refactor!: Pass the Codespaces body types by value and rename them to ...Request

[JamBalaya56562]

Updates #3644

This converts the last four CodespacesService body types, removing 8 entries from the paramcheck exception lists in .golangci.yml (4 from body-allowed-pointer-types and 4 from body-allowed-wrong-names).

Checking the four types against the OpenAPI descriptions turned up two problems that are fixed here in separate commits:

1. POST /user/codespaces takes a oneOf body: either repository_id (with an optional ref) or pull_request. CodespaceCreateForUserOptions declared both RepositoryID int64 \json:"repository_id"`andPullRequest *CodespacePullRequestOptions `json:"pull_request"`without an omit option, so every request sent"repository_id": 0and"pull_request": nulltogether, no matter which branch the caller meant.RepositoryIDis now*int64and both fields useomitempty, so only the branch the caller filled in is sent. That oneOf` is also why these two fields stay pointers instead of becoming required value types.
2. The PATCH /user/codespaces/{codespace_name} schema has machine, display_name and recent_folders, but DisplayName was missing from UpdateCodespaceOptions.

A few notes on what is deliberately not changed:

• CodespacesService.Publish keeps its name. The docs page is titled "Create a repository from an unpublished codespace", but the operation id is codespaces/publish-for-authenticated-user, so the current verb already matches the operation.
• CodespacePullRequestOptions keeps its name. It is a nested field type rather than a body parameter type, so the Options suffix rule does not apply to it.
• CreateCodespaceRequest stays shared between CreateInRepo and CreateFromPullRequest. The only difference between the two schemas is that the pull request variant has no ref, which is also why its //meta:schema annotation names only the repository operation.

BREAKING CHANGE: CreateCodespaceOptions, CodespaceCreateForUserOptions, UpdateCodespaceOptions and PublishCodespaceOptions are renamed to CreateCodespaceRequest, CreateCodespaceForUserRequest, UpdateCodespaceRequest and PublishCodespaceRequest; CodespacesService.CreateInRepo, CreateFromPullRequest, Create, Update and Publish now take them by value instead of by pointer; CreateCodespaceForUserRequest.RepositoryID is now *int64.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.54%. Comparing base (9b114ec) to head (9c7bdcd).
⚠️ Report is 3 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #4540   +/-   ##
=======================================
  Coverage   98.53%   98.54%           
=======================================
  Files         195      196    +1     
  Lines       17875    17938   +63     
=======================================
+ Hits        17614    17677   +63     
  Misses        261      261           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[gmlewis]

Thank you, @JamBalaya56562!
LGTM.
Awaiting second LGTM+Approval from any other contributor to this repo before merging.

cc: @stevehipwell - @Not-Dhananjay-Mishra

[Not-Dhananjay-Mishra]

Why are we adding meta:schema request ... to structs?

[JamBalaya56562]

Fair question. They are for the check-schema-fields command in #4375 (not merged yet): it checks an annotated struct's field optionality against the request body schema of the named operation in the OpenAPI descriptions, the same way //meta:operation ties a method to its operation. Running it on this branch is how the missing display_name and the oneOf problem here were found. #4519 added the first one on CreateDependencyGraphSnapshotRequest.

Until #4375 lands they are inert comments, so if you would rather not carry them in unrelated PRs before that, I'm happy to drop them from this PR.