Warn if yaml has duplicate keys

[Matistjati]

Jotting this down: I encountered an example in the wild with something along the lines of <problem>/data/testdata.yaml having

input_validator_flags: n=100
input_validator_flags: n=1000

Pyyaml silently resolves this by taking the last one. Checking "every key in every yaml file is unique" seems like a nice sanity check

[gkreitz]

Seems like a bug in upstream, yaml/pyyaml#41 and yaml/pyyaml#165 . Open since 2016, so I'm guessing they're not going to fix it.

The discussion there contains some suggested workarounds, I guess we'll have to adopt one of those.

[Aevumere]

I maintain a small free strict-YAML checker and tried the exact two-line sample from this issue. It reports found duplicate key input_validator_flags — line 2, column 1: https://yaml.aevumere.com/yaml-preflight?source=github-kattis

It is only a manual preflight, not a problemtools integration. The public checker now parses entirely in the browser: the YAML is not sent to the server; only a boolean valid/invalid usage signal is recorded.

[gkreitz]

Thanks for the suggestion. We will not upload all yaml we parse somewhere just to lint yaml files. :)