Upload failed static analysis parses to backend

[giovanni-guidini]

In codecov/codecov-cli#271 we are now ignoring some files that fail to be properly parsed by static analysis.

There's only 1 type of error that is ignored so far. This is on purpose because the parser is not suuuuper battle tested yet (even though it parses all of our code and all of Sentry's code and some other code, but still)

The point is, what to do with errors?
So far we are only printing them in the CLI output so users can know about them.

There was a suggestion by @matt-codecov that we should upload them to our backend services, which is good cause then we can present them in the UI or PR comment or somewhere else.

Granted, we are not a linter nor static analysis tool (at least yet), but maybe this is a good idea.

This info can be saved in StaticAnalysisSuite (which is per commit)
And uploaded with the initial "here are all the file hashes I've just found" request to api by the CLI.

Still up for debate.