Flex 4.5 Mobile

I have spent some time pouring over the specs trying to make sure I understand them and I plan to do a couple summary posts here in the near future. What I have found is that the idea of how to write an application for mobile is a little different from how we do these things on the desktop with Flex. The main idea is that there is really only one Section or View that is available at any one time. So the MobileApplication class builds on top of the Spark application to facilitate this by providing a ViewNavigator. The ViewNavigator is a new class in Flex 4.5 which appears to behave like the ViewStack of MX. The noticeable difference between ViewStack and ViewNavigator is that ViewStack can only change through content that was predefined, where as you can add new content to be displayed at runtime to the ViewNavigator. In fact with the ViewNavigator you use the views as a sort of stack where you push and pop the views to set what is displayed.

All of this is very interesting and I encourage you to go and read the spec documents on the open source site. There is lots of good information there and you can get a jump start on learning Flex 4.5 by digging into these documents now. Also as we have an Open Iteration meeting coming up if you dig into these docs now, then you might be able to ask questions during that call.

So get on over and checkout the Flex 4.5 mobile specs on the Adobe Open Source site.

Unless we get some very scary bugs, these files will be very close to what the final version of 4.1 will include. I think the only known change we have planned is building against the new Hamcrest that Drew tagged this weekend.

The team has already been hard at work prepping fresh features for the 4.2 release which has some new CI functionality as well as a bunch of new tests written with Mockolate. So despite being near a stable release for 4.1 the team is certainly not resting on its laurels.

So do your part to make FlexUnit 4.1 the best product it can be. Download the RC1 builds and give them a test. Let us know if you have any issues.

So you can guess just how excited I was when I noticed an announcement for the next Open Iteration Meeting has been announced on the Adobe open source site.

From the Hero Home Page:

Open Iteration Meetings

We will be having the first Hero Open Iteration Meeting on October 1, 2010. Please join the Hero product team as we cover work that has completed in past Hero iterations and discuss what’s in store for upcoming iterations.

Meeting: Hero Open Iteration Meeting
Date: October 1, 2010
Time: 8 – 9 AM PT (11 am ET, 4pm UK, 5pm CET)
Connect Details: http://my.adobe.acrobat.com/hero-open-source/

So mark that date on your calendar and make time to contribute back the the framework that has made the flash platform available to so many. Be heard and share your feedback with the Flex team.

However late last night a bunch of new information was posted on the flex wiki regarding the new DataGrid functionality in spark. This caused me to go digging around what else had been posted and I found that the site had been updated ( not sure when) to actually include a build of Hero (code name for Flex 4.5).

This build does not include the mobile items but does include many of the new spark controls that are detailed on the wiki. So if you have just been dieing to play with the new spark controls for Form, Image and DataGrid then hurry over to the Adobe Open source site and get you some Hero.

Hero Home Page
Hero Downloads

The really fantastic part of this giveaway is that if Dan gets enough signups he intends to start a scholarship for the next 360 so that someone who would not have previously had the ability to attend such an amazing and empowering conference would be granted the chance.

So again, if you have not bought your ticket to 360Flex in DC, Please do so through Dan Florio’s site.

So I put in the time and got this thing working great. It fires up and displays a pretty animation as a screen-saver, if you interact with the mouse it loads up the form for people to enter contact data and it was great. With one little exception. If you hit the escape key it took the application out of full screen and it didn’t work very well. Since this application is going to be running during a trade show they want to not make it very easy for the user to get out of the application. I thought this would be simple so I added a key down listener to the application and setup a listener with a break point so I could debug. The problem was unless i was out of full screen I was not hitting the break point. Which really doesn’t do me any good since I wanted to stop it from leaving full screen to begin with.

So I did a little googling and came up with a couple blog posts of people doing exactly what I wanted. The problem was when I ported their code into my application things were not working. You see I had an event handler on my WindowedApplication that called the init() function in response to the creationComplete. But when i got to init() my stage property had not yet been set. So I just used the nativeWindow.stage property thinking that they would end up being the same thing. Well as it turned out I was not catching any events on that stage for key presses.

It wasn’t until I posted on twitter thinking I had found a bug in the new 2.0 air stuff (both examples linked above were for the 1.5.2 release) did things become clear. You see the secret is in the stage. And the fact that I didn’t have a stage for my application should have been the big indicator. In a flex application creationComplete is the event that says all is well and you are ready to get started. However in an AIR application the final event that announces to the world you are ready to be used is the applicationComplete event. Of course the stage was not ready yet, the window had not been constructed at creationComplete.

So if you should come to building your own Kiosk and you need to catch key events when in full screen, remember to add the KeyboardEvent.KEY_DOWN event to your stage in your windowed application. And make sure you wait for applicationComplete to try to set the event listener.

