refactor: Remove bluebird from lib/plugins/aws/customResources

[juanjoDiaz]

Addresses: #7171

[codecov]

Codecov Report

Merging #9069 (ac9b0d7) into master (ccba70a) will increase coverage by 0.01%.
The diff coverage is 94.83%.

@@            Coverage Diff             @@
##           master    #9069      +/-   ##
==========================================
+ Coverage   87.15%   87.16%   +0.01%     
==========================================
  Files         288      288              
  Lines       11026    11067      +41     
==========================================
+ Hits         9610     9647      +37     
- Misses       1416     1420       +4     

Impacted Files	Coverage Δ
lib/utils/open.js	13.58% <66.66%> (ø)
lib/plugins/invoke.js	90.90% <75.00%> (-3.54%)	⬇️
lib/utils/createFromTemplate.js	83.33% <77.27%> (-11.67%)	⬇️
lib/plugins/config.js	81.33% <80.00%> (+0.77%)	⬆️
lib/plugins/aws/utils/credentials.js	91.93% <83.87%> (-2.51%)	⬇️
lib/plugins/create/create.js	90.69% <86.36%> (+0.69%)	⬆️
lib/plugins/aws/provider.js	95.10% <86.48%> (-0.04%)	⬇️
lib/plugins/package/lib/zipService.js	96.29% <93.65%> (-0.93%)	⬇️
lib/utils/yamlAstParser.js	97.32% <94.82%> (-0.87%)	⬇️
lib/plugins/interactiveCli/tabCompletion.js	97.50% <96.29%> (+0.20%)	⬆️
... and 20 more

---

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 ccba70a...d2f4c51. Read the comment docs.

[pgrzesik]

Hello @juanjoDiaz, with a heavy heart, due to a few regressions introduced with recent Bluebird refactorings(we've noticed way slower packaging, broken print command, and just today the issue with EMFILE error #9073 which suggests that we might be opening way more files than previously), we've decided to revert a few of the recent PRs with such refactorings and put on pause bigger changes related to Bluebird migration. Unfortunately, it's very easy to miss things during the review process of such refactors, some of the changes are also not caught by our test suite and we want to be more careful moving forward. We'd like to follow the process outlined in #7171 to limit the potential impacts and unexpected regressions for Framework users. We really appreciate all your efforts in pushing that refactoring forward, but we have to slow down a bit and introduce the changes in smaller chunks and we'd be thrilled to accept your contributions.

[juanjoDiaz]

Too bad.
But totally understandable.

I guess that a lesson learned from all this is that test coverage is high but not totally effective.

[pgrzesik]

Sorry, I've clicked "close" by mistake

I guess that a lesson learned from all this is that test coverage is high but not totally effective.

Unfortunately, some of the things are really hard to test, such as detecting the number of open files when packaging bigger service. But we also managed to improve our coverage a bit thanks to these findings e.g. by adding a sanity test for print command` that previous test cases didn't catch. Moving forward, we will dedicate more time to thoughful and detailed reviews to avoid such situations, but in order to do that, we want to keep the scope of the changes as small as possible to reduce the possibility of a bug sneaking in. Thanks for understanding and for all your efforts so far 🙇

[pgrzesik]

Thanks once again @juanjoDiaz - as mentioned in discussion above, we'd like to follow more iterative, small-steps path - I'm going to close this PR for now.