[Python] Nullable bugfix in model.mustache

[MalteEbner]

Description of the PR

Before bugfix: a required and nullable attribute set to null/None causes a ValueError when received by the client.
After bugfix: This error is only caused for properties being required and not nullable
General info: The python client cannot distinguish between unset attributes (value=None) and attributes set to null (value=None)

Issue to solve with a good description of the bug:

closes: #10535
closes: #8710

PR checklist

• Read the contribution guidelines.
• Copied the technical committee to review the pull request if your PR is targeting a particular programming language. @kenjones-cisco

Performed tests:

Tested with own api-specs and the swagger-cli:

• The generation works.
• The two lines causing the error are removed in the python client, nothing else changes.

[aarighi]

Hi, any update on this? Do you know when it will be merged? Thank you (:

[MalteEbner]

Hi, any update on this? Do you know when it will be merged? Thank you (:

Sorry, I would love to see it merged, but do not know get the maintainers to do it.

[aarighi]

Hi, any update on this? Do you know when it will be merged? Thank you (:

Sorry, I would love to see it merged, but do not know get the maintainers to do it.

I don't know if it may be relevant, but from the PR preview it looks like there are missing tasks required to close it

Any idea on how to fix this? I guess all 4 must be checked before the PR can be merged

[MalteEbner]

Any idea on how to fix this? I guess all 4 must be checked before the PR can be merged

The tasks are one for each checkbox in the PR description. They are not necessary to get it merged, see
#10985.

I guess the issue is rather that @kenjones-cisco did not find the time to review and he's responsible for the python part: https://github.com/swagger-api/swagger-codegen/#swagger-codegen-technical-committee

[fedepad]

Any progress here? Does something need to be changed in this PR or not to be good to merge and fix this problem?
Thanks for the info!

[MalteEbner]

Any progress here? Does something need to be changed in this PR or not to be good to merge and fix this problem?
Thanks for the info!

I guess the problem is that @kenjones-cisco is still not active. Perhaps you could update us here @HugoMario, as you have been the last to update this file, according to the git blame.

[bonnici]

I just ran into this problem as well, would be great if we can get this merged and released. I will have to work around it by just removing all the required attributes from the JSON file before generating code, which isn't ideal.

Here is the jq command I used to do that in case anyone else wants to do the same workaround:
jq 'del(.definitions | .. | .required?)' api.json > api.processed.json

Not sure who can approve but just from the recent people who have committed code, can anyone help? @frantuma, @HugoMario, @gracekarina

[ampuerod]

How can we get this change merged? I can't believe it's been there for more than a year.

[mhaley-tignis]

@HugoMario @keithbsb this seems like a pretty easy change for any maintainer to move along. Thoughts on helping us out?

[antoineauger]

We also ran into this bug.

Maybe @frantuma could unblock us all here? He seems to be the most recent active maintainer 😁

[DavidMetaphysic]

I'm also running into this issue.

[VKlueber]

I´m also running in this issue.

[mohit1337]

Can this be please merged? It is perfectly fine to have required fields which are nullable. Having a key marked as required just means it will be present in the output, it doesn't make any guarantees about the value, and null is a perfectly valid value.

[daniel-kmiecik]

I need to create a new PR request because there were problems building the PR from the fork. New PR is #12630. The author of change is still @MalteEbner. Thank you very much for your contribution and patience. Fix should be released in next week.

[thibgc]

#12630 was merged into master, and is now part of release 2.4.49 two days ago! Thank you @daniel-kmiecik! 🥳