The problem always seems to be Spotlight, which is used to locate the builds.

To cut a long story short: XCode places the dSYM files for the *current* build in ~/Library/Developer/Xcode/DerivedData and the dSYM’s for your archived builds in ~/Library/Developer/Xcode/Archives/ (if you’re using finder you’ll have to use ‘show package contents’ on the builds to see what’s inside).

Aside: is your Library directory hidden in Lion: chflags nohidden ~/Library/ to unhide.

Now the last time I was trying to symbolicate it was crash dumps, and some advice on SO advised me to add the DerivedData folder to my Spotlight exclusions so the symbolicator doesn’t get confused (I believe it worked at the time), of course this messes with symbolication Instruments, so be sure to remove it.

I love it when you can avoid Spotlight altogether, which just seems so flakey for things like this – and there is a good way (alt). I’ll summarise it here:

1. With Instruments stopped, click on File -> Re-Symbolicate Document
2. Search for your app name
3. Click “Locate” and choose your dSYM from ~/Library/Developer/DerivedData/APP_NAME-XXXXXXX/Build/Products/[BUILD_TYPE]-[DEVICE-TYPE]/
4. Click the Start button to begin profiling

With step 4, it’s possible you may have multiple folders with the same app name. The easiest way to find the one you need is to order by modification date, and pick the most recent.

For me, this fixed up Instruments on a per-execution basis, without needing to trick Spotlight into action. I expect once Spotlight has a chance to sort itself out, it may just work but you won’t find me holding my breath.

I’m happy I did, as these are some very convincing arguments against hashbangs, and for pushState. Using pushState alone means you can’t support IE, but as the author states “are page refreshes that bad?” – it is possible to make page loads fast, and this is a good example of something degrading gracefully.

Interesting summary of the conversation, and more pushState info.

here’s the solution, thanks to richard at 2006 dot atterer dot net

// $command is the command to run, $stdin is your input.  You get back $stdout and $stderr and $returnValue
// you'll probably want to wrap this in a function ;)

$descriptorSpec = array(0 => array("pipe", "r"),
                           1 => array('pipe', 'w'),
                           2 => array('pipe', 'w'));
  $process = proc_open($command, $descriptorSpec, $pipes);
  $txOff = 0; $txLen = strlen($stdin);
  $stdout = ''; $stdoutDone = FALSE;
  $stderr = ''; $stderrDone = FALSE;
  stream_set_blocking($pipes[0], 0); // Make stdin/stdout/stderr non-blocking
  stream_set_blocking($pipes[1], 0);
  stream_set_blocking($pipes[2], 0);
  if ($txLen == 0) fclose($pipes[0]);
  while (TRUE) {
    $rx = array(); // The program's stdout/stderr
    if (!$stdoutDone) $rx[] = $pipes[1];
    if (!$stderrDone) $rx[] = $pipes[2];
    $tx = array(); // The program's stdin
    if ($txOff < $txLen) $tx[] = $pipes[0];
    $ex = NULL;
    stream_select($rx, $tx, $ex, NULL, NULL); // Block til r/w possible
    if (!empty($tx)) {
      $txRet = fwrite($pipes[0], substr($stdin, $txOff, 8192));
      if ($txRet !== FALSE) $txOff += $txRet;
      if ($txOff >= $txLen) fclose($pipes[0]);
    }
    foreach ($rx as $r) {
      if ($r == $pipes[1]) {
        $stdout .= fread($pipes[1], 8192);
        if (feof($pipes[1])) { fclose($pipes[1]); $stdoutDone = TRUE; }
      } else if ($r == $pipes[2]) {
        $stderr .= fread($pipes[2], 8192);
        if (feof($pipes[2])) { fclose($pipes[2]); $stderrDone = TRUE; }
      }
    }
    if (!is_resource($process)) break;
    if ($txOff >= $txLen && $stdoutDone && $stderrDone) break;
  }
  $returnValue = proc_close($process);

I previously had an implementation using `fwrite` and `stream_get_contents` as some other people, and the docs suggested, which works fine for small inputs, but once you get something bigger than ~200k input, it can completely deadlock.

