Bug 16366
| Summary: | "Unsafe JavaScript" spew to console takes up CPU constantly while page loaded | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Mark Rowe (bdash) <mrowe> |
| Component: | WebCore JavaScript | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED WONTFIX | ||
| Severity: | Major | CC: | abarth |
| Priority: | P2 | Keywords: | InRadar |
| Version: | 528+ (Nightly build) | ||
| Hardware: | Mac | ||
| OS: | OS X 10.5 | ||
| URL: | http://sovietcomputing.com/node/2 | ||
Mark Rowe (bdash)
After loading the URL, http://sovietcomputing.com/node/2, in Safari 3 or TOT warnings similar to the following are written to the JavaScript console and system console repeatedly:
Unsafe JavaScript attempt to access frame with URL http://gmodules.com/ig/ifr?url=http://www.hitsalive.com/google_gadgets/russian_wotd.xml&up_promo22=1&up_promo6=1&up_numuses=0&synd=open&w=120&h=200&title=Russian+-+Audio+Word+of+the+Day&border=%23ffffff%7C0px%2C1px+solid+%23993333%7C0px%2C1px+solid+%23bb5555%7C0px%2C1px+solid+%23DD7777%7C0px%2C2px+solid+%23EE8888&source=http%3A%2F%2Fsovietcomputing.com%2Frobots.txt%23resize_iframe%2526remote_iframe_0%2526140%24#idi_hr=http%3A%2F%2Fsovietcomputing.com%2Frobots.txt from frame with URL http://sovietcomputing.com/robots.txt#resize_iframe%26remote_iframe_0%26140$. Domains, protocols and ports must match.
The warnings continue to be displayed many times per second until the page is closed. Safari's CPU usage spikes dramatically over that which is typical when viewing a page.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Mark Rowe (bdash)
<rdar://problem/5638289>
Geoffrey Garen
How do you know that the warnings are the source of the CPU usage? It seems to me that CPU is consumed by the script that is executing.
Mark Rowe (bdash)
I don't. I suspect that the warnings are merely a symptom of whatever is also causing the CPU usage. In other browsers I do not see the CPU usage nor the constant warnings.
Adam Barth
This site no longer exists. I think there is (or used to be) a real bug here, but there's not enough information in this report to track it down. :(