Repository navigation
Only enable diagnostics and fix settings change detection - #298
Conversation
…ld be disabled by this point
Jake Bailey (jakebailey)
changed the title
|
Test would be good. Like declare a variable, observer squiggly, change file, squiggly goes away, change back, it reappears. Alexander Sher (@AlexanderSher) - do we have something like this for unresolved imports? |
|
There are a few tests in LanguageServerTests already which do checks on the diagnostics. I can just add another one for the settings change. |
|
Eh, it's overloaded. Could still be improved. |
Jake Bailey (jakebailey)
changed the title
|
I've added a test which changes the settings and re-checks the diagnostics. The overall diagnostic testing code and settings editing code is really strange, though, and could probably be improved. All of the "work" for changing settings is actually done in |
) Fixes microsoft#242. microsoft#289 causes a few issues to do with parser-generated diagnostics, and there are more diagnostics I wasn't aware of coming from ErrorSink. This is a simplified PR that does the minimum to enable diagnostics (ignoring "Unspecified" at publish time) and fixes the setting change bug. We'll have to revisit where diagnostics should be stored to access via code actions in a different PR. There are only two diagnostics that actually come out of the server which can be disabled with `python.analysis.disabled` for now. I'll close microsoft#289 in favor of this one.
Fixes #242.
#289 causes a few issues to do with parser-generated diagnostics, and there are more diagnostics I wasn't aware of coming from ErrorSink. This is a simplified PR that does the minimum to enable diagnostics (ignoring "Unspecified" at publish time) and fixes the setting change bug.
We'll have to revisit where diagnostics should be stored to access via code actions in a different PR. There are only two diagnostics that actually come out of the server which can be disabled with
python.analysis.disabledfor now.I'll close #289 in favor of this one.