On 2018/12/01 17:57, Charles A Daniels wrote: > Thank you for your response! It's been very helpful. I've attached a > tarball incorporating your feedback. It seems like the test suite > is currently failing, so I am working on reaching out to the upstream > author to see what can be done. It looks like his test program > hard-codes expected values, and this is running into some floating > point rounding issues. The port appears to be working well enough to > be tested at this point, although obviously I think we want the test > suite passing before it gets committed. > > > > GH_ACCOUNT = vapier GH_PROJECT = pcalc GH_TAGNAME = v4 > > Either DISTNAME or GH_* is used, never both (is Makefile.template not > > clear enough about that?). > > I read through the relevant sections of the template and it was not clear to > me that the two were mutually exclusive. I would suggest adding the string > literal "Either DISTNAME or GH_* is used, never both" > to the template to that end.
That's not true. GH_COMMIT always needs DISTNAME. GH_TAGNAME may need DISTNAME if the autogenerated DISTNAME is bogus. Generally if GH_TAGNAME is of the format "1.0" or "v1.0" then the default is OK but if it's e.g. "zopfli-1.0.2" then the autogenerated one ends up as "zopfli-zopfli-1.0.2" so needs resetting. Easiest way to check is to not set DISTNAME in the Makefile and do "make show=DISTNAME". If that results in a sane name then keep it like that. Otherwise set DISTNAME as appropriate. > > Yes, `# GPLv2' is fine, but you can double check if it's GPLv2+ or > > GPLv3(+) really. > > As I noted in my update Makefile, I am going to reach out to the author to > clarify his intent. That's updated upstream now (in README.md committed today), "This project is licensed under GPLv2 (or later)", so please write that as "GPLv2+".
