[jira] Commented: (MYFACES-2965) MyFaces 2.0.2 AJAX crashes Safari on iPad

["Jan-Kees van Andel (JIRA)" <de...@myfaces.apache.org>]

    [ https://issues.apache.org/jira/browse/MYFACES-2965?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12929360#action_12929360 ] 

Jan-Kees van Andel commented on MYFACES-2965:
---------------------------------------------

Werner has been debugging his AJAX code. This is the result:

ah main difference between android and apple
both of them run their own javascript engine
they are not the same
it is the exists check for send as binary
it runs now
gotta send that bug to apple
all i do is 'undefined' == typeof xhr.sendAsBinary || null == xhr.sendAsBinary
apparently that is enough to crash the xhr object of safari as soon as send is called
this is a stupid bug on apples side
either way I will add mobile safari detection and remove that code for mobile safari
mobile safari should be sendAsBinary enabled anyway

> MyFaces 2.0.2 AJAX crashes Safari on iPad
> -----------------------------------------
>
>                 Key: MYFACES-2965
>                 URL: https://issues.apache.org/jira/browse/MYFACES-2965
>             Project: MyFaces Core
>          Issue Type: Bug
>          Components: JSR-314
>    Affects Versions: 2.0.2
>         Environment: iPad Safari
>            Reporter: Jan-Kees van Andel
>            Assignee: Werner Punz
>
> If you go to http://ipad.parleys.com/parleys-frontend-ipad-1.0/ on your iPad, and then click on the button "Latest", "Top Rated" or some other button that invokes an AJAX request, the browser crashes and goes back to the desktop.
> A simple debugging session pointed me to the jsf.ajax.request() invocation, but that's still pretty broad. Reverting to 2.0.1 solved the issue.
> The code can be viewed online at: http://code.google.com/p/parleys-html5/source/browse/#svn/trunk/web-ipad/src/main/webapp
> Talked offline with Werner about the issue and we think it's somewhere in the 2.0.2 optimizations.
> Werner's first guess: "If you want to check yourself you can set an alert in replacenode in Dom.js and see if the code is reached
> And see if it finishes.
> I am 100 percent sure the problem is there.
> But it is just sixth sense on my part.
> So just a guessing."

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.