Re: [PATCH] cat-file: fix mailmap application for different author and committer

From: Eric Sunshine <sunshine@sunshineco.com>
Date: Jun 11, 2025, 19:05 UTC
Message-ID: <CAPig+cTezW8XYjNo3L3Qy+f+aiCXppTj-Z=N0bBMG8mp9jJ8ZA@mail.gmail.com>
In-Reply-To: <xmqqy0tyi8aj.fsf@gitster.g>

On Wed, Jun 11, 2025 at 11:55 AM Junio C Hamano <gitster@pobox.com> wrote:

Show 9 quoted lines

> From: Siddharth Asthana <siddharthasthana31@gmail.com>
> > +     sed -n "/^author /s/\([^>]*>\).*/\1/p; /^committer /s/\([^>]*>\).*/\1/p" log >actual &&
>
> Perhaps just a  matter of taste, but
>
>         sed -n -e "/^author /s/>.*/>/p" -e "/^committer /s/>.*/>/p"
>
> may be easier to read and more portable (as some implementation of
> sed is picky about semicolon concatenated multiple commands).

For what it's worth, Git test scripts already contain a fair number of uses of semicolon-separated `sed` commands, and we haven't heard of any problems with them; not even from the very old and quite picky Solaris `sed` (or was it the ancient SunOS `sed`?).

The only case I can think of in which there was a semicolon-related problem (and perhaps what you're thinking of) was when a recent patch[*] neglected to insert a semicolon where it was expected. That particular case involved a missing semicolon before a closing brace:

    sed -n '/ version /{p;q}'

which should have been:

    sed -n '/ version /{p;q;}'

To summarize: Using semicolon-separated commands is safe and portable; I don't think there is any evidence that doing so would be problematic. Whether to use semicolon-separated commands or multiple `-e` arguments is subjective, and I don't believe the project has expressed a preference for one for or the other.

[*]: https://lore.kernel.org/git/CAPig+cR+ESNg4tV1G6jbKKeRKABD053qZcG0BoFuQ7aC+1tGYw@mail.gmail.com/