Sanitize output and chunk it at DATA_UPLOAD_MAX_MEMORY_SIZE

[humitos]

before sending it over the API and get a 400.

Related

• https://github.com/rtfd/readthedocs-ops/pull/444
• https://github.com/rtfd/readthedocs-corporate/issues/438
• https://sentry.io/read-the-docs/readthedocs-org/issues/712687019/activity/
• https://readthedocs.org/projects/test-builds/builds/8250601/

[codecov]

Codecov Report

Merging #4982 into master will decrease coverage by 0.02%.
The diff coverage is 50%.

@@            Coverage Diff             @@
##           master    #4982      +/-   ##
==========================================
- Coverage    76.8%   76.77%   -0.03%     
==========================================
  Files         158      158              
  Lines        9937     9948      +11     
  Branches     1241     1243       +2     
==========================================
+ Hits         7632     7638       +6     
- Misses       1973     1976       +3     
- Partials      332      334       +2

Impacted Files	Coverage Δ
readthedocs/doc_builder/environments.py	86.27% <50%> (-0.79%)	⬇️
readthedocs/vcs_support/backends/git.py	81.98% <0%> (-0.3%)	⬇️
readthedocs/projects/models.py	85.55% <0%> (ø)	⬆️

[humitos]

@stsewd thanks for the review! Since it's sensitive change, I'd like to wait for another review from the core team before merging.

[ericholscher]

Looks good. I'm curious how many of these we get -- perhaps we should log something about the project name and build step or something? I doubt we'll do anything about it, so happy to just hack it and be done also.

[humitos]

perhaps we should log something about the project name and build step or something?

Wow! I missed this. I 👍 on logging. I'm sending a new commit with this change. Thanks to raise this up!