I’m still not sure how I feel about #lang lindenmayer/turtle making its own main submodule in the user’s program. Trying to articulate why: https://github.com/greghendershott/racket-mode/issues/430#issuecomment-591409658

I definitely see the tension between “run the submodule under the cursor” and “run the program the way racket foo.rkt would” but I think I prefer F5 to be the latter

@samth I understand. On the GitHub discussion I suggested maybe what is today called racket-run should be renamed something like racket-enter-submodule-at-point. And a new racket-run should do the “main else file module” behavior. As you suggest. The rest is just a matter of default keybindings for e.g. C-c C-c and <F5>. Which users can change.

The two choices are similar to the two command-line choices: racket foo.rkt and racket plus ,enter foo.rkt.

Where the latter also doesn’t work with #lang lindenmayer/turtle.

My feeling is that is mostly a weakness in ,enter

Note that the implementation of xrepl may well predate submodules

Maybe one way to think about it is that, although module->namespace probably shouldn’t get contaminated with “magic” like “well I mean the main submodule if any, else the file module”, in a world with submodules there could exist some such function and things like xrepl could use it, idk.

It’s not a complicated function to write. It might be one of those functions whose value is mainly it exists and is in documentation so people notice it. :slightly_smiling_face:

Maybe the function is I mean ~= enter!, and that could use an enhancement.

:wave::skin-tone–3:, trying to find a typed definition of nodeset? . I created an opaque wrapper [#:opaque Nodeset nodeset?] but it’s not enough usable to manipulate its content (first, map second …) in a way that works in the untyped world. Or am I going to far and should I use Any?

Any is also not going to help

If it’s a list, then you probably want to use a list type

indeed

looks like some recursive list type