Jonathan Nieder [off-list ref] writes:
How is the cherry-pick event different from the send-by-mail-and-apply
event?
In both cases, the result has a distinct commit id and distinct
signoff and it is unlikely that the previous patch handler was testing
with the same tree as the next one. (And each patch handler should add
relevant comments if the new situation warrants that.)
Fair enough. If that is the direction we would want to go, perhaps it
suggests that we might eventually want to use "Cherry-picked-from: " as
the marker for this information?
And if we go that route, and if this information is being used, we have a
rather serious backward compatibility problem. Older scripts will break,
and this cannot be handwaved away with a configuration option or a command
line switch (even if you personally choose to keep using the old format,
you may get a commit with the new style trailer from elsewhere).
Hmm.