[CSHARP] Fix how the array type is set when using NULLABLE_REFERENCE_TYPES

[Mattias-Sehlstedt]

Fixes #22066

PR checklist

• Read the contribution guidelines.
• Pull Request title clearly describes the work in the pull request and Pull Request description provides details about how to validate the work. Missing information here may result in delayed response from the community.
• Run the following to build the project and update samples:

  ./mvnw clean package || exit
  ./bin/generate-samples.sh ./bin/configs/*.yaml || exit
  ./bin/utils/export_docs_generators.sh || exit
  

  (For Windows users, please run the script in WSL)
  Commit all changed files.
  This is important, as CI jobs will verify all generator outputs of your HEAD commit as it would merge with master.
  These must match the expectations made by your contribution.
  You may regenerate an individual generator by passing the relevant config(s) as an argument to the script, for example ./bin/generate-samples.sh bin/configs/java*.
  IMPORTANT: Do NOT purge/delete any folders/files (e.g. tests) when regenerating the samples as manually written tests may be removed.
• File the PR against the correct branch: master (upcoming 7.x.0 minor release - breaking changes with fallbacks), 8.0.x (breaking changes without fallbacks)
• If your PR solves a reported issue, reference it using GitHub's linking syntax (e.g., having "fixes #123" present in the PR description)
• If your PR is targeting a particular programming language, @mention the technical committee members, so they are more likely to review the pull request.

[wing328]

cc @mandrean (2017/08) @shibayan (2020/02) @Blackclaws (2021/03) @lucamazzanti (2021/05) @iBicha (2023/07)

cc @devhl-labs

[Mattias-Sehlstedt]

I have reverted this type calculation to the previous static array-mapping, since the more dynamic .getSchemaType(target) could modify the type with ? based upon NULLABLE_REFERENCE_TYPES (which is undesirable).

[devhl-labs]

What is different about this new spec? Can you just put the diff in the http-signature spec and revert this? If not, I suggest leaving the new spec and adding a new file to bin/configs, but still reverting this.

[Mattias-Sehlstedt]

For the new spec I have added objects that are inline aliases.

list:
  items:
    type: string
  type: array
deepList:
  items:
    items:
      type: string
    type: array
  type: array

I initially placed it in the original specification, but since that was used everywhere it caused such a huge delta that I thought it was clearer to isolate the representation in a separate specification.

But if we prefer to have fewer specifications that represent a wider array of scenarios I can change it so that these models are included in the main specification.

[devhl-labs]

I prefer fewer specs. If it doesnt work in RestSharp or HttpClient then we can use a new spec minimized as much as possible.

[Mattias-Sehlstedt]

Then I'll look into reverting the new specification and instead extending the existing one.

A question on another topic, how should we best handle the scenarios that have been presented here?

TL;DR: The NULLABLE_REFERENCE_TYPES has more scenarios now with deep aliases. These are difficult to represent with the current approach. How should one best leave them out to make it clear that they are not supported? Is the GitHub PR and issue ticket and our discussion here sufficient?

Longer: Previously the Csharp generator did not support deep aliases (more than one in depth), it failed by generating code that accidentally compiled (it just made the type List). Now with support for deep aliases, there are all of a sudden a lot more of possible NULLABLE_REFERENCE_TYPES scenarios (as per the comment linked earlier).

My current understanding of how the setting is realized is that it injects the config with the mustache file and !nrt at the end of the type. This has previously worked, since the depth was only one (so either inject it at the end, or not). Now with a dynamic depth, this is not really suitable for a mustache config anymore. It would be easier and clearer if the type was completely calculated in the Codegen. Implementing and testing this is a large task, and not something I can myself fix easily. I wonder how to best handle this?

The current solution presented would generate valid code, it is just that it will be Option<List<List<string>>?> when it should have been Option<List<List<string>?>?> for deep aliases (i.e., any ? that isn't the last will be omitted).

But as previously stated, this has never been supported before to begin with, so is it something that one should consider too much?