need to move vsc.install to its own repo

[stdweird]

any objections? the repo could also hold wrppers for running the jenkins unittests, whihc are unreviewed and look funny

[stdweird]

@JensTimmerman @boegel toughts?

[boegel]

I don't see the big advantage of fleshing it out into a separate repo/pkg, introducing our own dependency in vsc-base just so we can install/test it is weird?

Adding scripts for the Jenkins unit tests here still makes sense though.

[stdweird]

here's a good reason: stdweird@7cd32b5 (and some of it's implications you're probably not aware of)
the jenkins scripts should be part of vsc-base, should be the same script for anything that uses vsc.install (it's only a build dependency afterall)

[boegel]

OK, I guess that is a good reason. And indeed, since it's only a build dependency for vsc-base, it's OK-ish (and will lead to a cleaner vsc-base repo/install).

So: +1 on fleshing out vsc-install.

This does imply having both vsc-install and vsc-base on PyPI (too) btw, since this affects the EB bootstrap procedure too (but things should work just fine, no changes should be needed there).

[boegel]

@stdweird: What other implications are you referring to btw? Is there more going on that just making sure that import vsc.install works with that hack (which will no longer be needed if vsc-install is proper (build-only) dep)?

[JensTimmerman]

how does this solve anything? Aren't you just moving the problem?
vsc.install will still depend on vsc.install to install, and as such the hack will still be needed in it's setup.py file?

[stdweird]

yes, but only in that file and it will be a build_requires for vsc-base (i mean it's ok to have the hack for vsc.install, but is a but to coarse for vsc-base)

[JensTimmerman]

surely it will also be install_requires ? and as such will still be pulled in during installation?

[stdweird]

only build_reqiures, why would you need vsc.install to run anything?

[JensTimmerman]

installation requires it, you need it to run python setup.py install . in vsc-base