Bug 103791 – [v8?] DOM traversal test became bimodal

• Status: RESOLVED
• Resolution: WONTFIX
• Component: DOM

[Ojan Vafai]

Range: http://trac.webkit.org/log/?verbose=on&rev=136017&stop_rev=136010

Could it possibly be http://trac.webkit.org/changeset/136015/ ?

http://build.chromium.org/f/chromium/perf/mac-release-10.6-webkit-latest/dom_perf/report.html?rev=170457&graph=DOMDivWalk&history=50

http://build.chromium.org/f/chromium/perf/mac-release-10.6-webkit-latest/dom_perf/report.html?rev=170457&graph=DOMDivWalk&history=50

http://build.chromium.org/f/chromium/perf/mac-release-10.6-webkit-latest/dromaeo_domcoretraverse/report.html?rev=170457&graph=dom_traverse_nextSibling&history=50

[Ojan Vafai]

Actually, it looks like it might be an actual regression.

http://build.chromium.org/f/chromium/perf/mac-release-10.6/dromaeo_domcoretraverse/report.html?history=90&rev=170138&graph=dom_traverse_previousSibling

[Adam Barth]

Ok.  Let's roll http://trac.webkit.org/changeset/136015/ out.  It adds an instruction into a host code path, which is why I made the original change.

[Adam Barth]

Rolled out in http://trac.webkit.org/changeset/136313.  Let's see if that heals the bots.

[Anders Carlsson]

V8 is gone.