Make the pip hack script compatible with pip 18.1

[hartmantis]

It moved the from_line method in a backwards-incompatible way. This should resolve the remainder of #133.

My Python is pretty atrophied, but I tested this against the following recipe and it converged successfully.

apt_update 'default'

python_runtime '2'

python_virtualenv '/tmp/piplatest' do
  python '2'
end

python_virtualenv '/tmp/pip18' do
  python '2'
  pip_version '18.0'
end

python_virtualenv '/tmp/pip10' do
  python '2'
  pip_version '10.0.1'
end

python_virtualenv '/tmp/pip9' do
  python '2'
  pip_version '9.0.3'
end

[codecov-io]

Codecov Report

Merging #134 into master will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #134   +/-   ##
=======================================
  Coverage   55.53%   55.53%           
=======================================
  Files          21       21           
  Lines         632      632           
=======================================
  Hits          351      351           
  Misses        281      281

Impacted Files	Coverage Δ
lib/poise_python/resources/python_package.rb	37.6% <ø> (ø)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update b56d5d2...5fc90ef. Read the comment docs.

[whirsch-pubfactory]

I hacked something similar together, but your code is cleaner. This fixes the problem in my environment without any additional changes.

[powellchristoph]

Any idea when this can get released?

[coderanger]

@powellchristoph Unless you're planning to pay someone to do it, I wouldn't recommend ever asking that question in an open-source project.

[coderanger]

So this is currently jammed due to Travis changing their job timeouts from 2 hours to 1 for free tier projects. Unfortunately this means that the tests are unlikely to ever pass again, which makes releases quite dangerous. I'll maybe try to get things fixed in the future, but I would anticipate this being a permanent blocker given the amount of unpaid time I'm willing to put into this project going forward.

[keen99]

@coderanger while this is perhaps out of scope for this discussion - this means someone needs to feed $69/m to travis for poise-python, right? (vs rewriting the testing world) with the widespread impact of this, I'd think someone would be willing to step up and sponsor this... (or better - maybe it impacts someone at @travis-ci and they'll decide that their "always free for open source" definition needs some work...)

[coderanger]

If someone volunteers I wouldn't say no, but I don't have the time to do fundraising for my Chef projects anymore.

[willbarrett]

@coderanger in addition to the TravisCI (or presumably other CI solution) cost, how much would you be looking for to get this released?

[coderanger]

@willbarrett Given the impending release of Chef 15 will likely be the final nail in the coffin for my cookbooks, I think that ship has sailed. I would estimate something around $10-20k to get things back in a condition where they could be maintained through the Chef 15 release cycle. I suspect the sticker shock response of almost any company would be "what? lol no" but that's what several weeks of time costs.

[willbarrett]

OK, thanks! Makes sense.