InformAction Forums

Posted: **Wed Jun 23, 2010 9:29 am**

Giorgio Maone wrote:
al_9x wrote:Did you repro or were you able to guess what it was?
I couldn't reproduce, but I based my (quite complex) fix on the assumption you were getting a totally random order, rather than DOM-parsing based, in nsIContentPolicy calls (something I could never observe, but I guessed could be a byproduct of your system's slowness).

if you look at the bad dump, the following sequence repeats for each url, in descending order (4 - 1), same as the good dump:

Code: Select all

[NoScript] Content processing -- type: 5, location: http://url4.invalid/, origin: file:///C:/tmp/embedding.htm, ctx: <HTML Element>, mime: application/x-shockwave-flash, null
----------
[NoScript] Error()@:0
(5,[object XPCWrappedNative_NoHelper],[object XPCWrappedNative_NoHelper],[object XPCNativeWrapper],"application/x-shockwave-flash",null)@chrome://noscript/content/Policy.js:159

----------
[NoScript] [CP PASS 2] application/x-shockwave-flash*http://url4.invalid/, 5, null
----------
[NoScript] tagForReplacement
----------
[NoScript] Content BLOCKED Forbidden Content -- type: 5, location: http://url4.invalid/, origin: file:///C:/tmp/embedding.htm, ctx: <HTML Element>, mime: application/x-shockwave-flash, null
----------
[NoScript] Error()@:0
("Forbidden Content",[object Object])@chrome://noscript/content/Main.js:1862
(5,[object XPCWrappedNative_NoHelper],[object XPCWrappedNative_NoHelper],[object XPCNativeWrapper],"application/x-shockwave-flash",null)@chrome://noscript/content/Policy.js:619

if it were random, wouldn't the dump show it?

the difference starts on the line after the above sequences

good:

Code: Select all

[NoScript] Adding plugin sites: ["http://url4.invalid", "http://url3.invalid", "http://url2.invalid", "http://url1.invalid"] to ["file://"]

bad:

Code: Select all

[NoScript] Adding plugin sites: ["http://url4.invalid", "http://url3.invalid"] to ["file://"]

Posted: **Wed Jun 23, 2010 10:08 am**

al_9x wrote:if it were random, wouldn't the dump show it?

Probably yes, but since there are at least a couple other delayed code paths after blocking and before placeholders, the random order might have popped up elsewhere on your slow system.
If you make a diff, you'll see the changes to make all this more resilient are pretty extensive, so even though I can't tell for sure where your problem was exactly, I'm not surprised it's fixed anyway.

InformAction Forums

[WONTFIX] blocked object duplication

Re: [FIXED] blocked object duplication

Re: [WONTFIX] blocked object duplication