Kohsuke Kawaguchi

Unfortunately, too many whitespace changes make it difficult to review this diff

Thanks, but I'm afraid this is a breaking change that affects the current client applications, which means I cannot take it. I'm not sure what actual problem you are trying...

I can't fix this without breaking backward compatibility, as this commit changes the default handling of a method with `@Argument`. The closest change I can think of is to only...

This is great! I see that you've moved directory structures around, which looks to me like an unnecessary change. What is the motivation for that? I'd like to merge the...

`GHPullRequest` extends from `GHIssue`, so the design here is to call `GHPullRequest.fetchIssue()` at the right point to ensure methods defined on `GHIssue` return correct info when used on `GHPullRequest` instance,...

Translating the requirement: it'd be convenient to allow `PagedIterator` to skip forward without actually retrieving the objects in between.

Indeed. This comes from the fact that we rely on AST transformation to insert interceptors. Hence we generally do not know how the property access syntax like this is routed...

I started thinking what the same approach would mean for method calls. On the regular method call side, most regular calls result in `MetaMethodSite`, which invokes `MetaMethod.invoke()`. That would help...

on the method call side, the method to create a call site is strangely defined in `CallSiteArray.createCallSite`, which is a private method and there's no call path to use this...

I've released 1.21 with better implementation of this that hopefully hides the overhead. Keeping this pull request open in case he still complains, but we shouldn't merge this until that's...