The per-file CodeChecker Bazel rule doesn't generate metadata.json files. This is because every CodeChecker process generates its own metadata.json. See here

I'm not sure how to proceed with this.

We could, in theory, merge the contents of these metadata files into one.
(
This would entail:

• concat "result_source_files"
• For each analyzer: "analyzer_statistics"
  • add numbers in failed / successful
  • concat failed/successful_sources
• add number in "skipped"
• update "timestamps" by setting begin to the lowest, end to the highest value
• For everything else, assume they are the same (All analysis should run with the same checkers enabled (?))

)

Scope

-- Not sure

DoD

Per-file Codechekcer Bazel rule generates a single file named metadata.json