unity-dev team mailing list archive

[Ted Gould <ted@xxxxxxxxxx>]

On Mon, 2013-01-07 at 16:42 +0100, Didier Roche wrote:

> Le 07/01/2013 16:21, Ted Gould a écrit :
> > I think that's the issue, you should be saying "We need to fix bug 
> > 1234" not "Trunk was known broken, why aren't you focusing on my 
> > problem RIGHT NOW."  There might be other, higher priority work being 
> > done.  Or perhaps people are in the middle of longer term tasks that 
> > they need to complete.  Asking for the bug to prioritized to the next 
> > possible slot is a reasonable request, saying that development should 
> > have stopped is not.  It is reasonable to not fix bugs immediately, 
> > and do other work first in many cases even if that makes some options 
> > for fixing them more expensive (reverting is harder the longer you wait).
> 
> So, in that case, if a commit introduced a regression (regression being 
> maybe that we can't build in the next day trunks), then we revert the 
> guilty change. So reverting right away to not have this "reverting is 
> harder later on".


I think it makes more sense to file the bug, and then figure out what
the appropriate action is.  Certainly reverting might be the solution,
also just fixing it might be easier.  Reverting, to me, is saying "the
autolander isn't working".  If the autolander allows regularly something
to be committed to trunk that we need to revert we need to fix it; we
should be rejecting the merge not reverting later.

Reverting shouldn't happen on trunk, we should be rejecting earlier in
the development process.

Ted

Attachment: signature.asc
Description: This is a digitally signed message part