tech-pkg archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: compilation of x11/qt5-qtwebkit on pkgsrc-2020Q4 branch



Aymeric Vincent <aymericvincent%free.fr@localhost> writes:

> I started following the pkgsrc-2020Q4 branch and noticed that I need to
> put back the line
>
> PYTHON_VERSIONS_ACCEPTED=       27
>
> in the Makefile of x11/qt5-qtwebkit to make it compile[, possibly because
> PYTHON_VERSION_DEFAULT is set to 39 in my /etc/mk.conf].

FWIW, I have this built on a netbsd-9/adm64 that is just about at
2020Q4.  The pkgsrc checkout actually is a tiny bit newer, but I don't
think anything important has changed.  It is installed as a dependency
of merkaator and qgis, for me, and that all built fine as I kept doing
pkg_rr leading up to the branch.

It has also been building fine in the bulk builds of pkgsrc-current on
NetBSD 9.  A fail breaks about 75 packages, so it would be really
obvious vs a package breaking 10  being a big deal, and I don't remember
seeing it.

So it seems to build fine with 38, the default.

BTW, your bug report did not specify the NetBSD version and CPU type.

So this leads me to ask

  What do the upstream build instructions say about required python
  versions?  Is there a bug open on the upstream issue tracker?

  Did you try to debug this by trying to use the defaults, i.e. backing
  out your PYTHON_VERSION_DEFAULT setting?  If so, what happened?  (If
  not, please try it.)

  Assuming the previous step builds, what happens if you set
  PYTHON_VERSIONS_INCOMPATIBLE to 39?

> Otherwise the "Source/JavaScriptCore/generate-bytecode-files" python
> script fails when trying to pass the "encoding=" parameter to
> json.load(). This seems to be a python2-ism.
>
> I'm not sure if it's the correct fix, so I just mention this workaround
> FWIW. [Also, I'm not subscribed, cc me if you want to follow up.]

Attachment: signature.asc
Description: PGP signature



Home | Main Index | Thread Index | Old Index