123624 2013-11-01 10:47:18 -0700 Precision errors in Math.cbrt() on GTK, EFL 2013-11-04 10:01:27 -0800 1 1 1 Unclassified WebKit JavaScriptCore 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 zan zan commit-queue darin gyuyoung.kim msaboff oliver oldest_to_newest 945944 0 zan 2013-11-01 10:47:18 -0700 Math.cbrt(), introduced in r158401, is returning unexpected values in the js/math.html layout test (and the related stress tests) that are indicating precision errors. This only occurs on GTK and EFL. http://trac.webkit.org/changeset/158401 The precision error originates in the underlying cbrt() call in JSC::mathProtoFuncCbrt. pow() works as expected, FWIW. Diff: --- /home/slave/webkitgtk/gtk-linux-64-release/build/layout-test-results/js/math-expected.txt +++ /home/slave/webkitgtk/gtk-linux-64-release/build/layout-test-results/js/math-actual.txt @@ -184,8 +184,8 @@ PASS Math.cbrt(NaN) is NaN PASS Math.cbrt(0) is 0 PASS Math.cbrt(-0) is -0 -PASS Math.cbrt(27) is 3 -PASS Math.cbrt(-27) is -3 +FAIL Math.cbrt(27) should be 3. Was 3.0000000000000004. +FAIL Math.cbrt(-27) should be -3. Was -3.0000000000000004. PASS Math.cbrt(Infinity) is Infinity PASS Math.cbrt(-Infinity) is -Infinity PASS Math.cosh(NaN) is NaN 946057 1 oliver 2013-11-01 15:28:54 -0700 Try changing the test to +/-8 -> +/-2 946217 2 215808 zan 2013-11-02 08:20:07 -0700 Created attachment 215808 Patch 946489 3 215808 ossy 2013-11-03 13:56:03 -0800 Comment on attachment 215808 Patch LGTM 946494 4 215808 commit-queue 2013-11-03 14:24:57 -0800 Comment on attachment 215808 Patch Clearing flags on attachment: 215808 Committed r158545: <http://trac.webkit.org/changeset/158545> 946496 5 commit-queue 2013-11-03 14:25:00 -0800 All reviewed patches have been landed. Closing bug. 946518 6 darin 2013-11-03 17:02:06 -0800 This patch just stops testing for this bug, it doesn’t fix the bug. Do we want to fix the bug? 946711 7 oliver 2013-11-04 10:01:27 -0800 (In reply to comment #6) > This patch just stops testing for this bug, it doesn’t fix the bug. Do we want to fix the bug? The bug is lack of precision in the linux/gtk lib math, it can't be fixed in JSC short of us reimplementing cbrt entirely 215808 2013-11-02 08:20:07 -0700 2013-11-03 14:24:57 -0800 Patch bug-123624-20131102082007.patch text/plain 1934 zan SW5kZXg6IExheW91dFRlc3RzL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBMYXlvdXRUZXN0cy9D