Created attachment 210573 [details] Patch Update to 2.2.9 Use xz archive instead of bz2 Disable examples as they're not used Make "make test" work and only build tests if enabled Feel free to "backport" Makefile changes to 2.2.8 if you still think it's not worth updating to 2.2.9 as mentioned in PR#242215.
Created attachment 210574 [details] Poudriere log
Build info is available at https://gitlab.com/swills/freebsd-ports/pipelines/108228612
A note from maintainer. Please consider PR #242215 : https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=242215
How about staying on the current version, would that be acceptable?
(In reply to daniel.engberg.lists from comment #4) If this is a question for me, then could you kindly explain what this question means? What do you call "current version"? The latest commit? Have you read discussion from PR#242215? If you want this port to be updated to any version published by the upstream, and without thinking if it is necessary at all for this particular opsys, they you do it without me. Repeat, I do not approve this patch.
Created attachment 210639 [details] Patch Same as previous except for the version bump
Created attachment 210640 [details] Poudriere log
I apologize for not being clear about it, by current version I was referring to the current version in the ports tree (2.2.8). Also tested on 13-CURRENT r356392 (AMD64) Best regards, Daniel
Friendly ping, Sergei is the last patch acceptable to you? Best regards, Daniel
You are asking this for the 3-rd time, are you? Repeat from PR#242215: because of exprun, this port should not be touched unless absolutely necessary. You patch offers changes which are trivial, not needed and not justified, but I have noted your ideas. At the moment working on future version with multiple bug-fixes and multiple new features. In due time I will reconsider your ideas, and maybe reject some of them again. Let me kindly ask that somebody closes this PR as "rejected" or "not needed".
So I'm curious, what's so expensive about an exp-run? And if we're doing one anyway, why not bump the version? this seems like an overly-rude response to a reasonable request.
Also, as a side note, an exp-run is required when the version changes, not when doing cosmetic changes to the Makefile.
(In reply to Warner Losh from comment #11) 1. Let me apologize before any one, for whom my response seems to be rude. 2. About expenses of exp-run. Last time exp-run for this port spotted some problems with a port, which I never had heard about. Then nothing happend for a long time. Then I had to open another PR about this problematic port, invent a solution and suggest the solution to the maintainer of the problematic port. And only after that my initial patch was committed. Since then I am afraid of exp-runs.
Created attachment 218505 [details] Patch for expat2 v2 Additional changes: Update to 2.2.10 Addtional fixes to "make test" target Compile tested on FreeBSD 13.0-CURRENT #0 r364979 (amd64) (make + make check-plist + make test) Poudriere testport OK 12.1-RELEASE (amd64)
(In reply to daniel.engberg.lists from comment #14) Let me step down from maintainership of this port. Notes from former maintainer: 1) Target "post-install" should be changed into "post-install-DOCS-on". Or option DOCS has no sence. 2) Normally testing (make test) should be workable before "make install". But with the proposed solution it works only after "make install". To see this effect run poudriere in interactive mode with option TEST enabled.
Created attachment 218883 [details] Patch for expat v3 Return port to pool by request of maintainer Fix DOCS and add MANPAGES menu options
(In reply to daniel.engberg.lists from comment #16) Moin moin I don't quite see the point for the manapge option -- the manpage does not require any build/run dependencies, and it is just a small file -- so I don't see any benefit in not always installing it. mfg Tobias
Moin moin I would like to ask for an exp-run of the attached patch -- the maintainer will either be Daniel, desktop@ or myself if no other contender shows up. It won't be committed with ports@. mfg Tobias
(In reply to Tobias C. Berner from comment #17) While I don't have any strong opinion about it I don't see a reason why we shouldn't provide it in the first place / remove it.
I think we should reserve the MANPAGES option for instances where (a) nontrivial dependencies are needed to build the manpage(s), or (b) there are a nontrivial number of manpages. My fear is that if we use the MANPAGES option to guard trivial cases, we dilute both the meaning and the significance of the option. That said, that argument does not necessarily bear out in the rest of the tree. This is said without any hats; I'll leave the decision up to whomever will end up maintaining this thing. I urge consideration of the importance of consistency in end-user mechanisms, but this is definitely a bikeshed question anyway.
Exp-run looks fine
Thanks Antoine! As far as maintainer goes I think desktop@ would be a good candidate.
A commit references this bug: Author: tcberner Date: Mon Nov 16 18:15:16 UTC 2020 New revision: 555503 URL: https://svnweb.freebsd.org/changeset/ports/555503 Log: textproc/expat2: Update to 2.2.10 - give maintainership to desktop@ - add test target Changelog: https://github.com/libexpat/libexpat/blob/R_2_2_10/expat/Changes PR: 243228 Submitted by: daniel.engberg.lists@pyret.net Exp-run by: antoine Approved by: Sergei Vyshenski <svysh.fbsd@gmail.com> (previous maintainer) Changes: head/textproc/expat2/Makefile head/textproc/expat2/distinfo head/textproc/expat2/pkg-plist
Sergei Vyshenski, thanks for your maintainer ship so far :)