Fixing newlines and comments for if/else and do/catch #58
Conversation
|
While this improves things corpus-wise (trading two firefox errors for one new one), I suspect that new error might be higher-impact, and that it's not worth the tradeoff. Basically, we're now suppressing the newline between two declarations if the second has a comment, which seems fairly likely when items have documentation use cases. The fact that this doesn't fail more often tells me that the parser is able to figure this out most of the time, but I'm not sure what the missing case is. I'm going to hold off on merging this until I can either fix the |
Fixes #47 Adds `catch` to the semi suppressor. Also reworks the semi suppressor to not suppress _explicit_ semicolons - this doesn't matter for any corpus cases, but is theoretically more correct, and protects against a future failure involving comments.
Teaches the newline-suppressor to advance through a comment to see if the subsequent token is a newline-suppressing operator (or one of the single-char suppressors). This also requires us to parse multiline comments, since we would otherwise eat the first character of a multiline comment before that component got a chance to.
alex-pinkus
force-pushed
the
else-catch-newlines
branch
from
ededb4d
to
b96a976
Compare
|
OK, this fixes the comment-semi problem that the original commit introduced, so all should be well. |
Fixes #47
Adds
catchto the semi suppressor, and also suppresses semis at thestart of a comment. This also required a bit of rework that's been
overdue for a while, to not suppress explicit semicolons - those don't
get used much, but there was one test case where one of those was
followed by a comment.