351463 - Perl RegExp test 827: /([\d-\s]+)/.exec('a0- z')

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect)

Description

[Bob Clary [:bc] (inactive)]

expected 0- ,0-  (MSIE agrees)
actual   SyntaxError on line 1: invalid range in character class

there are several other character class tests some of which MSIE also flags as syntax errors that I haven't filed.

Comment 1

[Hixie (not reading bugmail)]

The syntax error is correct as far as I can tell and comes from ES4 15.10.2.15 NonemptyClassRanges.

Comment 2

[Hixie (not reading bugmail)]

The spec should be errata'ed. All the browsers seem to handle this fine now, they just treat hyphens next to \d \w \s etc as not being part of a range.

Comment 3

[Brian Crowder]

This is fixed by the patch in bug 375876 (we agree with MSIE6 here, now), and I believe rightly so.

Comment 4

[Hixie (not reading bugmail)]

If we don't raise a syntax error, it's not spec-compliant. I agree that we should change the spec though.

Comment 5

[Brian Crowder]

Yes, I agree; see bug 375876 comment #9

Comment 6

[Bob Clary [:bc] (inactive)]

/cvsroot/mozilla/js/tests/public-failures.txt,v  <--  public-failures.txt
new revision: 1.17; previous revision: 1.16

/cvsroot/mozilla/js/tests/js1_5/extensions/regress-351463-01.js,v  <--  regress-351463-01.js
initial revision: 1.1

Comment 7

[Bob Clary [:bc] (inactive)]

v

Comment 9

[x0]

Reopening this one then.

Either the fix was wrong, because it violates ECMA, or we do that by purpose, then a comment explaining why we do that is missing in the source, see bug 428703.

Comment 10

[Bob Clary [:bc] (inactive)]

this bug is fixed. if we need a comment explaining it or a devnote, please file a new bug.

Comment 11

[Daira Hopwood]

If this is an intentional non-compliance to ES3, please report it to the es3.x-discuss list, and perhaps the spec can be changed.