bug 373082 regressed e4x/extensions/regress-353165.js on 1.8.0, 1.8.1, 1.9.0: before BUGNUMBER: 353165 STATUS: Do not crash with xml_getMethod before 64624, after 64624, break 00000000 1: TypeError: set.unrooter is not a function PASSED! 1 before 516992, after 73856, break 00000000 actual: SOME RANDOM TEXT expected: SOME RANDOM TEXT PASSED! 2 after BUGNUMBER: 353165 STATUS: Do not crash with xml_getMethod 1: TypeError: set.unrooter is not a function PASSED! 1 2: TypeError: set.unrooter is not a function FAILED!: Section 2 of test - FAILED!: Expected value: FAILED!: true FAILED!: Actual value: FAILED!: false FAILED!

bclary nominated this as blocking the 1.8 releases. What's the significance of this regression? Is it easy to fix? Is this really functionality people are relying on?

(In reply to comment #1) > bclary nominated this as blocking the 1.8 releases. What's the significance of > this regression? Is it easy to fix? Is this really functionality people are > relying on? The test case relied on the details of the algorithms used to lookup XML functions which bug 373082 changed. So the fix is to change the test just to test for no crash, not to much the results against a particular algorithm.

(In reply to comment #2) > The test case relied on the details of the algorithms used to lookup XML > functions which bug 373082 changed. So the fix is to change the test just to > test for no crash, not to much the results against a particular algorithm. ^^^^^^^^^^^^^^^^^^^^^^ I meant "to match"

Checking in regress-353165.js; /cvsroot/mozilla/js/tests/e4x/extensions/regress-353165.js,v <-- regress-353165.js new revision: 1.2; previous revision: 1.1 -> INVALID.