Move to monaco-editor

[rumbin]

Current situation

There are several open issues that are related to the code editor, e.g.
#12429
#9990

Switching to monaco-editor has been discussed several times and was mentioned as a post-1.0.0 task: #10254, #10267, #12423.

Suggested change

Switch to monaco-editor, the editor behind VS Code.

Personally, I have worked with websites using this editor (dbt Cloud, dataform) and feel that it works very well.

Alternative

Stick with the current editor or consider others.

[rumbin]

I'd like to raise this issue again.
I just gave a Superset/SQL training to our users and I could see how they were unhappy about the Ace editor.
Some even resort to different tools like the Snowflake web frontend for this reason.

[villebro]

@betodealmeida thoughts?

[betodealmeida]

@villebro I think we should give it a try (I also mention it in SIP-68).

[ktmud]

I plan to work on this this year.

[srinify]

Any thoughts on CodeMirror? @ktmud @rumbin

https://blog.replit.com/code-editors#head-to-head

[ktmud]

@srinify Thanks for sharing the analysis! It seems their main reason of switching to CodeMirror 6 is that Monaco doesn't support mobile and they need to support a lot of languages. I don't think mobile support for code editing is that important for Superset so I'm still leaning towards Monaco for it's richness of config options and best-in-class autocompletion.

[srinify]

Ah interesting

[rusackas]

Anyone on this thread still have the intent to carry this forward and own the task? Wondering if whether should keep this open, convert this to a Discussion as a feature idea if we're looking for new volunteers, or close it if there's no longer any steam behind the initiative.

[rumbin]

This task exceeds my coding capabilities by far, but I'd assist with testing and discussion, if needed.

It would be so useful to switch to an editor that behaves like many of our users are used to from VS Code.

[ktmud]

Sorry for not carrying this forward. Had some priority shift internally and got pulled into other projects. Feel free to convert this into a discussion as it is a feature request and I think we now generally keep feature ideas as discussions.

If we ended finding a volunteer to work on this, I'd be happy to help with code review still.

[rusackas]

It's been another year since anything has happened on this thread. I think we should close this out as inactive/stale as well as because it's a feature request rather than a bug, per se (even if it would solve other reported issues). If anyone wants to pursue this, I think it might warrant a SIP or at least a discussion thread for lazy consensus, but there are indeed plenty of reasons to consider it. Say the word if you want to contribute and want any help in the process to make it happen!

[rumbin]

@rusackas I neither have the skills nor the capacity to make this happen. But still I feel that switching to Monaco editor would be a huge step forward for Superset/SQLLab.

[rusackas]

I don't imagine it'd be easy, but you're welcome to start a Discussion or take it to Slack and try to recruit interested parties around it, if you want to help make it happen :D

[chintanp]

I think closing this as "Completed" is incorrect. If it tracked elsewhere, then it is fine.as otherwise this is still a major issue as far as usability of Superset is considered. Code Editor is a very important requirement and the current editor is fairly buggy to be not useful. Please suggest alternatives, if possible. @rumbin have you found any other workarounds? Is the Snowflake web frontend you mention open-source and free to use?

[rumbin]

@chintanp our users are, to be honest, resorting to Snowflake's Snowsight web UI, as they are frustrated with SQLLab's editor.
Furthermore, we are using Grafana for time series charts if raw data and for histogram and CDF distribution charts. The OSS Grafana Snowflake connector is now using the Monaco editor...

So these are our unpleasant workarounds.

[rusackas]

Admittedly, I default to "completed" more often than I should, rather than selecting the other issue methods... I'll try to be better about that.

Nonetheless, if anyone wants to replace Ace Editor with something better, we're all for a SIP proposal to detail the UI changes/challenges, alternatives considered, and technical/migration plan. Just open an Issue ith the SIP template and you can spearhead that project.

Otherwise, if there are issues with the current editor making things unusable, we'd love to see those filed as issues so we can tackle them, whilst other grand plans are being scaffolded.

CC @yousoph @betodealmeida @eschutho @justinpark who might care about such a project.

[betodealmeida]

I'd love to see this, but I'm currently working on SIPs 85, 95, and 117. Hopefully someone can take this work.