signed-off-by: skip rule when PR-URL trailer is present not only Backport-PR-URL?

[thisalihassan]

The signed-off-by rule currently skips backports identified by a Backport-PR-URL: trailer. This works for already-landed backport commits, but fails for backport PRs at review time because the commit-queue only adds Backport-PR-URL: at land time. even though the contributing doc explicitly exempts backports.

Rule 1 says:
https://github.com/nodejs/node/blob/e15f90559bc62b1e9a8296f1fc8654f5be261baa/doc/contributing/pull-requests.md?plain=1#L207-L208

Rule 2 says:
https://github.com/nodejs/node/blob/e15f90559bc62b1e9a8296f1fc8654f5be261baa/doc/contributing/backporting-to-release-lines.md?plain=1#L77

The lint (in core-validate-commit) says: if no Signed-off-by: AND no Backport-PR-URL: (which is added by the commit-queue), fail.

Suggestion

Treat PR-URL: trailer as the same kind of skip signal as Backport-PR-URL:. Both trailers are added by the commit-queue at land time they're never contributor-authored. Their presence on a commit being validated is a strong heuristic that the commit was either cherry-picked from a previously-landed PR (the dominant case, all backports), or is itself a previously-landed commit having its message re-validated.

[thisalihassan]

cc: @aduh95 @jasnell @mcollina @richardlau

[aduh95]

I don't think PR-URL is the right proxy, it gives incentives for folks to add a dummy PR-URL trailer to shut off the linter. tbh I would be more inclinded to disable the commit linter on PR targeting staging branches.
Buut arguably the whole backport process could benefit from a complete overhaul, it's quite hard for humans to thoroughly review a a backport PR, and the linters are not well equiped to help with that