kobotoolbox / kpi / 16322266002

81%

master: 76%

push

github

fix(auditLogs): catch TypeErrors from copying request data DEV-711 (#5920)

### 📣 Summary
Fix an error that was preventing submissions with large uploads.


### 💭 Notes
Fixes an issue wherein submissions with large uploaded files attached
were failing. This was due to an issue in the audit log code, where we
copied request.data to the internal WSGI request for later processing by
the middleware. The bug emerged because Django uses
`TemporaryUploadedFile`s when uploading files larger than the configured
max size, and those cannot be copied using deepcopy.
This PR just catches that error, logs it, and moves on. It's not the
most foolproof of solutions if we ever need to log those requests but
for now we don't and this is blocking a fairly urgent fix (DEV-708). It
also means we don't have to go into the request.data() dictionary to
figure out what is in there before trying to copy it. Since
request.data() could be fairly large it will save some time.


### 👀 Preview steps

1. ℹ️ have an account and a project with an upload question
2. Update FILE_UPLOAD_MAX_MEMORY_SIZE to 100
3. Add a submission including a file
4. 🔴 [on release] The submission fails with `TypeError: cannot pickle
'_io.BufferedRandom' object`
5. 🟢 [on PR] The submission goes through

6147 of 9740 branches covered (63.11%)

6 of 6 new or added lines in 1 file covered. (100.0%)

23208 of 28731 relevant lines covered (80.78%)

0.81 hits per line