Re: [patch 0/3] readline-6.2 rebase

[Jan Kratochvil <jan.kratochvil@redhat.com>]

On Mon, 09 May 2011 19:00:13 +0200, Tom Tromey wrote:
> FWIW I think this is the best time to check this in, because the release
> branch has been made.

Do you think it should be checked in as separate patches (to unchanged
readline-6.2 with each of the patches applied afterwards) or as a single
commit?

The previous import was a single-commit one with sourceware patches already
contained:
	commit c6fad60c72bac0b0ba1b7912f866caa3a6cb744c
	Author: Daniel Jacobowitz <dan@debian.org>
	Date:   Fri May 5 18:26:14 2006 +0000
	    Readline 5.1 import for HEAD.

Advantage of the single commit is it does not introduce intermedia regressions
which makes it compatible with `git bisect'.  (OTOH git bisect has various
issues with building older snapshots on recent systems and also I am not sure
if all the multi-part patches are per-piece regression+buildability safe.)

Advantage of the multiple commits is sure more clear import for futher rebases
and reviews.  Porting the readline-5.1 patchset was a bit of reverse
engineering with GIT.

(I am more for the latter choice now; despite I proposed the former before.)


Thanks,
Jan