Edoardo Luppi
Edoardo Luppi
I'll be honest tho, the current actual behavior looks noticeable worse than what we'd expect. It's like seeing a lost parenthesis in the middle of the code.
Are you sure the double step is necessary? Given the alpha status I'd not even bother with deprecations and make them internal now. I don't think there are many people...
Thanks for the report! Will be fixed in the next release, which I'll publish on the weekend.
My bad, this seems to be a platform issue: https://youtrack.jetbrains.com/issue/IJPL-148354
It's not really a bug, but more of a deliberate choice to avoid null checking in custom actions. As you can see by the comment above in the .stg file,...
IIRC the MySQL grammar we use for benchmarking has got actions that reference the parameter. I'll have a look soon tho, don't have much free time lately.
I think I remember why we have that `UNSAFE_CALL`. Let's take this grammar. ```text assignment : ID ASSIGN expression { $text == "something" }?; ``` `$text == "something"` is translated...