aGFuZ2VMb2cJKHJldmlzaW9uIDE1ODQ4NykKKysrIExheW91dFRlc3RzL0NoYW5nZUxvZwkod29y a2luZyBjb3B5KQpAQCAtMSwzICsxLDE2IEBACisyMDEzLTExLTAyICBaYW4gRG9iZXJzZWsgIDx6 ZG9iZXJzZWtAaWdhbGlhLmNvbT4KKworICAgICAgICBQcmVjaXNpb24gZXJyb3JzIGluIE1hdGgu Y2JydCgpIG9uIEdUSywgRUZMCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3df YnVnLmNnaT9pZD0xMjM2MjQKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4K KworICAgICAgICBDYWxjdWxhdGUgY3ViZSByb290cyBvZiA4IGFuZCAtOCwgZXhwZWN0aW5nIDIg YW5kIC0yLCByZXNwZWN0aXZlbHkuIEN1YmUgcm9vdHMgb2YgdGhvc2UgdHdvCisgICAgICAgIG51 bWJlcnMgZG9uJ3QgY29udGFpbiBwcmVjaXNpb24gZXJyb3JzIG9uIHRoZSBHVEsgYW5kIEVGTCBw b3J0cyBhcyBpcyB0aGUgY2FzZSB3aXRoIDI3IGFuZCAtMjcuCisKKyAgICAgICAgKiBqcy9tYXRo LWV4cGVjdGVkLnR4dDoKKyAgICAgICAgKiBqcy9zY3JpcHQtdGVzdHMvbWF0aC5qczoKKwogMjAx My0xMS0wMSAgQWxleGV5IFByb3NrdXJ5YWtvdiAgPGFwQGFwcGxlLmNvbT4KIAogICAgICAgICBB ZGQgV2ViQ3J5cHRvIEFFUy1DQkMKSW5kZXg6IExheW91dFRlc3RzL2pzL21hdGgtZXhwZWN0ZWQu dHh0Cj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT0KLS0tIExheW91dFRlc3RzL2pzL21hdGgtZXhwZWN0ZWQudHh0CShyZXZp c2lvbiAxNTg0ODUpCisrKyBMYXlvdXRUZXN0cy9qcy9tYXRoLWV4cGVjdGVkLnR4dAkod29ya2lu ZyBjb3B5KQpAQCAtMTg0LDggKzE4NCw4IEBAIFBBU1MgTWF0aC5hdGFuaCgtSW5maW5pdHkpIGlz IE5hTgogUEFTUyBNYXRoLmNicnQoTmFOKSBpcyBOYU4KIFBBU1MgTWF0aC5jYnJ0KDApIGlzIDAK IFBBU1MgTWF0aC5jYnJ0KC0wKSBpcyAtMAotUEFTUyBNYXRoLmNicnQoMjcpIGlzIDMKLVBBU1Mg TWF0aC5jYnJ0KC0yNykgaXMgLTMKK1BBU1MgTWF0aC5jYnJ0KDgpIGlzIDIKK1BBU1MgTWF0aC5j YnJ0KC04KSBpcyAtMgogUEFTUyBNYXRoLmNicnQoSW5maW5pdHkpIGlzIEluZmluaXR5CiBQQVNT IE1hdGguY2JydCgtSW5maW5pdHkpIGlzIC1JbmZpbml0eQogUEFTUyBNYXRoLmNvc2goTmFOKSBp cyBOYU4KSW5kZXg6IExheW91dFRlc3RzL2pzL3NjcmlwdC10ZXN0cy9tYXRoLmpzCj09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT0KLS0tIExheW91dFRlc3RzL2pzL3NjcmlwdC10ZXN0cy9tYXRoLmpzCShyZXZpc2lvbiAxNTg0 ODUpCisrKyBMYXlvdXRUZXN0cy9qcy9zY3JpcHQtdGVzdHMvbWF0aC5qcwkod29ya2luZyBjb3B5 KQpAQCAtMjYxLDggKzI2MSw4IEBAIHNob3VsZEJlKCJNYXRoLmF0YW5oKC1JbmZpbml0eSkiLCAi TmFOIikKIHNob3VsZEJlKCJNYXRoLmNicnQoTmFOKSIsICJOYU4iKTsKIHNob3VsZEJlKCJNYXRo LmNicnQoMCkiLCAiMCIpOwogc2hvdWxkQmUoIk1hdGguY2JydCgtMCkiLCAiLTAiKTsKLXNob3Vs ZEJlKCJNYXRoLmNicnQoMjcpIiwgIjMiKTsKLXNob3VsZEJlKCJNYXRoLmNicnQoLTI3KSIsICIt MyIpOworc2hvdWxkQmUoIk1hdGguY2JydCg4KSIsICIyIik7CitzaG91bGRCZSgiTWF0aC5jYnJ0 KC04KSIsICItMiIpOwogc2hvdWxkQmUoIk1hdGguY2JydChJbmZpbml0eSkiLCAiSW5maW5pdHki KTsKIHNob3VsZEJlKCJNYXRoLmNicnQoLUluZmluaXR5KSIsICItSW5maW5pdHkiKTsKIAo=