[Webkit-unassigned] [Bug 132229] Coalesce responses on network process side

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=132229





--- Comment #14 from Antti Koivisto <koivisto at iki.fi>  2014-04-29 12:49:39 PST ---
> This is a missed optimization - a place where we could send a coalesced response, but don't.

Right, this path is an existing optimization that coalesces didReceiveData and didFinishLoading for disk cache hit (along with passing an mmapped buffer). It could be improved further but that is bit of a separate topic.

> This is not sufficient. You are adding new code, and we don't have to copy old mistakes unless there is a reason to.
> 
> Using signed values for sizes causes all sorts of bugs.

I disagree. Consistency is more important. We should fix this everywhere as a separate change. Conversions cause bugs.

> Do other functions make multiple calls to client? This new function is very different from others, as it's multi-step.

Yes, didReceiveResource does (which is a similar existing optimization, see above).

> I don't understand this answer. There can be multiple messages, but each gets a lot of additional checks before it reaches WebResourceLoader, they don't share the same stack when executed.

Maybe you could suggest concrete improvements to the patch as I'm not sure what you are after here.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.