Do not raise on missing contentChanges in didSave

[krassowski]

References

Fixes #185.

Code changes

Reading through the specs, now I would have implemented shadow_virtual_documents differently, but I guess it is too early for a larger refactor. Let it be a simple bugfix for now...

User-facing changes

None

Backwards-incompatible changes

None

Chores

• linted
• tested
• documented
• changelog entry

[krassowski]

@bollwyvl would it make sense to move jupyter-lsp out of beta after merging this?

[krassowski]

bump? just waiting for changelog and not sure which way to go ;)

[bollwyvl]

out of beta

Well, it's starting to get used in other places. The biggest things I could see changing before having something that we sort of commit to are:

• supporting tcp servers (mostly so that the pygls-in-IPython demo is more compelling)
• changing the URL schema to be based on the entry_point rather than mucking about with languages

But it's all beta anwyay!

Does this quiet down the log messages as well? I'm not too worried about it either way, just wondering.

[bollwyvl]

👍

[bollwyvl]

It's still all dark magic to me, but the test does test the condition.

[krassowski]

Thanks for the review!

Does this quiet down the log messages as well? I'm not too worried about it either way, just wondering.

Have not been able to reproduce on master; I have a suspicion that the report of excessive logging was due to my mistake and pushing to PyPI with one diagnostic print in the code (oops)... Should get fixed now either way.

The biggest things I could see changing before having something that we sort of commit to are:

I guess that we can use semantic versioning to announce that (i.e. 1.0).