Expose `safe_load`ing

[lbolla]

At the moment, yamlenv uses yaml.load to load .yaml files.
We should also support using yaml.safe_load.

[jeverling]

I'm not an expert at all in yaml parsing, but would it work as a workaround to use yml = yaml.safe_load(fp), then tmp = yaml.safe_dump(yml) and then yamlenv.load(tmp)?
I think this should prevent safety issues or am I wrong?

[lbolla]

@jeverling That's pretty inefficient, though. Also, I am not sure Python objects like tuples would be handled wrongly. E.g., see how load+dump a tuple transforms it in a list, when using "safe":

>>> yaml.safe_load(yaml.safe_dump((1,2,3)))
[1, 2, 3]
>>> yaml.load(yaml.dump((1,2,3)))
(1, 2, 3)

[jeverling]

Thanks, that makes sense!

[jeverling]

Thanks for the great project btw! I was looking at the way docker-compose interpolates variables in docker-compose.yml files, but your library is much nicer than trying to reuse what they do.

[lbolla]

Thanks! I am happy to accept a PR if you have time to work on it!

[jeverling]

Not right now, but I hope I can find the time in the next weeks.

[thisHermit]

Hey @lbolla, I was interested in working on this. I am not sure what to do though.
I need to solve to it for kibitzr over here.

Your help is much appreciated.

[lbolla]

@thisHermit I think supporting safe_load requires changing the Loader inheritance with SafeLoader.

Have a look at PyYaml docs on how to do it.

I'd be good to have a unittest that highlights the problem with the current implementation. I am OK with replacing the simple Loader with SafeLoader, btw: no need to support both, imo.