155289 - String.prototype.XXX.length has some wrong values

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect)

Description

[rogerl (gone)]

>String.prototype.indexOf.length
2  <--- should be 1 (ECMA 15.5.4.7)
>String.prototype.lastIndexOf.length
2  <--- should be 1 (ECMA 15.5.4.8)
>String.prototype.split.length
1  <--- should be 2 (ECMA 15.5.4.14)

Note: the wrong values are supported by the test suite.

Comment 1

[rogerl (gone)]

    Test numbers are:

    Testcase ecma/String/15.5.4.6-2.js failed
    Failure messages were:
    String.prototype.indexOf.length = 1 FAILED! expected: 2
    String.prototype.indexOf.length = null; String.prototype.indexOf.length = 1
FAILED! expected: 2
    delete String.prototype.indexOf.length; String.prototype.indexOf.length = 1
FAILED! expected: 2

    Testcase ecma/String/15.5.4.7-2.js failed
    Failure messages were:
    String.prototype.lastIndexOf.length = 1 FAILED! expected: 2
    delete String.prototype.lastIndexOf.length;
String.prototype.lastIndexOf.length = 1 FAILED! expected: 2

    Testcase ecma/String/15.5.4.8-1.js failed
    Failure messages were:
    String.prototype.split.length = 2 FAILED! expected: 1
    delete String.prototype.split.length; String.prototype.split.length = 2
FAILED! expected: 1

Comment 2

[rogerl (gone)]

Attachment: Fix length values

Comment 3

[Phil Schwartau]

cc'ing reviewers; adding js1.5 keyword; adding as dependency to bug 149801 -

Comment 4

[Brendan Eich [:brendan]]

Comment on attachment 89948 [details] [diff] [review]
Fix length values

sr=[e-mail] -- khanson, can you r= soon?  Or anyone who is cc'd, if
Kenton is on vacation (as I recall)?

/be

Comment 5

[Kenton Hanson (gone)]

Comment on attachment 89948 [details] [diff] [review]
Fix length values

r=khanson.  Brendan could you sr=  bug 145779, it is very similar to this
patch.

Thanks, Kenton

Comment 6

[Phil Schwartau]

Note rogerl is on vacation until July 8. I think we want this 
for the 1.0 branch, too, right?

Comment 7

[chris hofmann]

Comment on attachment 89948 [details] [diff] [review]
Fix length values

a=chofmann for 1.0.1

Comment 8

[Brendan Eich [:brendan]]

Fixed in branch and trunk.

/be

Comment 9

[Phil Schwartau]

Verified FIXED. All three testcases above corrected:

  mozilla/js/tests/ecma/String/15.5.4.6-2.js
  mozilla/js/tests/ecma/String/15.5.4.7-2.js
  mozilla/js/tests/ecma/String/15.5.4.8-1.js


These tests now pass in the debug/optimized JS shells, and in
the rhino/rhinoi shells, since Roger also made the fix there. 
Also fixed on MOZILLA_1_0_BRANCH; adding verified1.0.1 keyword -