)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":15197,"name":"Pierre Riteau","email":"pierre@stackhpc.com","username":"priteau","status":"StackHPC"},"change_message_id":"42349e2cecb6ab89a65594337fe042777fb5ec0b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"412b4b24_83d3724a","updated":"2022-05-31 08:39:32.000000000","message":"I\u0027ve looked more into this issue and I am not sure this is a bug in Voluptuous. Rather, I think we may be using this library in a wrong way.\n\nThe Coerce class expects a type as argument, but we are passing the function tzutils.dt_from_iso.\n\nSee docs: http://alecthomas.github.io/voluptuous/docs/_build/html/voluptuous.html#voluptuous.validators.Coerce","commit_id":"5726db9a443791e89677c7b908878cc9b730a899"},{"author":{"_account_id":15197,"name":"Pierre Riteau","email":"pierre@stackhpc.com","username":"priteau","status":"StackHPC"},"change_message_id":"fe68ca0335246d729224b480a676e6094d2641de","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"7df23884_1aef3213","updated":"2022-05-31 09:12:43.000000000","message":"Trying an alternative approach in https://review.opendev.org/c/openstack/cloudkitty/+/844015","commit_id":"5726db9a443791e89677c7b908878cc9b730a899"},{"author":{"_account_id":28356,"name":"Rafael Weingartner","email":"rafael@apache.org","username":"rafaelweingartner"},"change_message_id":"c944fce6610bb9db1b321f0a2a007629a495b275","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"84676108_02c24821","in_reply_to":"412b4b24_83d3724a","updated":"2022-05-31 11:26:52.000000000","message":"That was my doubt as well, if it really expects a type or a function. I also saw the commit that broke the code. Therefore, I assumed that a bug/limitation was introduced (by mistake) in voluptuous.validators.Coerce, and not the other way around (we using it incorrectly).\n\nI found on the Internet other people using Coerce with a function. Therefore, I assumed that it was expected to work with a function (which in fact, it does work). The only problem is that if we use the default error message (which was changed to handle Enums, and then broke its workings with functions), which we do not need. That is why I added a custom message to bypass that problem that was introduced into it.\n\nDo you think that the other is better? The use of schema actually gives me the idea of types being necessary, instead of a function.","commit_id":"5726db9a443791e89677c7b908878cc9b730a899"}]}