intellij-haskell
intellij-haskell copied to clipboard
rikvdkleij
Syntax check must treat dot and apostrophe as delimiters
Same goes for foo'barBazQux — the foo'bar part gets underlined, the BazQux is not.
Expected: Data.IO syntax should be checked as two words, Data and IO. In case of foo'bar, it should be treated as foo and bar.
The issue reproduction rate is not 100%, sadly.
Seems that the spellchecker is using camel case to parse the text. I always disable the spellchecker for code and literals.
Can you explain more?
Updated issue description
The PSI-tree … is correct
it is to me as well! Crazy, right?
I always disable the spellchecker for code and literals.
well, I don't as I've actually found it quite useful
There is also some hidden Java identifier "logic" hidden in IntelliJ where I do not get rid of.
I already spend so/too much time on it. See e.g. https://youtrack.jetbrains.com/issue/IDEA-204576
Oh, so it's the upstream issue, got it. Totes fine. Thank you for your work.
Anyway let's leave this issue written down here for the bookkeeping purposes. Also, since this issue doesn't seem to be duplicate of https://youtrack.jetbrains.com/issue/IDEA-204576, do you want me to mirror this one for them?
Also, since this issue doesn't seem to be duplicate of https://youtrack.jetbrains.com/issue/IDEA-204576
Yes, that's right.
do you want me to mirror this one for them?
Yes, maybe they can help. Thanks!
Still an issue?
yes
Did you create a Jetbrains IntelliJ issue?
nope, sorry
@rikvdkleij interestingly enough, the Scala plugin is affected as well. Guess it's really upstream.
@develop7 Thanks for your feedback!
Didn't you are also a Scala developer 😉
I've filed a follow-up issue at https://youtrack.jetbrains.com/issue/IDEA-254531, feel free to chime in just in case I've missed anything.
Didn't you are also a Scala developer
Since Scala is a requirement to contribute to ij-haskell, I am :)
Now the aforementioned upstream issue got resolved with the following answer:
In case of Haskell and Scala that behavior should be tuned on the language plugin side. There is SpellcheckingStrategy that makes possible introduction of additional splitters of words before spellchecker will see them.
Which puts the ball back to our side