When opening the nikon website a strange painting error occurs with the safari window. Not even sure how to describe it.
Confirming bug. This is not a regression--it happens on Safari 2.0.3 (417.8) on 10.4.4 as well. I would describe the behavior as "bouncing", although the entire window contents bounces inside its boundary, not just the HTML content area. FREAKY!
I'll reduce this one... Assigning to myself :P
Created attachment 6281 [details] Testcase This testcase shows issues with how we handle setTimeout, don't know if this is a bug. This occcurs repeatedly in the code for the site: if ( ( ismc ) && ( document.all || issafari ) ) { clearTimeout(macTimeout1); clearTimeout(macTimeout2); macTimeout1 = setTimeout("window.resizeBy(1,1);",100); macTimeout2 = setTimeout("window.resizeBy(-1,-1);",200); } this causes the "shake". The testcase shows this as well and shows that we handle timing differently then Firefox.
As the reduction shows, this is not a painting issue, the scripts do multiple resizes on purpose and just for Safari. If i use User-Agent Mozilla 1.1, this does not occur and the site works fine. So these guys need to be evangelized, CC-ing Mark Malone for that. The other issue here is that the timers fire in the wrong order it seems. Renaming, attaching better testcase with larger timeouts in a sec.
Created attachment 6282 [details] Testcase with larger time outs 1000 instead of 100 and 2000 instead of 200 ms.
Created attachment 6283 [details] Better testcase Now correctly.
Ok, this bug will be on resizing the window. The upper left corner isn't anchored and resizing in general looks weird, flashy and slow and not clean like f/i Firefox. This bug will track that. Timing issues, if any, will be tracked in another bug. Reassigning back to webkit-unassigned.
The timing issues are filed in bug 7099.
*** Bug 9285 has been marked as a duplicate of this bug. ***
This actually looks like a Safari issue: the calculations that WebKit performs before asking Safari to resize the window seem to be correct. Also, Safari is apparently setting window minimum size as a result of this test - manual resizing no longer works correctly.
<rdar://problem/5665174>
*** Bug 31033 has been marked as a duplicate of this bug. ***
*** Bug 68419 has been marked as a duplicate of this bug. ***
Created attachment 120996 [details] A test-case exaggerating the behavior. Attached another test case which exaggerate the behavior: Resize the browser window (making it smaller) using window.resizeTo() will cause window's title to be clipped. To reproduce 1. Open the attached test-case in a new browser window (not in a tab, because window.resizeTo() doesn't work in a tab). 2. The test-case will automatically resize the browser for 5 times with an interval of 1-2 seconds. 3. Observes the window's title being clipped during each resize.
Is this still a valid bug? I can't recreate it at all but I'm not sure that I'm doing it right!
@Alan - I think browsers are now evolved and we have hardware which can manage these resizing issues. Plus I am not able to reproduce this bug, can we mark this as 'RESOLVED CONFIGURATION CHANGED' or 'RESOLVED WONTFIX'?
The attached test case looks performant to me.