Re: [gui] Implement RescaleOp

[Andrew Haley <aph at redhat dot com>]

Mark Wielaard writes:
 > 
 > On Thu, 2004-09-30 at 11:52, Andrew Haley wrote:
 > > Mark Wielaard writes:
 > 
 > >  > As long as we are not completely merged and gcj doesn't just use GNU
 > >  > Classpath glibj out of the box we really need people to post the actual
 > >  > patches so they can be checked and merged with upstream easily.
 > >  > 
 > >  > That, or we will have to ask libgcj hackers to explicitly propose and
 > >  > post patches on classpath-patches if they want them to be integrated.
 > > 
 > > The patches mailing list is for people to read.  We try to make sure
 > > that patches are in a form that can be merged, but its primary purpose
 > > is for patch discussion.  
 > 
 > Fair. But a little note when what/if you commit without posting the
 > precise patch would be appreciated.
 > 
 > > As far as I am awre, it has always been customary to accept patches
 > > with a caveat of "please fix the indentation before check-in" and such
 > > patches are not re-posted.  You are proposing a change of policy.
 > 
 > Maybe. I always understood a "OK, but with these changes" to mean "OK,
 > if you post the cleaned up patch, with these changes, you can commit it
 > without waiting for approval".

Maybe you did, but I believe you're wrong because that this has never
been the case, and you're also wrong in that adopting this policy
would add clutter to the list.  

If I notice a small indentation error in a patch I wrote or a small
error in a comment I won't re-post the patch because it would waste a
lot of valuable people's time.

Of course, this problem is caused by the fact that CVS doesn't have
atomic multi-file commits.  However, it ought to be possible to write
a tool that extracts patches from CVS logs, which is presumably what
is really needed.

Andrew.