Re: auto-generated parts in posted patches (Re: [patch 2/2+rfc+doc] Install gcore by default (+new man page))

[Tom Tromey <tromey@redhat.com>]

>>>>> "Pedro" == Pedro Alves <palves@redhat.com> writes:

Jan> Forgot to state here were always replies one should strip all
Jan> auto-generated parts from posted patches - so I do it.  But I find
Jan> that inconvenient.  It is easier to skip a part of patch when
Jan> reviewing it than to regenerate it hopefully the same way as the
Jan> patch submitter because the source tree requires special autotools
Jan> (old version + unpatched).

Pedro> I fully agree.  It's also very useful to see the generated bits
Pedro> to check whether a submitter who has commit/write access is using
Pedro> the right versions of the tools to regenerate the patches, and
Pedro> check whether the correct set of files is being regenerated.

I'm also not averse to changing this rule.

One drawback is that it would mean a different rule for gdb than for
other parts of the toolchain.  But that is already true in various ways.

I'd like it if non-trivial series were available in git somewhere.  Then
if someone really wanted to try the patch out it would be available
without re-applying.

Tom