Re: [RFC] series headers

Daniel Barkalow [off-list ref] writes:

I'd like to be able to get format-patch to produce a [PATCH 0/N] message, 
with a message that's actually in my repository, plus various goodies 
generated either from diffing the ends of the series or by running through 
the log an extra time to pick up summary information.

So far, so good.

As far as implementing this... would it be sane to make struct 
rev_info.commit_format a callback, so that the code to generate an email 
message can be somewhere that's easy to use to generate an email that 
isn't for a commit in the log? I don't *think* git's quite fast enough for 
the indirect jump to a callback instead of an if tree for an enum will 
actually hurt us.

I suspect that temptation to touch rev_info.commit_format arises
purely because you are thinking about making 0/N a (perhaps
fake) commit.  I do not see a point in that.

What is the workflow?

 $ work work work, commit commit commit, reorder and perfect
 $ git tag end-of-series
 ... in $EDITOR, edit the [0/N] message
 $ git format-patch origin..end-of-series
 ... which notices end-of-series, and perhaps internally runs
 ... $ git-shortlog origin..end-of-series
 ... $ git diff --stat --summary origin..end-of-series
 ... $ git cat-file tag end-of-series
 ... to create 0/N which it did not do so far in
 ... 0000-cover-letter.txt
 $ git-send-email 0*.txt

Would it be so much better than this workflow which would
probably not need to touch much of the 'commit formatting" code
that is used for [1/N]..[N/N] messages?

 $ work work work, commit commit commit, reorder and perfect
 $ git format-patch --with-cover origin..HEAD
 ... which notices --with-cover, and perhaps does
 ... $ git-shortlog origin..HEAD
 ... $ git diff --stat --summary origin..HEAD
 ... $ echo "*** BLURB HERE ***"
 ... to create 0/N which it did not do so far in
 ... 0000-cover-letter.txt
 $ $EDITOR 0000-cover-letter.txt
 $ git-send-email 0*.txt