Created attachment 188243 [details] math/py-pandas: Update to 0.21.0 Changelog ========= * Update to 0.21.0 QA == * portlint: OK * poudriere: OK -- testport on 10.4-RELEASE-p3 amd64 w/py27 (default), py34, py35, py36 * pytest: OK -- all OPTIONS enabled if possible (ALCHEMY11) on 10.4-STABLE (r325957) amd64 w/py27 (default), py34, py35, py36 Notes ===== testport performed with default options. Residual pytest errors are attributable to defects in a dependent port (devel/py-tables).
A commit references this bug: Author: tcberner Date: Sat Nov 25 18:39:42 UTC 2017 New revision: 454890 URL: https://svnweb.freebsd.org/changeset/ports/454890 Log: Update math/py-pandas to 0.21.0 PR: 223849 Submitted by: John W. O'Brien <john@saltant.com> (maintainer) Changes: head/math/py-pandas/Makefile head/math/py-pandas/distinfo head/math/py-pandas/files/patch-pandas_tests_io_test__html.py head/math/py-pandas/files/patch-pandas_tests_io_test__sql.py
Committed. Thanks.
(In reply to Tobias C. Berner from comment #2) Thank you for taking this bug, Tobias. Why did you change PORTVERSION to DISTVERSION?
Hi John It's the preferred way to set the version. mfg Tobias
Preferred by whom? Your claim is unsupported. $ find /usr/ports/ -depth 3 -name Makefile \ | xargs egrep -o "(PORT|DIST)VERSION" -h \ | sort \ | uniq -c 4435 DISTVERSION 34367 PORTVERSION Please revert this gratuitous change and in the future avoid committing changes that are neither approved by the maintainer nor directed by portmgr. Thank you.
Hi John Preferred according to the porters handbook [1]. The scale of these two numbers will certainly change in the future. Sorry, I did not mean to step on your toes as the maintainer :) If you still want a revert of this change, I can do it. mfg Tobias [1] https://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/makefile-naming.html#makefile-versions
That's a relatively recent change [0] and I can find no traffic at all on the ports list or ports-announce describing it, let alone asking for existing ports to be changed. What's the benefit of changing? What problem does it solve? Is it a problem with this port that I'm not aware of? Can a larger problem only be solved if this port changes? It seems like it creates needless churn, which frustrates me. Furthermore, without sufficient reason, nor the coordinated effort required to convert tens of thousands of ports to the newly preferred style, it also seems likely to fizzle out and leave the tree in a more inconsistent state than when we started, which frustrates me more. I can appreciate your disinclination to revert the change. Nevertheless, it leaves a bad taste in my mouth. [0] https://svnweb.freebsd.org/doc/head/en_US.ISO8859-1/books/porters-handbook/makefiles/chapter.xml?r1=50627&r2=50628&