Last modified: 2014-01-05 05:02:46 UTC
Broken out from bug 34458: (In reply to bug 34458 comment #8) > as of Feb 21, both logged-in and not-logged-in users may edit on test2 using > IE7. > > however, IE7 still shows the "Access denied" error for "Line 51" upon loading > any page. This is fairly annoying, especially for any users who may have > "Always display... errors" checked in the browser. Line: 51 Char: 3 Error: Access is denied. Code: 0 URL: http://test2.wikipedia.org/wiki/Test_page Line: 9 Char: 520 Error: Could not complete the operation due to error 80020101. Code: 0 URL: http://test2.wikipedia.org/wiki/Main_Page I see these in IE7, not in IE8. I See it on test2 but also on en.wikipedia.org, so this isn't 1.19-specific.
which makes sense. Road identified Line 51: RoanKattouw: Line 51: [4:46pm] RoanKattouw: </style><!--[if lt IE 7]><style type="text/css">body{behavior:url("/w/skins-1.19/vector/csshover.min.htc")}</style><![endif]--></head>
That should devolve to nothing on IE 7 though, since the comment says it's only for less than IE 7...?
I suspect the problem is actually the line immediately before that one. On test2.wikipedia.org that's where CentralNotice's banner controller loads. On non-updated wikis, it loads at the bottom of the page. The loading hook is BeforePageDisplay in both cases (which it's been using for years). Did someone recently change (fix?) the behavior of that hook? We might just need to switch the hook used in CentralNotice from BeforePageDisplay to SkinAfterBottomScripts, as it really should load at the bottom. (Although I'm not sure why loading at the top would ever cause an error.)
I just tried running CentralNotice inside 1.19wmf1 locally and it works as expected (loads from the bottom). Is it possible test2.wikipedia.org is running some weird version of CentralNotice? (or has some strange hook configuration?)
It should be running the same centralnotice version as whats in the 1.19wmf1 branch.
I'm running the 1.19wmf1 branch version of CentralNotice and it's behaving differently that the one on test2 (loading at the bottom rather than from the head). Whatever the problem is, it appears to be specific to test2.wikipedia.org. Revising summary and setting priority as low...
It looks like test2 has wgResourceLoaderExperimentalAsyncLoading turned on. I suspect this is what is causing the problem.
