discuss mailing list
opengrok-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/opengrok-discuss
