Rebooting Flash 17 times causes error # 2046 and requires a browser restart

Question

I encounter very strange behavior with the Flex 4.1 application that I am writing that interferes with testing. It seems that I can restart the application 16 times, and then on the 17th, the download process failed with

Error #2046: The loaded file did not have a valid signature 

This seems to be happening on the 17th reboot on both Firefox 5.0 and Chrome 12. I'm not sure if this is relevant, but I am running Flash Player v10.2.159.1 (also happening from 10.3.181.34) on Ubuntu 10 , 04. It happens with both regular and debuggers. When I run the application in Windows FF5, it looks like this is not happening. Closing the current browser window does not seem to fix it. The only way is to completely close all browser windows and restart the browser. And again after 16 successful loads, the 17th failure.

At this point, I am thinking of flagging it as a Linux Flash error, but I would like to make sure and check if anyone knows if there is anything that I should have done to prevent this.

The user from this post seems to have had the same problem, but I think he didn’t notice the pattern I have.

Any help would be greatly appreciated. Ruy

== UPDATE ==

I just realized that after my application starts throwing error 2046, an attempt to load any other Flash that uses signed RSL also shows error 2046 (for example, this application ), which means that the problem is not specific to my application and, most likely related to flash cache or something like that.

Answer 1

Disclosure: I am a developer of Flash Player in Adobe. This is unlikely to attract much attention, since it is only Linux and something like an edge: it may be annoying while dev is running, but very few users reload the same page more than 16 times. This may be a problem with the browser. But this is probably us :) I will look at the jira tomorrow and see if I can raise it a little, but I will be honest that this is really an extreme case and is unlikely to receive much love. If you want to increase your chances, be sure to add the simplest test example .swf, which you can add to the error. Also, please double check to see if everything is happening with the latest beta.
I also just looked at earlier error messages and forum posts, probably you should post this as a Flash Player error, not Flex.

Answer 2

Guess the long shot, but it looks like the problem we had .... in the project properties - Flex Build Path - Framework Linkage - change to "merged into code". This fixes a problem very similar to what you are describing, although I am sorry that I do not know exactly what the reason is. Good luck

Answer 3

tl; dr: I don’t know the reason, sending a random opportunity in the hope that this might give someone else an idea or two for testing.

Given that this is apparently an unresolved error in Adobe's bug tracking program, it is unlikely that you will get a final answer here. Considering that this happens both in Firefox and in Chrome, allow to exclude browser errors and assume that this is either some kind of shared library (Flash) or OS API (Linux kernel implementation). The comment in one of the questions about jira specifically mentions the killing of the Flash process, fixes it, so its a problem with Flash, not an OS error.

The most interesting thing that I see here is your observation that it manages exactly 16 times before it boots. Time for some specs here, from someone who has never worked on a kernel or cryptocurrency:

With a 2048-bit RSA key and 32k cache for storing them, 16 keys will correspond before another one crashes, so one hypothesis is that every time this file is downloaded, Flash caches a signed value (possibly a hashed version) for some reason - maybe track allowed and used security permissions, etc.? If this record is not deleted, then once it is filled, all file downloads will fail if the signature caching is part of the verification.

Things you can experiment with:

• Reduce the size of the application to see if it is possible to reload the page more often (as suggested by stackfish)
• Count the number of signed RSLs and if it has a capacity / multiple of 2 (maybe others get an error after loading 32 pages if they use half the number of signed libs?)
• Make sure the Linux Flash plugin has some ability to increase the credential cache or something (or decrease it, just to see if it affects the number of downloads - if so, this may be due to a problem).

I expect that to actually find a solution, you will need to dive into the library loading code and look at all the constants associated with downloading signed libs that are 4, 16 or several of 16 to see if they can be responsible - in short saying it is unlikely to be allowed by other people outside the Flash dev imho command: /

Answer 4

This behavior may be due to a memory leak caused by either a Flex implementation or a browser plugin. Firefox, as you know, does not clear the memory in any case, and the trace will continue to grow, the longer you open the browser window.

If you reduce the size of your flex application to create something very small, the number of times you can reload the page increases?

Answer 5

Error # 2046 on win vista, a 64-bit machine with 1000 MB ati radeon videocard, the problem occurs only in msn video sofar