Jeff King <peff@xxxxxxxx> writes: > This started off with Michael's patch to sign_buffer, which is at the > tip, and then me trying to address the possible deadlocks there and in > verify_signed_buffer. While I was in the area, I took the opportunity to > do a few cleanups. > > It's unclear to me whether the deadlocks are possible in practice; see > patch 5 for discussion. I do recall thinking about the verification side and coming up with the same conclusion as yours when we queued that code (i.e. they need to read the whole thing before checking). > My guess is probably not, but the amount of code > to support doing it right is not all that much. But if we don't like it, > we can drop 4-6. Let's keep all of them; they all looked reasonable. > Patch 7 is still authored by Michael, but has been massaged greatly by > me. I'll comment more directly on the changes there. > > [1/7]: gpg-interface: use child_process.args > [2/7]: verify_signed_buffer: drop pbuf variable > [3/7]: verify_signed_buffer: use tempfile object > [4/7]: run-command: add pipe_command helper > [5/7]: verify_signed_buffer: use pipe_command > [6/7]: sign_buffer: use pipe_command > [7/7]: gpg-interface: check gpg signature creation status Thanks. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html