This implementation is better as it reads and writes the data when needed, keeping everything flowing and alive.

With that in mind, I created a replacement method for UIDevice: UDID. It uses the device’s MAC address to generate a unique ID that is the same length as the old UDIDs. I was inspired by some other people’s solutions, but rolled my own implementation which I feel is neater, and more compatible (as it generates similar 40char UDIDs). You can fork it on github.

It’s not a drop-in replacement, for fear of monkey patching over an existing method. Just do a find & replace with: [[UIDevice currentDevice] UDID].

Of course: consider if device identifiers are right for you. To identify a specific install of your app (which can be migrated from one iPhone another due to loss/upgrading), generating a UUID and saving that as a NSUserDefault would work best (since it will transfer with the app data).

License: public domain.

The most awesome is KDiff3 (search for ‘Apple Mac OSX binary’ on that page for the Mac binary). This is a three-way merge tool with an editing panel below. Seriously where has this been all my life? Amazing.

Download it, and install – make sure it’s in your system PATH. Then simply run git config --global merge.tool kdiff3 to set it as your default merge tool. Now every time you get a conflict, your git mergetool command will actually be rather helpful! Basically you are shown the base revision (that both branches split from) in the middle, one branch either side, and an editor below allowing you to pick which side you want and/or edit the file directly.

The second tool I (re)discovered was SourceTree (Mac App Store DL Link). It is a Mac GUI client for git. I’m pretty comfortable with the command line, but what I love about this tool is it allows me to review my changes before I commit them (either staged or unstaged), discard hunks I don’t want, and of course open the file in question for editing. Very convenient. git diff is nice and all, but I love having that handy ‘discard hunk’ button in SourceTree so I can skip my silly changes. For me this is the most useful feature, though no doubt there are others. Oh and it’s free (it’s a bit confusing, as there is a “trial” period, but after the trial you simply register for free, and are given a license).

Refs:
1. Some useful git config tips
2. Graphical Diff Tools for Mac OS X
3. Setting up mergetool on windows

ActiveRecord::Base.connection.select_one('SELECT COUNT(*) FROM mytable')
ActiveRecord::Base.connection.execute('SELECT * FROM mytable')

Thanks revgeorge.

An iOS team is actually in two parts: the iOS Developer Team, accessed via the Member Center, and a completely separate iTunes Connect team.

It turns out that the one AppleID can belong to multiple iOS developer teams, and Apple asks you when you login to the member center which team you want to use. However iTunes Connect has no such feature, and each AppleID can only belong to one iTunes Connect “team”.

So you can re-use AppleIDs for developer accounts, but not itunes connect accounts. It’s up to you I guess if you want a totally separate AppleID for each team, or just each iTunes connect account.

I set it up with the help of this fantastic tutorial. My steps after the break.

Install

yum install monit

Configure on boot

/sbin/service chkconfig --levels 235 monit on

Create the config file

vi /etc/monit.d/monitrc

With something like:

set daemon 30
set logfile syslog facility log_daemon
set mailserver localhost
set mail-format { from: monit@www.example.com }
set alert you@example.com

check process sshd with pidfile /var/run/sshd.pid
start program "/etc/init.d/sshd start"
stop program "/etc/init.d/sshd stop"
if failed port 22 protocol ssh then restart
if 5 restarts within 5 cycles then timeout

check process apache with pidfile /var/run/httpd.pid
group www
start program = "/etc/init.d/httpd start"
stop program = "/etc/init.d/httpd stop"
if failed host www.example.com port 80 protocol http
and request "/monit/token" then restart
if cpu is greater than 60% for 2 cycles then alert
if cpu > 80% for 5 cycles then restart
if totalmem > 500 MB for 5 cycles then restart
if children > 250 then restart
if loadavg(5min) greater than 10 for 8 cycles then stop
if 3 restarts within 5 cycles then timeout

set daemon is used to specify the number of seconds between each check.

As per the docs, it is here.

Anything in single quotes is put through as-is.

Look no further than this awesome test app! Market link here.

Some high quality phone image marketing resources. I like this one.