On Fri, Mar 23, 2018 at 10:12 AM, Johannes Schindelin <johannes.schinde...@gmx.de> wrote: > Hi Wink, > > On Thu, 22 Mar 2018, Wink Saville wrote: > >> The backend scriptlets for "git rebase" were structured in a >> bit unusual way for historical reasons. Originally, it was >> designed in such a way that dot-sourcing them from "git >> rebase" would be sufficient to invoke the specific backend. >> >> When it was discovered that some shell implementations >> (e.g. FreeBSD 9.x) misbehaved when exiting with a "return" >> is executed at the top level of a dot-sourced script (the >> original was expecting that the control returns to the next >> command in "git rebase" after dot-sourcing the scriptlet). >> >> To fix this issue the whole body of git-rebase--$backend.sh >> was made into a shell function git_rebase__$backend and then >> the last statement of the scriptlet would invoke the function. >> >> Here the call is moved to "git rebase" side, instead of at the >> end of each scriptlet. This give us a more normal arrangement >> where the scriptlet function library and allows multiple functions >> to be implemented in a scriptlet. >> >> Signed-off-by: Wink Saville <w...@saville.com> >> Reviewed-by: Junio C Hamano <gits...@pobox.com> >> Reviewed-by: Eric Sunsine <suns...@sunshineco.com> >> --- >> git-rebase--am.sh | 11 ----------- >> git-rebase--interactive.sh | 11 ----------- >> git-rebase--merge.sh | 11 ----------- >> git-rebase.sh | 2 ++ > > The patch makes sense to me. > > Thanks, > Johannes
Junio, Eric and Johannes, thanks for the help!!! I've created v5 with the two patches, what is the suggested format-patch/send-email command(s)? Here is one possibility: git format-patch --cover-letter --rfc --thread -v 5 --to=git@vger.kernel.org --cc=sunsh...@sunshineco.com --cc=johannes.schinde...@gmx.de -o patches/v5 master..v5-2 If this was the first version then the above would seem to be a reasonable choice. But this is version 5 and maybe I don't need --cover-letter which, I think means I don't want to use --thread. If that's the case should I add --in-reply-to? But that leads to the question. from which message should I get the Message-Id? More likely I'm totally wrong and should do something completely different, advice appreciated. -- Wink
