tristanlatr

> with accurate line number information we could pick the right slice from the original source file No idea how you would do that! This sounds like a very hard...

We could try to integrate with ASTTokens: https://asttokens.readthedocs.io/en/latest/api-index.html#asttokens.ASTTokens.get_tokens and use it in the colorizer. This probably means larger refactorings.

Ok, the issue happens only in IOS safari: https://css-tricks.com/annoying-mobile-double-tap-link-issue/ Maybe I won’t revert the whole PR and prepare a fix for it only.

This means major refactoring of the docstrig linker in order to make it work with docutils nodes instead of stan.

See https://github.com/twisted/pydoctor/pull/417#issuecomment-888516326 about type annotations in signatures. I think it’s a good idea to show them I the signature.

Let's talk about rendering the type annotation inside signatures, @cretz, here is what's my old code (#417) currently generates for a random function (not overloaded): As talked about with @glyph,...

There is an implementation of detecting a exception here https://github.com/tristanlatr/pydocspec/blob/1efc4261a83c306391a97a61112d88b4beef7488/pydocspec/__init__.py#L650

Here is my proposition for tackle this issue: For `System.allobjects` and for `Documentable.contents`, we adopt a version of `dict` that maintains a stack of duplicates objects, that we can request...

Ok thanks for the feedback, let me reduce the scope to a set of fields that, I my sens, should really be supported in pydoctor: - `@kwtype`: Alias for `@type`:...

We would need to re-write the epytext parser which is inherited from epydoc, so this is probably a hard task for little benefits. So marking this issue as not being...