Summary: | REGRESSION: Browser window shrinks while visiting aljazeera.net story | ||
---|---|---|---|
Product: | WebKit | Reporter: | Ahmed Al-Saadi <ahmed.alsaadi> |
Component: | New Bugs | Assignee: | Nobody <webkit-unassigned> |
Status: | RESOLVED FIXED | ||
Severity: | Normal | CC: | adele, chris.messina, ddkilzer, ismail, mrowe |
Priority: | P1 | Keywords: | HasReduction, InRadar, Regression |
Version: | 523.x (Safari 3) | ||
Hardware: | Mac | ||
OS: | OS X 10.4 | ||
URL: | http://www.aljazeera.net/NR/exeres/BA8EC33F-EC9B-44F1-A51F-CD7F4C079BEF.htm |
Description
Ahmed Al-Saadi
2007-09-20 22:27:06 PDT
Confirmed with a local debug build of WebKit r25670 with Safari 3 Public Beta v. 3.0.3 (522.12.1) on Mac OS X 10.4.10 (8R218). I'm also going to call this a regression since loading the same page with Safari 2.0.4 (419.3) with original WebKit on Mac OS X 10.4.10 (8R218) does not exhibit this behavior. (Of course there are more errors in the JavaScript Console, but it still doesn't make the window tiny or cause tabs to behave strangely.) The symptoms suggest to me that this may be an issue in Safari rather than WebKit. A malfunctioning tab bar, for instance, would seem to have little to do with WebKit. It looks like there are at least two issues here. Clearly Safari should not be getting itself into a state where the tab bar is misrendering. I'll write up a separate radar to cover that issue shortly. The second is that when the page loads, an iframe at <http://jazad.aljazeera.net/jazcommerce/default.aspx?sThisPage=_1sis_reportsandanalysis_&BannerType=SponsorCMS> is loaded. This calls through to a function named unresizeFrame which is defined in <http://jazad.aljazeera.net/jazcommerce/CommerceFunctions.js> as: function unresizeFrame(){ if(window.top.sponsorAd) window.top.sponsorAd.resizeTo(0,0) } It is this resize that is leading to problem. Reduction is at http://bdash.net.nz/files/iframe-resizeTo/. DOMWindow.resizeTo appears to always affect the outermost frame. In IE 7, the reduction results in the red iframe being resized to 10x10. In Safari it currently attempts to resize the browser window. window.top.myIframe is the DOMWindow associated with the iframe in each case. I also discovered this problem on a malware site: http://iexplorer-security.org/?id=987650077 (clicking this link will resize the browser, popup a Javascript prompt [1], and if you click anywhere on the resulting page [2], a .exe file will be downloaded). However, if you resize the shrunken window, the tabs are clipped: http://www.flickr.com/photos/factoryjoe/2521951566/ [1] http://www.flickr.com/photos/factoryjoe/2521108253/in/photostream/ [2] http://www.flickr.com/photos/factoryjoe/2521113925/in/photostream/ (In reply to comment #5) > Reduction is at http://bdash.net.nz/files/iframe-resizeTo/. Mark, do you still have this reduction sitting around anywhere? The URL is now invalid. (In reply to comment #8) > (In reply to comment #5) > > Reduction is at http://bdash.net.nz/files/iframe-resizeTo/. > > Mark, do you still have this reduction sitting around anywhere? The URL is now > invalid. It’s there again now. (In reply to comment #9) > (In reply to comment #8) > > (In reply to comment #5) > > > Reduction is at http://bdash.net.nz/files/iframe-resizeTo/. > > > > Mark, do you still have this reduction sitting around anywhere? The URL is now > > invalid. > > It’s there again now. Thanks! The bisect-builds script says this was fixed (progressed) between: Fails: r36519 Works: r36635 (In reply to comment #10) > The bisect-builds script says this was fixed (progressed) between: > > Fails: r36519 Works: r36635 Revision r36560 looks awfully suspicious: <http://trac.webkit.org/changeset/36560> |