I'm running into an intermittent problem in my selenium tests. It appears that respondToReadyState is invoked with readyState of 4, but this.transport.readyState is still 1. This results in a call to success(), which in turns calls this.transport.status which throws an exception with NS_ERROR_NOT_AVAILABLE. With help of firebug, I can see that readyState and this.transport.readyState are different quite often in respondToREadyState. So, I assume that's the desired behavior. However, it seems like there is a race condition that in certain cases causes this.transport.status to fail. I'm using Prototype 22.214.171.124 and Firefox 126.96.36.199. I can provide additional debug info if someone can help.
on 2008-06-25 05:46
on 2008-06-25 19:26
LOL ... we actually run into the same problem as we reported before (see http://prototype.lighthouseapp.com/projects/8886/t..., which is fixed for 188.8.131.52). However, we accidentally upgraded prototype which removed our patch. No one realized we are running into the same exact issue since we assumed we already had the patch :).
on 2008-06-25 19:33
It's better if you can "monkey patch" if possible - load an external script that patches Prototype with your fix after it's been loaded. Not that a Prototype upgrade will never cause your patch to break, but in many cases, it won't, and in the case of a big upgrade you'll probably end up having to tweak your patch anyway. -Fred On Wed, Jun 25, 2008 at 12:26 PM, Paul <email@example.com> wrote: > > LOL ... we actually run into the same problem as we reported before > (see > http://prototype.lighthouseapp.com/projects/8886/t... > , > which is fixed for 184.108.40.206). However, we accidentally upgraded > prototype which removed our patch. No one realized we are running into > the same exact issue since we assumed we already had the patch :). -- Science answers questions; philosophy questions answers.