Edward Kandrot’s Programming Blog

Ran into this today, I wanted to put an image picker onscreen. Sounds easy, lots of examples out there, but the problem is that they do not seem to work on Universal apps.

According the Apple docs for the iPhone/iPod, one creates an UIImagePickerController. The code sample for that work great on the iPhone/iPod. When I tried to run it under the simulator for the iPad, it crashes with this message: “On iPad, UIImagePickerController must be presented via UIPopoverController”. Kind of odd, I would think that the API designer would have made it so that it would just work on the new devices, but they failed to make it so. Digging around in the Apple docs, which is the largest part of my development day, since the APIs for the iPhone are… well, too complicated to get stuff do with.

Anyway, I found the iPad dev docs, which showed how to do this on the iPad: one queries to see if the class exists like this,
Class popoverClass = NSClassFromString(@”UIPopoverController”);
if popoverClass is not nil, then go for it. That seems to work on the iPad. But now iOS4 is out, so I decide to give the code a try - it crashed with this error: “-[UIPopoverController initWithContentViewController:] called when not running under UIUserInterfaceIdiomPad.”

So, to create a Universal iPad, iPhone, iPod, iOS4 app, query the interface is no longer good enough, since the new methods exist under iOS4, they just crash. Searching online for others with the same API problems, which is how I spend the rest of my day usually, I found that one also has to check if it is running on an iPad as well.

Summary: with the latest APIs, you have to know if the class exists, whether the method under the class exists, and which device you are running on, before you call any API. Since they can change any existing API to suddenly crash on a new device, how can one write code that is future-proof? I haven’t figured it out within Apple’s API design philosophy. Stuff that could easily be hidden or made to work as default, but allow for overriding if the programmer wants more control, seems to be a foreign concept at Apple. Code that would work fine and be 100% correct for the iPhone/iPod will just fail on the iPod - not sure I understand their thinking on this.

Edward

PS By the way, the problem with adding the “is iPad” check to the exist code is that it now leaves open the door for Apple to release a device that isn’t an iPad, but requires the UIPopoverController in order to access the UIImagePickerController. I have found no way in the docs or in searching online to write a simple program to display the picker that will work on current and future devices. Ideas?

I’ve been working on a few different iPhone/iPad apps recently: games, graphics, and utilities. I decided to put what I’ve learned to the test and write a simple fun graphic toy for the iPad. I submitted it last week and it is now for sale on iTunes! It is called StaticFlow - be sure to check it out (I’ve added a screenshot and will create a video soon).

StaticFlow on iTunes

I also launched a website for my iPhone/iPad apps: prey4.us. I originally registered the domain for a game I was working on, it was only in the planning stages. I’m thinking about merging ArcaneFuture.com and prey4.us to make an all encompassing website for all of my shipping apps, code snippets, and web apps. I’ll let you know if I do that.

While waiting for approval, I decided to enhance the app - it now works on the iPhone and the new iPhone 4. I added high res graphics in OpenGL, which took longer than I thought it would. The documents were a little fuzzy, sending me back and forth between different APIs, but I figured out that in my view initialization I had to add these lines:

NSString *reqSysVerScale = @”4.0″;
if ([currSysVer compare:reqSysVerScale options:NSNumericSearch] != NSOrderedAscending) {
self.contentScaleFactor = UIScale_;
}

where UIScale_ was the scale I was going to set. The reason I did that rather than just slap a 2.0f in there was that during the touch events, it appears that the locations are not scaled. So, if the code gets a touch event at (101, 102), the code has to pass along (202, 204) to my hit detection code. I was unsure of what scale I could support without slowing down the iPhone, so I the view calculates the scaling the the processor can support, sets UIScale_, which is then used to notify the view of the new scaling, and is used later on by the touch handling code to get the correct hit detection coordinates.

It appears the scaling is easy if one isn’t using OpenGL on the iPhone iOS4 - it all happens behind the scenes by the programmer having images at both normal and scaled resolutions, then the OS selections the correct graphics for the device. Generating the graphics in OpenGL on the fly is the uncommon case, so it requires more work by the programmer, and it is less documented.

Edward

Just ran into this today while working on an iPhone app. I created some shaders for OpenGLES 2.0 and added them to the project. When I built, it would complain with this:

warning: no rule to process file ‘$(PROJECT_DIR)/Shaders/Simple.vert’ of type sourcecode.glsl for architecture i386

I wasn’t sure what it meant (searching online wasn’t illuminating), so I moved on to running my app. When I ran it, it would fail. I tracked down the failure to this line:

vertShaderPathname = [[NSBundle mainBundle] pathForResource:@”Simple” ofType:@”vert”];

which was returning nil. I did some searching online and found people talking about similar problems, but their solutions didn’t work for my problem.

I’m new to XCode, so I still do not know all of what it can do and how to set its options, so this is probably something most XCode developers know, though since I couldn’t find this solution anywhere else online, I am writing it down here.

Under XCode, there is a Targets turn-down under Groups & Files. Under Targets, there is the target name, which is also a turn-down. Under there are “Copy Bundle Resources” and “Compile Sources”, which lead to my solution. Sure enough, under “Compile Sources” were my shaders! So, I moved them to “Copy Bundle Resources”, recompiled, and ran it.

