Use Unicode 6.1.0 when determining whether an identifier is acceptable or not
Project Member Reported by math...@qiwi.be, Feb 21 2012
v8 currently uses Unicode 5.0.0. Here are some examples of identifiers that are currently failing because of this, even though they’re valid according to ES 5.1/Unicode 6.1: * `var \u0cf1;` — http://mothereff.in/js-variables#%5Cu0cf1 * `var \ua7aa;` — http://mothereff.in/js-variables#%5Cua7aa * `var \u1bba;` — http://mothereff.in/js-variables#%5Cu1bba * `var a\ua674;` — http://mothereff.in/js-variables#a%5Cua674 Of course, there are many more. Is the list of allowed characters in IdentifierStart and IdentifierPart auto-generated based on a UnicodeData.txt file, or how is this done in v8?
Feb 22 2012,
Feb 24 2012,
I'll handle this one.
Mar 6 2012,
Fixed in r10933. And to answer your original question: Yes, those tables are automatically generated using the Unicode specification.
Mar 6 2012,
Thanks for fixing this so quickly! > Yes, those tables are automatically generated using the Unicode specification. Would you mind pointing me to the script that generates `src/unicode.cc`? I couldn’t find this in the repository.
Mar 7 2012,
We use an internal tool to generate these tables which is not part of the open-source V8 release.
Mar 23 2017,
Sign in to add a comment