#50199: gnuplot: opportunistically uses gtk2 when building with wxwidgets enabled -----------------------+--------------------- Reporter: devans@… | Owner: mojca@… Type: defect | Status: closed Priority: Normal | Milestone: Component: ports | Version: Resolution: fixed | Keywords: Port: gnuplot | -----------------------+--------------------- Changes (by devans@…): * status: new => closed * resolution: => fixed Comment: Replying to [comment:9 mojca@…]:
Replying to [comment:8 devans@…]:
This patch really is a shotgun approach because only one of configure or configure.in will actually be effective in any given case. Your patch for configure.in should work but you need to use autoreconf in this case to create a corresponding configure file. If you patch configure then it will work if you DON'T use autoreconf or it will get over-written.
I usually patch configure.in (or configure.ac in more modern ports) and autoreconf as the code is easier to understand without all the expanded macros in the way.
Did you actually test or are you speculating? I didn't patch `configure.in` (at least not intentionally as I don't want to run autoconf if there's no need to do so), I just tried to find the least intrusive patch that hopefully fixes the problem (I could have removed the whole chunk of code testing for GTK2, but this would require way more massive patching of `configure`).
I do think that a rev bump is appropriate as this port is not distributable and people current port may link or not with gtk2 depending on what was active build time. With the patch, you guarantee that they are not linking with gtk2. I noticed the problem because, my gnuplot was built with gtk2 but I inadvertantly deactivated it during testing of p5 -graphics-gnuplotif and got run time errors concerning the missing gtk2 libraries and their dependencies.
If you are happy with the patch, feel free to revbump. If you are not, I want further explanation :)
I tested it and verified that it builds without reference to gtk2 even if it is present. You are patching configure but the patchfile also includes changes to configure.in that have been commented out (so not functional). I was confused.
I think this issue has been beaten into submission at this point.
What does that mean?
I was being too colloquial, sorry. I mean the subject has been discussed thoroughly. At your invitation, revision incremented in r144230. Case closed. -- Ticket URL: <https://trac.macports.org/ticket/50199#comment:10> MacPorts <https://www.macports.org/> Ports system for OS X