Is there any hurdle in suggesting the patch to upstream (as a pull request, perhaps)?
From my PoV, it could just be merged.
A first step perhaps is just to open an issue with the current patch as attachement
and asking for any opinion on why someone would not like those changes. Perhaps
upstream wants to make the changes their own way.
Important points are:
Support Boost::pythonXY component as opposed to Boost::python or Boost
python3. This is the present/future.
Use current (less deprecated) cmake ways to detect python.
Less redundancy in Python checking in CMakeLists at deeper level.
Add support for BLAS_LIBS and LAPACK_LIBS, as this is customary in HPC
installations to choose the (optimized) BLAS implementation in an environment
that offers multiple options.
If they agree that these are worthy goals, they could just merge the patch file, or
ask (you?) to fork their repo and a prepare a pull request, as it is customary in
the git(hub, lab) world nowadays.
Hi Thomas, the new version of DP3 brakes our installation script again,
I will put together an updated patch. Did you manage to start
propagating your fixes to the DP3 repo?
Hi Thomas, the new version of DP3 brakes our installation script again,
I will put together an updated patch. Did you manage to start
propagating your fixes to the DP3 repo?
I wanted to today … you got your update ready yet?
Well, this just looks like the DP3 patch has been fully merged upstream, see https://github.com/lofar-astron/DP3/pull/238 . So we can drop the patch. I don't see any hunks left, presuming your log is complete. You might try preparing a commit for that yourself.
Henrik, you should be able to push to the repo now. You could remove the patch (or just make it empty). The EveryBeam thing … it's only in the dev branch, right? Maybe you can work on supporting that one, too.
It is only on the dev branch, yes. But afaikt this is the only branch where the your pull request was merged.
So for the current master we would still need the patch.