incubator-callback-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Grieve <agri...@chromium.org>
Subject Re: Performane in FileAPI
Date Sat, 22 Sep 2012 14:39:46 GMT
I'd love to see this benchmark checked into mobile-spec.

Probably we should create a sub-page for benchmarks, just like there is for
Automated Tests.


On Sat, Sep 22, 2012 at 10:20 AM, Simon MacDonald <simon.macdonald@gmail.com
> wrote:

> Here's my two cents. The current bridge in 2.0/2.1 is much slower than what
> we are now defaulting to in the upcoming 2.2 release. That change improves
> the FileReader performance greatly. However, switching to XHR speeds things
> up by an order of magnitude. Using my quick patch the FileReader passes all
> mobile spec test cases except the error conditions which need to be
> converted. It looks like a major win for android if we make the switch.
>
> Folks on other platforms should run your benchmark to see if there is a
> similar improvement and then run my patch through mobile spec to see of
> there are any failures.
>
> With regards to CI yes we really do need this but I don't think we can set
> it up in time to test this issue.
>
> Going back to hacking the camera now and I'll get off my soapbox.
>
> Simo
>
> On Friday, September 21, 2012, Daniel Kurka wrote:
>
> > Hello everyone,
> >
> > in several situations I got frustrated with the performance of the
> > phonegap File API. For some time we have used XHR as a replacement for
> > FileReader within phonegap apps.
> > I would really like to see the IO performance of phonegap to improve.
> > First step in that way might be to use XHR instead of the bridge.
> >
> > To get this started I talked with Brian at Phonegap Day EU and he decided
> > the best way to do this is to actual get some data first, this is what
> > CB-1478 is about.
> > So far it seems that we can make a major improvement on Android, BB.
> > (other still need testing though)
> >
> > As Filip already pointed out some kind of benchmark should go into CI, so
> > that we will not regress at any point. What could I do about that?
> > Simon already proposed a patch which might work, but I wanted to get some
> > feedback on this first (
> >
> https://github.com/macdonst/incubator-cordova-js/commit/4c0134b827386d028f53370b3f5006b222939e26
> > )
> >
> > So what do you guys think what the best way forward is?
> >
> > -Daniel
> >
> >
> >
> >
>
> --
> Simon Mac Donald
> http://hi.im/simonmacdonald
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message