After this simple change, the bundle code worked correctly because XCode was finally copying over the files to be included in my final app. I didn’t find anything about this while searching, it is probably all there somewhere, but without the warning or keywords I’ve tried to add to this post. I hope this helps someone else!

Summary: the warning was caused by XCode trying to compile my shaders rather than just copying them to my app. Moving the files under the Targets from “Compile Sources” to “Copy Bundle Resources” fixed the warnings and allowed the app to locate the files when running.

-Edward

yahoo! destroyed 10 years of movie club data the other day. All of the movie reviews from club members, all of the movies we have seen, etc, gone. So, I decided to create a new version of the club on my server. There was something called phpBB already there, so I used that. Wow, does it really lack some nice features, but at least it is functional.

As soon as I created the club, I noticed a stream of spammers signing up. Very strange, since at the time I didn’t have any links to it, and searching for the spammers IP and email addresses, I found that they are commonly known spammers. Makes me think that phpBB gets a kickback…

I noticed that it was hard to block spammers with phpBB. Since I couldn’t block them by keyword, they could create user accounts, which would show up in the Member List as well as on the front page as New Accounts. So, they were still getting their message out, even though I deleted them as soon as I could.

So, today, I decided to put an end to this! I went into the php/sql code for phpBB and made it so that un-activated accounts do not show up in the Member List, nor in the count of active users, nor on the front page as new users. Until I active someone, it is like they never signed up!

I got the idea from this blog software. I have 2 or 3 spammers a day who create accounts here and post spam. But this blog software doesn’t show the world their spam until I approve it, which I never do. I am not sure why the default on phpBB isn’t to disable unapproved members - it is kind of weird in these days of spammers. Actually, it is weirder than that, they do not even have the option to do it - I had to learn enough php/sql to add this option for myself.

If anyone is interested in the changes, I will gladly give these back to the community - maybe 10+ lines worth of changes across 4-ish files to get it all working correctly for admin vs normal user.

-Edward

PS The movie club site: http://theedward.net/movieclub/ which is open to all who want to post about movies, or go see a movie with us on Tuesdays in the SF Bay Area (we try to go every Tuesday).

I was working on a new problem from the Euler Project - they post new problems every week-ish.  I decided to try this one in Erlang, to give myself more exposure to the language to see what I think of it.  I realized that something was missing, and actually, I think it is missing from most languages:  GCD.

 Alex Stepanov has given many talks about the importance of GCD - I think there are a few four hour talks out there that have been recorded and are very interesting to watch.  I realize more and more how right he is about this.  (His paper on GCD is linked to at Stepanov Papers, which I have linked to on the right - some great reading on that site.)

It seems to be a useful function, one that should be built in.  Given that it is not, I’ve included one in this blog post.  Code like this should only be written once, programmers should be able to focus on the problem at hand, not worry about missing common functions - it takes away from focus.  Sure, everyone could write their own, but the same is true for sin, cos, etc.  GCD should be a standard, it is in J the programming language.

Before I wrote this, I did some searching, and found that a few people had written a GCD and posted their code.  It was after looking at their code that I decided to open a dialog about this, since some of the ones I found were wrong, or longer, or more complex than the actual GCD calculation.

-Edward

gcd(M, 0) -> M;
gcd(M, N) -> gcd(N, M rem N).

I updated my Arcane Future site today, so this blog will now be linked to it!  I am using DreamWeaver to create and maintain my website.  The program looks like it can do a lot of things, but the interface makes it really hard for a beginner to figure out how to make it do those things.  I even watched some of the videos at Adobe.com, but most assume that one has created a website previously.  Then there is this blog software…  it removes all of my carriage returns putting everything in one long paragraph.  I found a way to trick it into doing what I want some of the time… 

My question is:  why does most software make the user struggle to get things done?  I only use a limited subset of software out there, mostly for programming, math, spreadsheets, and games.  And in every case, the struggle isn’t with the concepts of the application, but with the User Interface. 

I am not a UI person, my passion is for code performance.  But at some point, doesn’t UI factor into this?  If it takes a person 2x to 10x longer to get to the part where they are being creative, doesn’t that overwhelm the time I can gain by making the program 2x to 10x faster? 

Which is worse:  finding a program that does what you want, but is unusable because the UI totally obscures the  functionality, or finding a program that does what you want but it takes hours to run? 

-Edward 

I had been obsessed with Ray Tracing the past week-ish.  I wrote a basic one in Erlang, so I could learn more about this interesting language.  It was kind of slow, it took 50 seconds to generate this image:

   Then I picked up a book on F# last Sunday and started over.  F# is kind of interesting, and definitely generates much faster code.  I was able to get many more features into my F# version.  See these images:

 I will post more about my experiences with these languages as well as more about the images.  Mostly, I am testing to see if the blog software works. It doesn’t.

-Edward

This is my first post to my Arcane programming blog. My plan is to fill this with my experiences while programming as well as share some of my past experiences. I hope you enjoy it.  -Edward