Created attachment 200647 [details] patch This port fails to build on 12 since about August 2018. It turns out that a dangling #define of ksize_t in the port sources is the culprit. The attached simple patch will fix it, please apply and bump PORTREVISION. (Note: portlint utters a few harmless warnings but no errors.)
Maintainer informed via mail
Thank you for the patch Helge. Could you please: - Update the patch to include the PORTREVISION bump - Update ether your Bugzilla email address to match the ports current MAINTAINER line, or update the MAINTAINER line to match your current Bugzilla email address - Confirm that this change passes QA (poudriere in particular) - Set the maintainer-approval flag to + on attachments for ports you are maintainer of. Attachment -> Details -> maintainer-approval [+]
Created attachment 200653 [details] patch +revision
(In reply to Kubilay Kocak from comment #2) - done - will not do - I am intentionally using different email addresses - no clue how to achieve that :-( - done
(In reply to Helge Oldach from comment #4) Without the emails matching, we: - can't verify maintainer'ship or patch provenance - can't reliably obtain approvals from the maintainer for patches provided by people other than the maintainer - can't have bugzilla set system/issue fields appropriately or automatically We need the emails to match, for whichever email is preferred by the maintainer. Details and instructions for testing ports can be found here: https://www.freebsd.org/doc/en/books/porters-handbook/testing.html Note also the distinction between maintainer-feedback flag (at the issue level), and maintainer-approval flag at the attachment level. One is to acknowledge feedback, the other is to approve (or not) patches (attachments). See: Attachment -> Details -> maintainer-approval flag
(In reply to Kubilay Kocak from comment #5) Again, I won't change email addresses. Note I am the only user at oldach.net so your administrative concerns are certainly met. I won't install poudriere as my machine is not capable enough and installation of it is obviously overkill for this single line patch. I already noted about portlint results. Flag semantics groked.
(In reply to Helge Oldach from comment #6) I understand, though unfortunately there's no way to objectively or reliably verify that. Part of maintainership involves following the processes of the project one is contributing to. Certainly not for the sake of process, it reduces ambiguity, manual handling, saves us time, a resource that is limited, and ultimately benefits you, getting your changes addressed and into the hands of users sooner. This is a question of collaboration, not enforcement. I hope you'll reconsider.
(In reply to Helge Oldach from comment #6) For details on why change size is orthogonal to QA requirements, see: https://lists.freebsd.org/pipermail/freebsd-ports/2018-December/115041.html
(In reply to Kubilay Kocak from comment #7) OK, so in case you still need approval from the port maintainer and don't trust my claim that he already approved the patch, I suggest that you obtain approval from him by email.
Created attachment 200676 [details] patch +revision corrected patch: the dangling ksize_t was introduced by ports r335979, hence only consider for the related __FreeBSD_version onwards.
(In reply to Helge Oldach from comment #10) base r335979 of course
testbuilds are fine.
A commit references this bug: Author: pi Date: Sat Jan 12 19:34:15 UTC 2019 New revision: 490075 URL: https://svnweb.freebsd.org/changeset/ports/490075 Log: www/sitecopy: unbreak on 12 and later PR: 234524 MFH: 2019Q1 Submitted by: Helge Oldach <freebsd@oldach.net> (maintainer) Changes: head/www/sitecopy/Makefile
A commit references this bug: Author: pi Date: Sat Jan 12 19:38:55 UTC 2019 New revision: 490076 URL: https://svnweb.freebsd.org/changeset/ports/490076 Log: MFH: r490075 www/sitecopy: unbreak on 12 and later PR: 234524 Submitted by: Helge Oldach <freebsd@oldach.net> (maintainer) Approved by: portmgr (unbreak blanket) Changes: _U branches/2019Q1/ branches/2019Q1/www/sitecopy/Makefile
Committed, thanks!