Kirill Müller
Kirill Müller
From https://github.com/igraph/rigraph/pull/1602#discussion_r2000607282: > Should we leave deprecated functions unchanged? To be clear: I don't think we should ever touch deprecated functions. Just ensure they are as stable as we can...
To fix this, we'd need a class for strings that contain URLs. I don't find one in https://github.com/krlmlr/awesome-vctrs/, but perhaps there is one? We could then use link texts for...
🤖 This issue has been assigned to Copilot for automated handling.
I like the initiative, only minor remarks: - `supports_table_alias_with_as` and `supports_window_clause` : I don't understand what's special about these, compared to, e.g., the translations. Why not also store the translations...
We can call `rlang::arg_match()` in the manual wrapper in such cases, before forwarding to `_impl()`.
Scratch that, we probably need `igraph.match.arg()` also in the manual wrapper.
It's a case-insensitive version of `match.arg()`, and `rlang::arg_match()` is a version of that with better behavior and error messages.
Now using the new `switch_igraph_arg()` for translating enums (=R strings) to integers.
Thanks. When you run interactively, what do you see with `getOption("width")` ? Can you replicate with a reprex by setting the corresponding width?
Thanks. I see the error in Positron too.