Enigmatic Thought » FlexSDK

Bugquash badges fixed

Matt C — Tue, 29 Dec 2009 20:29:37 +0000

Nate has fixed the BugQuash badge display, so everyone who has contributed to the Flex SDK can show the world. I’ve added mine back to the sidebar and I need to hunt another bug to earn my orange belt. As thanks for harassing him (idling in a chat room and asking every so often) he added me to the Test Page.

Flex 3.5 finally released

Matt C — Tue, 15 Dec 2009 16:49:34 +0000

And that means the 3.4 showstopper, double responses to HTTPService have been fixed. You can get 3.5 direct from the source.. I don’t know what the deal is with the build date, or why they waited 3 months for a full 3.x release rather than a 3.4.1 patch release. Especially since the bug was fixed a week after 3.4 came out.

Scientific Notation Patch Accepted

Matt C — Wed, 18 Nov 2009 16:21:40 +0000

The patch I mentioned way back in Let’s patch: Fun with Scientific Notation! was finally accepted and is now live in the 3.x branch. This fixes SDK-16241, which dealt with scientific notation.

I removed the Bug Quash badge since it’s misreporting, but this is my fourth patch accepted. I was hoping for 5, but if you followed the Bug Hunting series, my big fix for object comparisons was accepted with some caveats, meaning I didn’t get credit.

The 4 month delay is obviously due to the push for Flex 4. Though the trunk is somewhat off limits to community patches depending on your reading of the Submitting a Patch page on Adobe’s site. “We are starting Flex 4 development in /flex/sdk/trunk, so the trunk is in flux and patches submitted against it are less likely to be accepted for now.”

HTTPService bug crushed in 3.x branch

Matt C — Thu, 10 Sep 2009 14:05:50 +0000

Since I yelled about the HTTPService Bug I figure I might as well applaud Adobe for getting it fixed. The 3.x branch was updated Tuesday afternoon and the bug (SDK-22883) was closed yesterday. So, if you’re wondering what the fix is, Open up your HTTPService.as file and replace the dispatchRpcEvent method in the HTTPOperation with this:

^?View Code ACTIONSCRIPT3

    override mx_internal function dispatchRpcEvent(event:AbstractEvent):void
    {
        if (hasEventListener(event.type))
        {
            event.mx_internal::callTokenResponders();
            if (!event.isDefaultPrevented())
                dispatchEvent(event);
        }
        else
        {
            if (httpService != null)
                httpService.mx_internal::dispatchRpcEvent(event);
            else
                event.mx_internal::callTokenResponders();
        }
    }

(or monkey patch it in). 3.5 is obviously coming, and you can see the targeted fix list here

Hold the HTTPService: Flex 3.4 gets a showstopper

Matt C — Thu, 03 Sep 2009 14:54:54 +0000

It turns out that Flex 3.4 may not have been ready for the spotlight. Sure, I’m as eager as anyone to get a hold on it, especially since I had a part in its creation. However, there’s a horrible bug in the HTTPService world.

If you add a responder to a token created by the send() method, you’ll end up with two calls to the responder for every return by the server. The workaround is use event listeners rather than responders, but a lot of people (myself included) aren’t interested in rewriting legacy code to work around something that worked in a prior version.

Enigmatic Thought » FlexSDK

Bugquash badges fixed

Related Posts (generated):

Flex 3.5 finally released

Related Posts (generated):

Scientific Notation Patch Accepted

Related Posts (generated):

HTTPService bug crushed in 3.x branch

Related Posts (generated):

Hold the HTTPService: Flex 3.4 gets a showstopper

Related Posts (generated):