a test case to illustrate issue #14871

[parmentelat]

References

issue #14871

Code changes

one more test case, with related input and outputs

notes

• the test right now breaks, because of Find/replace highlighting keeps coming back up after find prompt closed #14871 precisely
• also at first the outputs are made on darwin - will get rid of that once the correct linux outputs are produced

User-facing changes

none

Backwards-incompatible changes

none

[jupyterlab-probot]

Thanks for making a pull request to jupyterlab!
To try out this branch on binder, follow this link:

[parmentelat]

also I had to run git commit --no-verify, the prettier stage was failing, even though jlpm run prettier seemed happy
please let me know how to deal with that

[krassowski]

jlpm run prettier is only prettier, there is also eslint which might have caught that getSelectionRange is unused and needs to be used or removed. In general jlpm run lint from root of the repo should work (it includes both prettier and eslint).

[parmentelat]

I must have misunderstood, I had the impression it was the prettier stage that was wrong..
anyway, I have removed the unused function and it is now all good wrt committing
thanks @krassowski

[parmentelat]

trying to improve the situation, here's a few thoughts to share

level1

the first-level bug has to do with the search dialog having lasting effects
and indeed, it will be an obvious improvement
if giving up on the search dialog (using either Escape or the cross icon)
brings us back in a state where nothing is being searched

level2

however, playing with the search a little more, I came to realize that,
when the search dialog is still on, it feels like the job in charge of outlining stuff
sometimes has a very intrusive behaviour, notably putting some cells in edit mode while they were not,
which is imho wrong in itself, and worse still it often leads to arbitrary cells becoming in focus (in the viewport)
as a result this intrusion is very often devastatingly confusing for the user,
because it makes one lose one's train of thought by scrolling up or down with no apparent sense

my own experience has been to face level2, so I tried to get rid off the issue by turning off the search dialog,
but it was not enough because of level1

So I guess what I'm saying is, I might eventually be able to fix stage1 as part of this PR, I'll try at least
but even so, there will remain some agreement to reach on how the search outlining business
is expected to behave, and imho the expected behaviour would be:

• no change of editing mode
• no scrolling whatsoever

but I guess that's just me
and apologies for getting a little off-track as compared to the initial target of this PR

[krassowski]

Thanks again for your work here @parmentelat 🎉, closing this in favour of #15443 which includes your commits from this PR. Please feel welcome to review the new PR :).