[Midnightbsd-cvs] [MidnightBSD/mports] 1e92d7: ai/codex: de-duplicate distinfo

[Lucas Holt]

  Branch: refs/heads/ai/codex
  Home:   https://github.com/MidnightBSD/mports
  Commit: 1e92d7e83aa3cbc8d81da682eef07f68ce3fffff
      https://github.com/MidnightBSD/mports/commit/1e92d7e83aa3cbc8d81da682eef07f68ce3fffff
  Author: Lucas Holt <luke at foolishgames.com>
  Date:   2026-08-21 (Fri, 21 Aug 2026)

  Changed paths:
    M ai/codex/distinfo

  Log Message:
  -----------
  ai/codex: de-duplicate distinfo

distinfo contained the entire distfile list twice: 2424 SHA256 entries for
1212 distinct files, the second block byte-identical to the first. It was
introduced in 017ad5ddb3 (0.147); the file was clean at e292b391b8 (0.146.0),
so makesum appended rather than replaced, or ran twice.

Duplicated entries break fetching, because DISTINFO_DATA (Mk/bsd.mport.mk)
prints every matching line:

    CKSIZE=$(distinfo_data SIZE "${full_file}")   -> "17438 17438"

Mk/scripts/do-fetch.sh then interpolates that unquoted into "-S ${CKSIZE}",
so fetch(1) receives a stray extra argument. Because the crate paths contain
a '/', do-fetch.sh also passes "-o <path>", and fetch requires -o to name a
directory once it sees more than one URL:

    fetch: rust/crates/Inflector-0.11.4.crate is not a directory

Every mirror then fails identically, which is how this surfaced in Magus
run 647 as a fetch failure rather than a malformed distinfo.

Keeping only the first block leaves the file byte-identical to a correct
makesum run. Verified the 1212 remaining entries match DISTFILES exactly
(none missing, none extra) and that "bmake checksum" passes for all of them.

AI-Assisted-by: Claude Opus 5 <noreply at anthropic.com>
Signed-off-by: Lucas Holt <luke at foolishgames.com>



To unsubscribe from these emails, change your notification settings at https://github.com/MidnightBSD/mports/settings/notifications