changes to make app.current_request thread safe

[jtetrault]

• fixes race conditions that can occur when chalice is being run locally

Related Issue: #759

Wraps app.current_request access in a dict that is automatically keyed by the current thread's id, with each thread getting its own app.current_request.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[codecov-io]

Codecov Report

Merging #1358 into master will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #1358      +/-   ##
==========================================
+ Coverage   96.11%   96.12%   +0.01%     
==========================================
  Files          28       28              
  Lines        5324     5342      +18     
  Branches      682      682              
==========================================
+ Hits         5117     5135      +18     
  Misses        136      136              
  Partials       71       71              

Impacted Files	Coverage Δ
chalice/local.py	99.40% <100.00%> (+0.03%)	⬆️

---

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 e2eac21...33e7702. Read the comment docs.

[jamesls]

Thanks for the pull request! Reading over the discussion in #759, I'd prefer to make this change in the local mode code because it only affects local mode.

[jtetrault]

Thanks for the pull request! Reading over the discussion in #759, I'd prefer to make this change in the local mode code because it only affects local mode.

That makes sense. I made some changes so that only code running in local mode will be affected. Unfortunately I had to resort to a bit of a hack because the Chalice instance that gets passed around is instantiated in the user's app.py module, so there is no chance to subclass it properly.

[jamesls]

Thanks for the update. It looks like this failed the travis build so we'll need to get that resolved before we can merge.

[jtetrault]

I've resolved most of the mypy issues, but I think the one remaining is actually a known bug with mypy.
python/mypy#4125

[jamesls]

Thanks for updating. I'd still like to see the travis build passing before merging so I'm open to ideas on how to do that. What if instead of subclassing we just had LocalChalice take a reference to the original Chalice object and override the current_request property but then delegating everything else via getattr(). Maybe that could work?

[jtetrault]

Yeah, I had considered that approach as well. I can give that a try. I am not sure if mypy will be alright with it.

[jtetrault]

Looks like things are passing now. Thanks for the suggestion! I added a basic unittest as well to satisfy codecov.

[jamesls]

Thanks for updating! Looks good to me.

[jamesls]

Merged in 417e73e