Per our IRC conversation - from time to time BrowserStack has an out-of-date Java which causes the test run to hang, causing failures for the whole module that uses it. It seems it's just the jQuery.acceptData test:
|
test("jQuery.acceptData", function() { |
|
expect( 11 ); |
|
|
|
var flash, applet; |
|
|
|
ok( jQuery.acceptData( document ), "document" ); |
|
ok( jQuery.acceptData( document.documentElement ), "documentElement" ); |
|
ok( jQuery.acceptData( {} ), "object" ); |
|
ok( !jQuery.acceptData( document.createElement( "embed" ) ), "embed" ); |
|
ok( !jQuery.acceptData( document.createElement( "applet" ) ), "applet" ); |
|
|
|
flash = document.createElement( "object" ); |
|
flash.setAttribute( "classid", "clsid:D27CDB6E-AE6D-11cf-96B8-444553540000" ); |
|
ok( jQuery.acceptData( flash ), "flash" ); |
|
|
|
applet = document.createElement( "object" ); |
|
applet.setAttribute( "classid", "clsid:8AD9C840-044E-11D1-B3E9-00805F499D93" ); |
|
ok( !jQuery.acceptData( applet ), "applet" ); |
|
|
|
ok( !jQuery.acceptData( document.createComment( "" ) ), "comment" ); |
|
ok( !jQuery.acceptData( document.createTextNode( "" ) ), "text" ); |
|
ok( !jQuery.acceptData( document.createDocumentFragment() ), "documentFragment" ); |
|
|
|
ok( jQuery.acceptData( |
|
jQuery( "#form" ).append( "<input id='nodeType'/><input id='nodeName'/>" )[ 0 ] ), |
|
"form with aliased DOM properties" ); |
|
}); |
This seems to matter for IE8 only, the master branch doesn't have any special logic for classid handling. It seems there's a hight chance someone from the team will run the test suite on IE8 manually at least once in a while which would catch potential problems; we should disable this test just on BrowserStack then.
Per our IRC conversation - from time to time BrowserStack has an out-of-date Java which causes the test run to hang, causing failures for the whole module that uses it. It seems it's just the
jQuery.acceptDatatest:jquery/test/unit/data.js
Lines 136 to 162 in d39fef8
This seems to matter for IE8 only, the
masterbranch doesn't have any special logic forclassidhandling. It seems there's a hight chance someone from the team will run the test suite on IE8 manually at least once in a while which would catch potential problems; we should disable this test just on BrowserStack then.