feat: implement skipping pointer type

[gabor-boros]

This PR implements skipping pointer types for optional fields to address #1039 and #1061. A new extension is defined to achieve the desired behaviour: x-go-type-skip-optional-pointer.

The extension's name maybe longer than usual, but at least descriptive. If set to true, the type will not be a pointer if the field is optional or nullable. If set to false, the type will be a pointer.

In the case of custom types, like what was proposed in the mentioned issues (Optional[string] or null.Bool), the extension is respected, so it is controllable what to generate (ie. pointer or non-pointer type).

I have already tested this on my own and works very well. Also, added some tests to cover value parsing and code generation. Probably there are cases I didn't prepare for, but I'm happy to add more scenarios.

Upon merge, this PR fixes #388, #382, #823, #1039, #1061 and closes #266, #479

[gabor-boros]

@alexandear I assume you are a maintainer in the repo. May I ask a review from you? I found no contribution guidelines.

[alexandear]

@gabor-boros in fact, I'm not an official oapi-codegen maintainer. And don't have permission to merge this PR. So, we should wait for maintainers.

[gabor-boros]

Ah! I was misunderstanding this from the number of contributions. Sorry for the noise in this case and thank you for the replies!

[gabor-boros]

@alexandear do you have any suggestion who to ask about feedback on this PR?

[alexandear]

eepmap-marcinr is one of the maintainers.

[gabor-boros]

Great suggestion @alexandear, thank you!

@deepmap-marcinr is there anything I could help out here with the review?

[notgurev]

This would be a huge relief from a lot of unnecessary boilerplate.

@deepmap-marcinr this PR is backwards-compatible and shouldn't cause any hate mail (unlike similar PRs you commented on). Any chance it could be merged?

[deepmap-marcinr]

Thank you.