Spellcheck runs in code cells but not markdown cells

[psychemedia]

In JupyterLab 4.1.8, the spellchecker is (correctly) highlighting spelling errors in code cells, but not identifying any spelling errors in markdown cells.

In the devtools console, I see the following warnings as I type, though it's not clear to me what might be raising them or whether that issue relates to the spellchecker's incorrect behaviour?

highlights.ts:343 Could not get highlights: TypeError: Cannot read properties of undefined (reading 'isReady')
    at Qt.requestHighlights (highlights.ts:217:20)
    at a.factory [as _factory] (jlab_core.b91074b1e57242ddcd31.js?v=b91074b1e57242ddcd31:1:1758915)
    at a._execute (jlab_core.b91074b1e57242ddcd31.js?v=b91074b1e57242ddcd31:1:1757383)
    at o (jlab_core.b91074b1e57242ddcd31.js?v=b91074b1e57242ddcd31:1:1756803)

[krassowski]

For me spellchecking in markdown works great. This error you show comes from jupyterlab-lsp, not from spellchecker; it is possible that that error could be breaking the editor logic and thus the spellchecker.

https://github.com/jupyter-lsp/jupyterlab-lsp/blob/b159ae2736b26463d8cc8f0ef78f4b2ce9913370/packages/jupyterlab-lsp/src/features/highlights.ts#L208

Can you open an issue on https://github.com/jupyter-lsp/jupyterlab-lsp showing how to reproduce this error (and listing versions of the extension, and language servers installed).

[psychemedia]

@krassowski ok, thanks (though I would expect it to be more likley that any jupyterlab-lsp error would cause issues with code cells rather than markdown cells?)

[krassowski]

In Lab 4 LSP is also supported in Markdown cells.

[psychemedia]

It seems there's a conflict with jupysql-plugin that breaks the spell checker in markdown cells (issue)