I think the change proposed by Allen is fine. The main point of the new definition of IdentityEscape is to reserve \p, \X, and other escape sequences involving ASCII letters, to which we may want to assign different interpretations in the future. Allowing \- does not conflict with this.

Norbert


> On Jan 14, 2015, at 0:20 , Mathias Bynens <mathias at qiwi.be> wrote:
> 
> 
>> On 13 Jan 2015, at 22:23, Allen Wirfs-Brock <allen at wirfs-brock.com> wrote:
>> 
>> Would those of you who consider yourselves RegExp experts take a look at https://bugs.ecmascript.org/show_bug.cgi?id=3519  Is this a bug? If so, what is the fix?
>> 
>> This construction for Identity Escape goes back to Norbert's original proposal http://norbertlindenberg.com/2012/05/ecmascript-supplementary-characters/index.html 
>> 
>> Perhaps we need to add a:
>>  ClassAttom[U] :: [+U]  \-
>> 
>> production or some such to the pattern grammar.
> 
> I think it’s a bug — see https://codereview.chromium.org/788043005/diff/220001/src/parser.cc#newcode4354 for the discussion that led to this report.
> 
> Your change would allow developers to use an escaped `-` in a character class, e.g. `/[a-f\-A-Z]/u`, rather than having to move it to the beginning (i.e. `/[-a-fA-Z]/u` or end (`/[a-fA-Z-]/u`) of the character class, as is possible today without the `u` flag. That is a good thing IMHO.
> _______________________________________________
> es-discuss mailing list
> es-discuss at mozilla.org
> https://mail.mozilla.org/listinfo/es-discuss

I think the change proposed by Allen is fine. The main point of the new definition of IdentityEscape is to reserve \p, \X, and other escape sequences involving ASCII letters, to which we may want to assign different interpretations in the future. Allowing \- does not conflict with this.