Bug 209630
| Summary: | [maintainer-update] science/py-h5py: update to 2.6.0 | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Johannes Jost Meixner <xmj> | ||||
| Component: | Individual Port(s) | Assignee: | Kurt Jaeger <pi> | ||||
| Status: | Closed FIXED | ||||||
| Severity: | Affects Only Me | CC: | pi, xmj | ||||
| Priority: | --- | Flags: | pi:
maintainer-feedback+
|
||||
| Version: | Latest | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Bug Depends on: | |||||||
| Bug Blocks: | 209746 | ||||||
| Attachments: |
|
||||||
testbuilds in poudriere are OK, but it has a strange failure mode: http://people.freebsd.org/~pi/logs/h5py-build-issues.txt See http://people.freebsd.org/~pi/logs/h5py-build-issues-2.txt for a second syndrome: The files of the port itself are mode 644 before the build, 444 after the build. What's going on ? (In reply to Kurt Jaeger from comment #1) I can't reproduce this error with recent portstree in a poudriere jail set up with poudriere testport -I -n -j 11amd64 -p svnports -o science/py-h5py. I'm wondering, do you have OPTIONS_UNSET= NLS somewhere in a make.conf? It's not in poudriere, it's on a populated host (fc.opsec.eu). /etc/make.conf: WITH_PKGNG=devel DEVELOPER=yes DEFAULT_VERSIONS= perl5=5.22 python=2.7 python3=3.4 ruby=2.3 pgsql=9.5 php=7.0 'populated host' means:
pkg info | wc -l
1941
(In reply to Kurt Jaeger from comment #2) I can't reproduce this error with poudriere either. I made a test with make -d all In line 691 of http://people.freebsd.org/~pi/logs/h5py-build-issues-3.txt this is executed: (cd /home/pi/myp/science/py-h5py/work/h5py-2.6.0/docs) && /bin/sh -c '(/usr/bin/ find -d $0 $2 | /usr/bin/cpio -dumpl $1 >/dev/null 2>&1) && /usr/bin/find -d $ 0 $2 -type d -exec chmod 755 $1/{} \; && /usr/bin/find -d $0 $2 -type f -exec c hmod 444 $1/{} \;' -- . /home/pi/myp/science/py-h5py/work/stage/usr/local/share/ doc/h5py There are many (>700) lines of failing chmod following. (In reply to Kurt Jaeger from comment #7) So, the line in question within the Makefile is this: @(cd ${WRKSRC}/docs) && ${COPYTREE_SHARE} . ${STAGEDIR}${DOCSDIR} Which, on my system, works because it doesn't try to create the DOCSDIR twice chmod: /home/pi/myp/science/py-h5py/work/stage/usr/local/share/doc/h5py/./work/stage/usr/local/share/doc/h5py/work/h5py-2.6.0/docs/high: No such file or directory it looks like your system is trying to copy everything into ${DOCSDIR}${STAGEDIR}${DOCSDIR} which it shouldn't do. To verify, can you run `make -V STAGEDIR -V DOCSDIR`? It should look like this - which is why your poudriere install works: root@11amd64-svnports:/usr/ports/science/py-h5py # make -V STAGEDIR -V DOCSDIR /wrkdirs/usr/ports/science/py-h5py/work/stage /usr/local/share/doc/h5py How do we move this forward? (In reply to comment #8) make -V STAGEDIR -V DOCSDIR says: /home/pi/myp/science/py-h5py/work/stage /usr/local/share/doc/h5py (In reply to Kurt Jaeger from comment #10) Before: @(cd ${WRKSRC}/docs) && ${COPYTREE_SHARE} . ${STAGEDIR}${DOCSDIR} After: ${COPYTREE_SHARE} ${WRKSRC}/docs ${STAGEDIR}${DOCSDIR} works like a charm. Committed, thanks. I did not find any changelog -- any ideas where it's hidden ? A commit references this bug: Author: pi Date: Wed May 25 16:32:24 UTC 2016 New revision: 415847 URL: https://svnweb.freebsd.org/changeset/ports/415847 Log: science/py-h5py: 2.5.0 -> 2.6.0 PR: 209630 Submitted by: Johannes Jost Meixner <xmj@chaot.net> (maintainer) Changes: head/science/py-h5py/Makefile head/science/py-h5py/distinfo |
Created attachment 170471 [details] update py-h5py to 2.6.0 Patch attached updates science/py-h5py to 2.6.0. Poudriere & portlint-tested.