REGRESSION(r183570): jslib-traverse-jquery is 22% slower

• Component: New Bugs
• Status: RESOLVED FIXED

[Geoffrey Garen]

REGRESSION(r183570): jslib-traverse-jquery is 22% slower

[Geoffrey Garen]

Created attachment 252100 [details]
Patch

[Geoffrey Garen]

Comment on attachment 252100 [details]
Patch

Looks like I broke everything :(.

[Geoffrey Garen]

Created attachment 252112 [details]
Patch

[Mark Lam]

Comment on attachment 252112 [details]
Patch

r=me too.

[Andreas Kling]

Comment on attachment 252112 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252112&action=review

> Source/JavaScriptCore/ChangeLog:13
> +            (b) the nodes are usually already sorted.

I think they will always be sorted, originating from querySelector APIs which always return nodes in document order. :)

> Source/JavaScriptCore/ChangeLog:22
> +        82% speedup on jslib-traverse-jquery, bringing us to 40% faster than
> +        the unregressed baseline.

Kickass!

[Geoffrey Garen]

Comment on attachment 252112 [details]
Patch

Still broken :(.

BTW, the nodes in the worst tests are never in order or reverse order. I don't know why yet.

[Geoffrey Garen]

Created attachment 252162 [details]
Patch

[Build Bot]

Comment on attachment 252162 [details]
Patch

Attachment 252162 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/5776924949348352

New failing tests:
fast/dom/navigator-detached-no-crash.html
js/sort-large-array.html
transitions/transition-end-event-multiple-03.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.1_T3.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.2_T1.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.2_T3.html
fast/xmlhttprequest/xmlhttprequest-get.xhtml
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.1_T1.html
http/tests/workers/worker-importScriptsOnError.html
js/dom/regexp-caching.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A3_T2.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A3_T1.html
js/reserved-words-strict.html
fast/dom/plugin-attributes-enumeration.html
js/dom/global-constructors-attributes-dedicated-worker.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.1_T2.html
fast/dom/event-handler-attributes.html
js/comparefn-sort-stability.html
js/dom/global-constructors-attributes.html
sputnik/Conformance/13_Function_Definition/13.2_Creating_Function_Objects/S13.2.1_A5_T1.html
jquery/traversing.html
js/Object-getOwnPropertyNames.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.2_T2.html

[Build Bot]

Created attachment 252165 [details]
Archive of layout-test-results from ews104 for mac-mavericks-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews104  Port: mac-mavericks-wk2  Platform: Mac OS X 10.9.5

[Build Bot]

Comment on attachment 252162 [details]
Patch

Attachment 252162 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/6060547409707008

New failing tests:
fast/dom/navigator-detached-no-crash.html
js/sort-large-array.html
transitions/transition-end-event-multiple-03.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.1_T3.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.2_T1.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.2_T3.html
fast/xmlhttprequest/xmlhttprequest-get.xhtml
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.1_T1.html
http/tests/workers/worker-importScriptsOnError.html
js/dom/regexp-caching.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A3_T2.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A3_T1.html
js/reserved-words-strict.html
fast/dom/plugin-attributes-enumeration.html
js/dom/global-constructors-attributes-dedicated-worker.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.1_T2.html
fast/dom/event-handler-attributes.html
js/comparefn-sort-stability.html
js/dom/global-constructors-attributes.html
sputnik/Conformance/13_Function_Definition/13.2_Creating_Function_Objects/S13.2.1_A5_T1.html
jquery/traversing.html
js/Object-getOwnPropertyNames.html
sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.11_Array_prototype_sort/S15.4.4.11_A2.2_T2.html

[Build Bot]

Created attachment 252166 [details]
Archive of layout-test-results from ews102 for mac-mavericks

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: ews102  Port: mac-mavericks  Platform: Mac OS X 10.9.5

[Geoffrey Garen]

Created attachment 252195 [details]
Patch

[Geoffrey Garen]

Created attachment 252250 [details]
Patch

[Andreas Kling]

Comment on attachment 252250 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252250&action=review

> Source/JavaScriptCore/ChangeLog:29
> +        The solution is simple enough: Call compareDocumentPosition(b, a) instead.

Amazing. r=me too.

> Source/JavaScriptCore/ChangeLog:48
> +        Another option is for elements to keep a compareDocumentPosition cache,
> +        like a node list cache, which allows you to determine the absolute
> +        position of a node using a hash lookup. I will leave this as an exercise
> +        for kling.

Or we can just wait for anttik to finish his legendary CoreNode refactoring where container nodes have arrays of children, and compareDocumentPosition can be implemented using simple pointer comparison for siblings. :]

[Geoffrey Garen]

Committed r183749: <http://trac.webkit.org/changeset/183749>

[Simon Fraser (smfr)]

Comment on attachment 252250 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=252250&action=review

> Source/JavaScriptCore/ChangeLog:40
> +        A fully principled soultion to this problem would probably do something

soultion