Under The Bridge

Now here’s something we’d sorta been pondering ourselves might be worth doing; this fellow’s taken the GPL’d crossplatform game Tux Racer and ported it to the iPhone …

… for $3.99. Whilst providing a desktop build for free, and source downloads at the bottom of the product page which apparently this fellow feels meet the terms of the GPLv2 sufficiently.

We say “apparently”, because there is a body of opinion out there that since the casual consumer cannot edit the source code to an App Store-bought application and run the edited program directly on their phone, it is impossible to comply with the terms of the GPL with an application distributed through the App Store. We don’t subscribe to that personally, mind you; first off, nowhere in GPLv2 does it explicitly state that you must be able to run it on the distributed device — so being able to run your modified code in the simulator, which requires nothing but the freely downloadable Xcode toolchain, arguably satisfies the license adequately. And even if you do insist that the terms implicitly require that you run the modified version on the distributed device, then we don’t think that the $99/year developer program fee to Apple to be able to run whatever source you like on your phone is a condition that amounts to a GPL violation, any more than needing to buy a commercial compiler to do something useful with any desktop GPL’d source would violate it. But as the Free Software community is disproportionately populated with Apple-haters, mere logic is not always sufficient to appease them; and as this (to our knowledge) is the first GPL-licensed iPhone program that anyone has attempted to charge money for, if there is anyone out there willing to test their anti-iPhone interpretation of the GPL in court, well, this is whom they’d probably pick for a test case.

So we’ll be watching to see if any of the huffers and puffers get all indignant about this fellow’s entrepreneurial bent; and if not, we’ll probably look for some GPL’d game to port ourselves, as we would rather like to have an open source iPhone project to direct those who would like to see some demonstration of our coding practices to, and hey if it might make a buck or two that’s always great as well! In the meantime, if you’re a desktop Tux Racer fan, and not a programmer yourself — buy it to encourage more GPL’d source ports!

Colophon:Well, today was to be the final client-provided post of The Great Wordpress Client Test, finishing up with über-text-editor

TextMate 1.5.8!

but the actual testing ends up a DNF, as we couldn’t be bothered actually getting TextMate to work. See, its blogging extension’s bundle’s idea of configuration is to have you put your blog’s XMLRPC address in a text file. We gave it a couple tries at what we think that address should be for this blog, but it failed to make them work. And we figured we just wouldn’t bother to keep trying, since obviously there wasn’t going to be any useful conveniences here at all, so at best it would make it into the same class as the other non-WYSIWIG tools. So we’ll award it 2/10 to put it at the bottom of our ratings due to its completely beyond the pale requirement of “needing a clue about blog internals” and leave it at that.

So one of the most difficult things about being an iPhone programmer is figuring out just how darn much memory you’re actually using. And by way of this post on leak finding in an iPhone app, here’s a snippet for accomplishing that:

#import <mach/mach.h>
#import <mach/mach_host.h>

static void print_free_memory () {
    mach_port_t host_port;
    mach_msg_type_number_t host_size;
    vm_size_t pagesize;

host_port = mach_host_self();
    host_size = sizeof(vm_statistics_data_t) / sizeof(integer_t);
    host_page_size(host_port, &pagesize);        

    vm_statistics_data_t vm_stat;

    if (host_statistics(host_port, HOST_VM_INFO, (host_info_t)&vm_stat, &host_size) != KERN_SUCCESS)
        NSLog(@"Failed to fetch vm statistics");

    /* Stats in bytes */
    natural_t mem_used = (vm_stat.active_count +
                          vm_stat.inactive_count +
                          vm_stat.wire_count) * pagesize;
    natural_t mem_free = vm_stat.free_count * pagesize;
    natural_t mem_total = mem_used + mem_free;
    NSLog(@"used: %u free: %u total: %u", mem_used, mem_free, mem_total);
}

If your application has non-trivial memory needs, you might be well served to bring that up at launch and tell the user to reboot if memory is low, for instance!

Colophon:
For Day 1 of The Great Wordpress Client Test, this post is brought to you by:

Blogo 1.2.8!

Assuming, you know, you’re actually reading this, it does indeed set up and work transparently with WordPress 2.8, but there’s a few things that less than thrilled us:

• We’re not overly impressed with the HTML it produces compared to the online editor. We found extraneous paragraphs and the like once uploaded and we noticed that it broke the #include statements.
• Switch from rich text to HTML mode and back, or to edit mode and back, and you can’t undo. The online editor gets that right.
• Pasting and formatting that code snippet had a variety of associated annoyances; pasting in the wrong place, thinking the #includes were followed by tags or something. Eventually we deformatted the copied text in TextEdit and pasted it into the HTML view to get it to not do things wrong. That’s worse than the online editor too.
• We’re just not really fans of the interface. Maybe it grows on you.
• It completely failed to post the above picture — we added that online after posting!
• It ate the filenames in the #include statements during upload, apparently thinking they were HTML tags; we fixed that manually with the online editor. It should be smarter about sorting that kind of thing out.
• It mucked up the assigned categories when posting, and we fixed that manually too.

So there’s a number of reasons to look skeptically at it, and that picture problem is pretty much a dealbreaker, since we do like to put the odd picture in our posts. On the plus side, it does have a lot of nifty conveniences, the previous post editing looks nicely worked out and the setup guidance was a veritable exemplar of how to get people started properly … but we’re just not terribly thrilled with it overall. Out of ten, we’ll give it a four, and if someone happens to know The Magic Trick We’re Missing for getting the embedded picture to upload, then we’d give it a seven. But we’re certainly hoping that the remaining six work out better, even if the image upload fail is something easily correctable … because if not, sticking with the online editor is looking like a better idea.