Created attachment 161108 [details]
Add QT4AGGBACKEND option, use find to strip shared libraries
Hoping that it's not too much of a bother so soon after the previous report, this one will hopefully be easier now the Makefile was cleaned.
I'd like for the Qt4Agg backend to be optionally available (it's the recommended one by upstream after all).
I only added py-qt4-ui as dependency, which ~1 month of (obviously not exhaustive) usage showed no problem.
Qt5Agg backend would be desirable, but as there is (AFAIK) no py-qt5 port, that will have to wait.
The patch also convert library stripping to a find command, because build could fail with non-existent libraries with an explicit and complete list.
As plist-ing is automatic, I'm not sure that make check-plist has a sense, but for completeness' sake:
===> Parsing plist
===> Checking for items in STAGEDIR missing from pkg-plist
===> Checking for items in pkg-plist which are not in STAGEDIR
===> No pkg-plist issues found (check-plist)
No runtime issues either after pkg autoremove'ing build dependencies.
Thank you for your time
I approve the patch.
Created attachment 161906 [details]
Add QT4AGGBACKEND option, revised
Updating the patch since another and different fix was applied to resolve shared lib stripping.
Looks good---I approve the patch.
Hmmm, maybe I missed something, but I did not see any commit to the port?
A commit references this bug:
Date: Thu Oct 29 13:47:20 UTC 2015
New revision: 400457
Add Qt4Agg backend option.
Simplify post-install stripping.
Submitted by: Matthieu Volat
Approved by: maintainer
Sponsored by: Absolight
(In reply to Matthieu Volat from comment #4)
> Hmmm, maybe I missed something, but I did not see any commit to the port?
Some of us use Git on top of Subversion. I close the PR when I commit the change in my git repo, and issue a git svn dcommit at the end of the day, so it can be a few hours, but it's there.
Ho, ok, I understand. Thank you very much!