gpgme-1
- Path: arch/maintained/gpgme-1
- Build failure during the 2026-07-07 “build all packages for x86_64”
private-repo release pass (see
repo_release_logs/maintained_gpgme-1.log).
- Purely transient:
fatal: unable to access 'https://dev.gnupg.org/source/gpgme.git/': The requested URL returned error: 429
(rate-limited by upstream). Confirms the new archlinuxaba-x86_64-build
chroot is active (Building in chroot for [archlinuxaba] (x86_64),
archlinuxaba downloading...).
- Not a real defect – just needs a plain retry later once the rate limit clears.
Not retried yet as part of this pass.
- 2026-07-28: recurred with the exact same 429 in the routine
repo_release_changed.sh batch, three weeks later – this is not a one-off,
dev.gnupg.org seems to rate-limit this build host persistently rather than
transiently. Blocks qgpgme1-qt5 too (see its own memory file) since that
package depends on this one being published first. Still not retried
standalone – worth trying a repo_release.sh maintained/gpgme-1 in isolation
sometime outside a full batch run, in case the 429 is triggered by the batch’s
overall request volume rather than being unconditional.
- Retried standalone right after the batch, same session: still 429 on both i686
and i486 (default
repo_release.sh arch set, no REPO_BUILD_ONLY_ARCH
override this time) – so it’s not batch request volume, dev.gnupg.org is
just persistently blocking this host’s IP/UA for
dev.gnupg.org/source/gpgme.git. Genuinely blocked, not a quick win – stop
retrying same-session, revisit next cycle or consider whether this source URL
has a mirror.
- 2026-07-29: user fixed manually – swapped the VCS source from
git+https://dev.gnupg.org/source/gpgme.git to the GitHub mirror
git+https://github.com/gpg/gpgme.git#tag=gpgme-1.24.3?signed, bumped
pkgrel 4 -> 5 (packaging-only change, source unchanged), and committed it
in the submodule (“using github, gpg git is constantly down”). .SRCINFO
already regenerated to match. If this recurs for other gnupg.org-sourced
packages (e.g. check qgpgme1-qt5, which depends on this one), the GitHub
mirror is the fix, not another retry.
- 2026-07-29 rebuild via
repo_release.sh maintained/gpgme-1: x86_64
built/signed/published fine (github source fix confirmed working, no more
429s). pentium4/i686/i486 all failed identically in doc/
(make[3]: *** [Makefile:502: gpgme.info] Error 1), root cause
reader_element/TreeElementXS.c: loadable library and perl binaries are mismatched (got first handshake key ..., needed ...)
– a Perl XS ABI mismatch inside those three extra-<arch>-build chroots
(texinfo’s makeinfo vs the chroot’s perl), not a gpgme PKGBUILD problem.
Looks like chroot package drift (partial upgrade) rather than anything
source-specific – would need pacman -Syu (or a rebuild) inside each of those
chroots to confirm/fix, not attempted this session since it’s an environment
issue outside this package’s scope and no chroot-update tooling exists yet in
scripts/. Expect this to recur on any other package whose doc/ build
invokes makeinfo in these three chroots – worth checking if it’s already hit
others before assuming it’s new.