Created attachment 237329 [details] Patch for py-metadata-cleaner Use USE_GITLAB to follow Porters Handbook more closely and be consistent with ports tree Compile tested on FreeBSD 13.1-STABLE (amd64)
I've also submitted a small patch for meson so we can drop BINARY_ALIAS and use shebangfix once it's merged. https://gitlab.com/rmnvgr/metadata-cleaner/-/merge_requests/168
Comment on attachment 237329 [details] Patch for py-metadata-cleaner Unfortunately I cannot approve of this change as the source tree snapshot is not the release tarball. So as per policy, it should not be used for the distfile. Please see bug #264893 for where I performed the change you suggested in the exact opposite direction. However, thank you also for submitting an update of the port to 2.2.5. I give you approval to commit just the update to that version and the corresponding changes without the change to which distfile is used.
Gitlab can be a bit confusing, "Assets" are not release tarballs and generated on the fly. Compare that to static tarballs attached to the release such as https://gitlab.com/sane-project/backends/-/releases/1.1.1 (Other) or https://gitlab.com/LibreGames/jumpnbump/-/releases (Downloads). Since upstream doesn't supply one USE_GITLAB should be uses for consistency with the rest of the tree (we have ~50 ports that violates this policy).
(In reply to Daniel Engberg from comment #3) Is the tarball produced with USE_GITLAB bit for bit identical to the one obtained from the assets list? It does not seem like it is from you setting WRKSRC explicitly. As it is not the same tarball, I cannot approve of the change.
(In reply to Daniel Engberg from comment #3) > (we have ~50 ports that violates this policy). Please point out the policy stating under which circumstances USE_GITLAB must be used. As I read it, USE_GITLAB is only provided as a convenience, not as a requirement.
Comment on attachment 237329 [details] Patch for py-metadata-cleaner Maintainer approval denied: I do not approve of the change in distfile source. The other changes are okay and can be committed.
We're aligning the tree for consistency, there's work being done in improving USE_GITLAB to use release tags (like USE_GITHUB) however we want (and most likely will) prepare the tree for future changes. Since Gitlab instances generates tarballs differently depending on filename and url there will be differences in checksums and directory naming.
(In reply to Daniel Engberg from comment #7) If this is in preparation of a future improvement, then please proceed.
Comment on attachment 237329 [details] Patch for py-metadata-cleaner Approval granted.
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=d1797aa3bc7c30b16f936f068c5380965190bf69 commit d1797aa3bc7c30b16f936f068c5380965190bf69 Author: Daniel Engberg <diizzy@FreeBSD.org> AuthorDate: 2022-11-06 10:21:04 +0000 Commit: Daniel Engberg <diizzy@FreeBSD.org> CommitDate: 2022-11-06 10:46:53 +0000 multimedia/py-metadata-cleaner: Update to 2.2.5 Changelog: https://gitlab.com/rmnvgr/metadata-cleaner/-/blob/0aad256555bf8fea0015532d204869987e3cfa8c/CHANGELOG.md PR: 267076 Reviewed by: Robert Clausecker <fuz@fuz.su> (maintainer) multimedia/py-metadata-cleaner/Makefile | 11 +++++++---- multimedia/py-metadata-cleaner/distinfo | 6 +++--- multimedia/py-metadata-cleaner/pkg-plist | 32 ++++++++++++++++++++++++++++++++ 3 files changed, 42 insertions(+), 7 deletions(-)
Committed (added back DISIVERSIONPREFIX and updated disthash), thanks