From: Yann Dirson Date: Sun, 08 Jan 2006 10:06:54 GMT Subject: Re: [PATCH 7/7] Add an optional limit to git-rebase Message-ID: <20060108100654.GF32585@nowhere.earth> In-Reply-To: <7vk6db12z7.fsf@assigned-by-dhcp.cox.net> On Sun, Jan 08, 2006 at 01:30:20AM -0800, Junio C Hamano wrote: > Yann Dirson writes: > > > This patch adds a limit to restrict the list of patches to rebase. > > > > This is useful when a set of patches were done against a given head, > > but needs to be ported to a different head, as opposed to being ported > > to a descendant of the original head. In such a case we only want to > > port our own patches, not those that make the two branches different > > upstream. > > This usage makes sense, independent from your git-format-patch > "limit" patch. The patched argument to git-format-patch does > not match your git-format-patch changes AFAICT ;-). D'oh, I knew I should always written a testsuite when changing the implementation of a feature at such a late time in night. > > -git-format-patch -k --stdout --full-index "$other" ORIG_HEAD | > > +git-format-patch -k --stdout --full-index "$other" ORIG_HEAD $limit | > > If all you want to do is to move the cut-off point more recent > than the merge base is, how about (without your git-format-patch > changes) doing something stupid like this instead? > > git-format-patch -k --stdout --full-index ${limit:-"$other"} ORIG_HEAD Right, that would even be closer to the version that had indeed been tested - it does have the advantage of passing t3401, though :) -- Yann Dirson | Debian-related: | Support Debian GNU/Linux: | Freedom, Power, Stability, Gratis http://ydirson.free.fr/ | Check