Meta-issue for improving the REPL experience #50817
Comments
|
I am so onboard with this plan! Not confident how much I can promise to work on the code directly, but I'm happy to support via Because Maybe we can add another item to the list:
Because last time I looked at the REPL.jl completion code it's ... a bit of a scary disaster area of regexes and heuristics IIUC |
|
Another item to add to the list:
|
|
I'm wondering, would it be worth making a channel on Zulip/etc. for chatting about this? I think it could be worth making a space for discussion around this, and I'm not sure that just this issue suffices. |
|
@tecosaur I think that's a good idea. I was surprised that a #repl channel doesn't already exist in the Slack |
|
Personally, I'm much more a fan of the Zulip, so if other people are fine with that (yes?), I'll put in a request for a #repl channel there. |
|
I've now made a Zulip stream: https://julialang.zulipchat.com/#narrow/stream/404905-repl 🙂 |
|
FYI, there is now also a Slack channel (that currently basically says "go to Zulip"), and the Slack/Zulip channels should be listed as part of the new "REPL working group" soon (with me as the contact). Meetings etc. TBD, but for now I thought it would be a nice start to better collect people interested in driving this aspect of the Julia experience forward 🙂. |
Fix #51194 This PR fixes a regression introduced in #49294, so I believe it should be backported to v1.10. In the current code, completion of `qux(foo, bar.` is detected by parsing `foo(qux, bar` as an incomplete expression, and then looking for the sub-expression to complete (here, `bar.`). This approach fails however for infix calls, since completing `foo + bar.` starts by parsing `foo + bar`, which is a complete call expression, and so the code behaves as if completing `(foo + bar).` instead of `bar.`. This leads to the current problematic behaviour: ```julia julia> Complex(1, 3) + (4//5).#TAB im re ``` which would be correct for `(Complex(1, 3) + (4//5)).#TAB`, but here we expect ```julia julia> Complex(1, 3) + (4//5).#TAB den num ``` This PR fixes that by trying to detect infix calls. In the long term, all this ad-hoc and probably somewhat wrong string processing should be replaced by proper use of `JuliaSyntax` (as mentioned in #49294 (comment), #50817 (comment) and probably other places), but for now at least this fixes the regression.
Fix #51194 This PR fixes a regression introduced in #49294, so I believe it should be backported to v1.10. In the current code, completion of `qux(foo, bar.` is detected by parsing `foo(qux, bar` as an incomplete expression, and then looking for the sub-expression to complete (here, `bar.`). This approach fails however for infix calls, since completing `foo + bar.` starts by parsing `foo + bar`, which is a complete call expression, and so the code behaves as if completing `(foo + bar).` instead of `bar.`. This leads to the current problematic behaviour: ```julia julia> Complex(1, 3) + (4//5).#TAB im re ``` which would be correct for `(Complex(1, 3) + (4//5)).#TAB`, but here we expect ```julia julia> Complex(1, 3) + (4//5).#TAB den num ``` This PR fixes that by trying to detect infix calls. In the long term, all this ad-hoc and probably somewhat wrong string processing should be replaced by proper use of `JuliaSyntax` (as mentioned in #49294 (comment), #50817 (comment) and probably other places), but for now at least this fixes the regression. (cherry picked from commit e949236)
|
I've added a plot to the end of the initial message that people subscribed to this issue may find of interest, I've also plopped it in a zulip thread. |
|
From slack, Ruby's interactive REPL has had a fun update that we might want to look at in this effort: https://railsatscale.com/2023-12-19-irb-for-ruby-3-3/ |
|
Thanks to a bout of recent merging, this effort seems to be lurching forwards again 😀 |
|
I'm starting an effort to provide a fresh take on |
I've ended up on a path to improve the REPL experience in a few ways, and thought I'd actually write up the overall plan.
Markdown.MDStyled markdown, with a few tweaks #51928REPLstdlib (see Upgradable stdlibs #50697)Here's the vision for the REPL I currently have in mind:
GraphViz code
The text was updated successfully, but these errors were encountered: