[CS2] Unicode code point escapes on 2, fixes #4248 #4520
Conversation
|
@helixbass Could you ping me when the diff is readable? :) |
|
@lydell sorry I wrongly assumed that once |
src/lexer.coffee
Outdated
| replaceUnicodeCodePointEscapes: (str, options) -> | ||
| shouldReplace = options.flags? and options.flags.indexOf('u') is -1 |
There was a problem hiding this comment.
@lydell as currently written, options won't always contain a flags key (only does if this is for a regex, in which case flags will be set to the regex flags string) and not in seems to break with an undefined arg. Does seem like there should be a simpler way to write this though - I guess you could do something like options.flags ?= 'u' before 'u' not in options.flags to handle the string case but that seems confusing semantically. Or could refactor to pass in eg dontReplace option instead of flags but it seemed very clean to just pass around existing flags and then have the logic for interpreting it with regards to code point escape replacement be contained in this method
There was a problem hiding this comment.
I meant only replacing the right side of the and.
There was a problem hiding this comment.
@lydell ok yup that reads well, updated
@GeoffreyBooth requested that I refactor toJS() into a shared test helper so I opened a pull request #4522 for that (against master). Perhaps wait to merge this one until we've gone through the cycle of merging that into master, merging master into 2, and re-merging 2 here (or it may not matter if that refactor doesn't cause any merge conflicts here)?
GeoffreyBooth
changed the title
|
Unfortunately, there are merge conflicts now :( |
Fixes #4248: contains the Unicode code point escape handling specific to
2branch (as suggested by @lydell)This currently also includes some commits from
master(including my code point escape stuff targeted againstmasterie #4498) that haven't been merged into2yet. If it's preferable to separate that out, I can open a pull request for my2_merged_masterbranch (against2branch), which just containsmastermerged into2But otherwise here is the diff of just my
2-specific changes. Basically we just have to pass around the (he)regexflagsin order to be able to distinguish non-u-regexes (whose code point escapes should be rewritten to normal Unicode escapes) from strings/u-regexes (whose code point escapes should not be rewritten)