Does anyone have any suggestions?
maybe cross domain policy
Use a network monitor to compare traffic between the working and non-working
Relying on getStackTrace(), trace(), System.gc() and other debugging APIs is
certainly going to cause trouble.
If you are using metadata, make sure you keep the metadata you need. By
default, almost everything is stripped from a release build.
Flex harUI, thanks for the response. We trace() a lot in our application, so I'll try disabling all our calls to it. I didn't realize that it could actually cause problems, I thought trace() is safely ignored when running the application via a non-debug version of Flash. Is there a webpage that contains all the debugging API calls that I should not be using?
Incidentally, I did use wireshark to inspect traffic over the wire and found that the application's doesn't get around to sending anything at all.
The problem with using trace is that I'm not sure what happens if someone
does something like
That should return the child removed in the debugger, but the entire call
might get deleted in release.
I don't know of a page that lists all debugger-only calls. There aren't
that many, but flash.sampler package is another set of APIs that are
If there is no network traffic, are you putting the release SWF in the exact
same URL as the debug SWF? Otherwise there could be issues with the
configuration being hardwired to something.
Had an issue where my Flash application was failing (crashing) in the production player, and working fine in the debugger version of the player. Since my experience was different, thought I would add a comment.
I assumed that the issue was related to using some sort of Debugger Player specific calls, like mentioned above. However, this was not the case.
It was a 'while' function, that apparently died silently in the Debugger Player, yet failed, froze, and crashed the production player. The only way I found this sort of phantom bug, was to check the net calls in the browser. Using the last net action made by the player directly before it crashed, I worked forward from that point in the code where that net action took place. Literally just commenting out different things in the workflow until I found it.
Not sure if this could possibly help anyone, but figured I'd share anyway.