Robert Haas escribió:
> But if there are patches against that code, then they've been broken
> now and they will break again when the pgindent run is done. If the
> indentation is fixed at commit-time (or before someone goes to the
> trouble of fixing them), then they get broken only once. I guess it's
> not the end of the world, but it sure seems like the less work
> pgindent does when it is run, the better.
I think that we should be looking at making pgindent runnable by lone
hackers at home in their patched trees. That way they fix their patches
by simply running it when it's run on the CVS tree, before doing an
"update". That should remove/reduce the merge problems.
... at least in CVS; not sure what would happen if this was done in a
GIT repository. (It would probably require a rebase, but then what do I
know about GIT?).
BTW I think it's better to redirect this kind of discussion to -hackers.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support