[MutationObservers] Microbenchmarks for appendChild, removeChild, and innerHTML

[Adam Klein]

[MutationObservers] Microbenchmarks for appendChild, removeChild, and innerHTML

[Adam Klein]

Created attachment 114324 [details]
Patch

[Adam Klein]

Created attachment 114330 [details]
Bumped depth to 200

[Adam Klein]

Created attachment 114335 [details]
Remove braces

[Ojan Vafai]

Comment on attachment 114335 [details]
Remove braces

View in context: https://bugs.webkit.org/attachment.cgi?id=114335&action=review

Looks fine. I have some personal stylistic nits that would make the code seem more natural to me, but you can ignore them. I only listed them for the first tests, but obviously the same nits apply to all the tests.

> PerformanceTests/Mutation/append-child-deep.html:23
> +    window.start = null;
> +    window.numRuns = 25;
> +    window.times = [];

Using "window." here seems weird to me. Just use start/numRuns/times. It'll do the same thing right?

> PerformanceTests/Mutation/append-child-deep.html:32
> +        var time = new Date().getTime() - start;

Nit: Use Date.now()

> PerformanceTests/Mutation/append-child-deep.html:43
> +        start = new Date().getTime();

ditto

[Adam Klein]

Comment on attachment 114335 [details]
Remove braces

View in context: https://bugs.webkit.org/attachment.cgi?id=114335&action=review

>> PerformanceTests/Mutation/append-child-deep.html:23
>> +    window.times = [];
> 
> Using "window." here seems weird to me. Just use start/numRuns/times. It'll do the same thing right?

Just a clash of stylistic preferences, I guess: I wanted to avoid declaring these variables, but still wanted to make it clear that I _knew_ they were being set on the global scope.

>> PerformanceTests/Mutation/append-child-deep.html:32
>> +        var time = new Date().getTime() - start;
> 
> Nit: Use Date.now()

Ah, yes, nicer.

[Adam Klein]

Committed r99879: <http://trac.webkit.org/changeset/99879>