Big props to Nate Beck for reminding me that creationComplete was just not gonna cut it. :)

So all of that works wonderfully. I have been working on the application on my desktop and everything has been great. But tonight I wanted to work on the application from my laptop. Heroku uses SSH security to allow you access to your application. On github there is a section in your account where you can upload your ssh keys so they can be tied to your account. But in looking through the admin pages on Heroku I could find no such section. In fact after thinking about it a bit I could not remember setting one up on my desktop the first time.

The trick here is that the Heroku app does this for you when you initialize the application. So to add a second computer to your account on Heroku you create a new git repository, then from that directory run heroku config. This will guide you through logging into the Heroku platform as well as uploading your ssh keys so you can access the repository. So the actual commands would look something like

1
2
3
4
5
6

mkdir myapp
cd myapp
git init
heroku config
git remote add heroku git@heroku.com:myapp.git
git pull remote master

Hope that helps if you need to use multiple computers to work on one application on Heroku.

PS. if you want to add additional contributors that is actually covered very clearly in their docs here

Android 2.2 has not been officially released yet (although some people have gotten OTA updates). However the rom has been leaked and I have found several places to download it. Unfortunately after trying to flash the rom to my devices and getting a failure, I read a few posts on forums and found that this rom won’t work for AT&T devices.

The root cause is that the build number required to take the upgrade is not what ships on the AT&T N1. And as such the upgrade fails.

There are many people out there that have gotten around this issue by “rooting” their phone. What this essentially means is that they unlock the bootloader and load an alternate boot loader that is able to bypass some of the restrictions of the native loader that ships with the N1. Unfortunately unlocking your N1 voids its warranty and at this point there is no way to re-lock your bootloader. So it is not a reversible change. For the record I have a HTC Hero that I have rooted and am running Android 2.1 on. There are benefits to rooting your device, I am just not ready to bite that bullet for this brand new piece of hardware that *should* work just fine the way it is.

So in the end if you are Flash Developer who wants to play with FP 10.1 in the browser either buy the TMobile version of the N1 or you will have to wait for an official update for the rom. Luckily in the mean time we can still use the new AIR bits because they work with Android 2.1 just fine.

Just trying to save someone the heartache I felt when I realized I just bought this phone that can’t actually do what I want yet. It will… I just have to be patient. Something I am not very good at.

Beyond the additions to the actual SDK to make our lives easier, Flex Builder has been updated and re-branded Flash Builder. The re-branding is essential in understanding the scope of what that product does, and helping Flex developers is just one small part of that.

Adobe has created a comparison chart that includes features of Flex Builder 3, Flash Builder Standard, Flash Builder Premium, and the Flex 4 SDK. It does a great job of showing what features are included in what product and also which existing features have been updated or enhanced.

Check out the comparison chart here.

Apple goes out and builds this great platform that revolutionizes how people view their phone. Every wireless company and handset manufacturer in the world has been trying to tell us for years that the mobile space is where its at. People spend very little time with their personal computers, but their cell phone is with them all day long. Adobe spent 4 MAX’s conferences in a row telling us “this is the year for mobile”. Yet it never happened. People wanted the cheapest phone they could get from their carrier and the carriers had to discount nicer phones to get people to upgrade and extend their contracts. Then the iphone came out. Not only are people jumping ship from their cell carrier and their crappy free phone, they are willing to stand in long lines and pay $700 for a phone! The world turned on its ear and overnight every other handset manufacturer and wireless carrier was behind in the race. Developers start making money writing apps. Users start buying apps and they spend hours wasting time reading twitter and facebook wherever they are. The world is changed, we even get the “real” internet with us in our pocket every minute of every day. Except that its only mostly the “real” internet. We see most of the content, but all over there are these little boxes with a blue lego on them. I wonder what was supposed to go there? Oh, that’s right! That is all the flash content from the “real” web that is missing from the “i” experience. So the users demand, the developers cry, but apple says no. Flash is to resource intensive and buggy to run on their blessed device. The best hand held phone in the world, the “jesus” phone can’t run flash. And that is where the secret plan started.

In that little statement Apple challenged the world. The most amazing piece of technology every made can’t run flash. Every competitor in the market is trying to figure out how to get their market share back and actually challenge the dominance of the iPhone. And now they have their chance. If anyone can build a handset that can run flash, and the iPhone can’t, then it must be a better phone than the iPhone! And so now, Sony Ericsson, HTC, Blackberry, Nokia…. 19 of the top 20 handset manufacturers are releasing phones this year that run Flash Player in the browser! 19 of 20! Blackberry for gods sakes, they can’t even make a browser work worth a damn but now they are gonna have Flash Player!

So after seeing this all come together I have to assume that this was a marketing move that was designed by Steve Jobs himself. Because the sure success of this endeavor could only have been put together by the mastermind that is responsible for Apples recovery in the world. There is no way that by accident, Apple could have caused the proliferation of the Flash Platform to every other device on the market and there by ensuring its continued dominance and the success of the Platform.

No, that “no flash on the iPhone” plan could not possibly have backfired at all. :)