I had tried to do this over a year ago, but couldn’t quite get it working correctly. I had a number of problems before, notably that my hosting server only ran PHP4, whereas (at the time) most OpenID plugins seemed designed for PHP5, or a specific version of PHP4 that I wasn’t running. The other main factor was that most of the OpenID plugins at the time were very early (beta or earlier) versions, so may have been a little buggy still, whereas the plugin I installed today was very easy to get set up and running.

Anyone using a verified OpenID to post a comment on this site will find that their name and email is not required, plus your comments skip moderation.

Please let me know if you find any problems in using your OpenID on this site.

I have apilib.php which contains the PHP calls to the functions, then service.php is a wrapper around the apilib.php and provides the actual REST based services. Service.php is basically just a big switch statement, so, since PHPDocumentor needs the function declarations I can’t actually base the service documentation on service.php, so I’m using apilib.php as the basis for the service documentation.

I created my own PHPDocumentor template, which has been straightforward enough, though I did come across a couple of little issues – nothing too big and nothing that can’t be worked around .

These were the 2 little hacks I needed to put into my template:

1. in apilib, I’ve named the functions using camel case, whereas the actual method calls in service.php need to be lower case – so I amended the stylesheet to use: {text-transform:lowercase;} where the function name was displayed
2. I also needed to remove the dollar signs from the front of all the PHP parameters, so to do this I just used the Smarty templating command ‘replace’: {$functions[func].params[params].var|replace:’$':”}

I still have some of the documentation to actually write into apilib.php – most notably example service calls for each method and double checking the return descriptions are consistent (and correct). All is working out well and you can see the generated docs on the Cohere site.

This one takes an associative array of objects to turn into JSON object:

/**
* @param $objects object to turn into JSON
* @param $name overall name of the JSON object
* @param $callback name of the callback function
* @return string of the JSON object
*/
function phpToJSON($objects,$name,$callback=''){
$str = '';
if ($callback != ''){
$str .= $callback.'(';
}
if ($objects){
$str .= '{"'.$name.'":[';
$okeys = array_keys($objects);
for ($i=0; $i< sizeof($okeys); $i++){
$myobj = $objects[$okeys[$i]];
$attr = get_object_vars($myobj);
$str .= '{';
$keys = array_keys($attr);
for($j=0;$j< sizeof($keys); $j++){
$str .= '"'. $keys[$j] .'":"'. $attr[$keys[$j]] .'"';
if ($j != (sizeof($keys)-1)){
$str .= ',';
}
}
$str .= '}';
if ($i != (sizeof($objects)-1)){
$str .= ',';
}
}
$str .= ']}';
}
if ($callback != ''){
$str .= ');';
}
return $str;
}

and this one takes an ADOdb recordset and turns it into JSON object:

/**
* @param $objects object to turn into JSON
* @param $name overall name of the JSON object
* @param $callback name of the callback function
* @return string of the JSON object
*/
function phpRSToJSON($rs,$name,$callback=''){
$str = '';
if ($callback != ''){
$str .= $callback.'(';
}
if ($rs){

$str .= ']}';
}
if ($callback != ''){
$str .= ');';
}
return $str;
}

Hope you find these useful, and if you spot anything wrong (or that could be done better!) – please feel free to let me know

Essentially what’s been happening is that if the remote url is accessible (i.e. an fsockopen connection can be made to it), but the remote url takes a long time to respond, then the connection will just hang until a response is given – even if this exceeds the given timeout setting.

So if I have my timeout set to 10 seconds, but the remote URL takes 15 seconds to give a response, the script will NOT timeout after 10 seconds, instead it will wait until the remote URL responds. So my impression is that the timeout setting is only a connection timeout, not a general timeout – which was contrary to how I thought the timeout setting functioned. Thinking about if after this kinda makes sense, as the timeout is only on opening the connection, not on fopen – so it can’t know the remote url is taking longer than expected to respond.

As I can’t see a way to set a ‘real’ timeout on an fsock or fopen connection, I had a go using the cURL library. This appears to work in the way I expected – i.e. setting a timeout of 10 seconds would mean the script would stop after 10 seconds as the remote URL had not responded.

ScriptAlias /php/ “c:/PHP/”

AddType application/x-httpd-php .php .php5

Action application/x-httpd-php “/php/php-cgi.exe”

SetEnv PHPRC “C:/php”

I then also needed to update the directory security section in the httpd.conf file too, to allow Apache access to the PHP directory.

For info, this was the best tutorial I found for getting all this set up – although it’s for MySQL 4.1, it still holds for MySQL 5.

So it looks like we’ve got 3 options: 1, install the software on the same server so the url is the same and everything will be happy, 2, make proxy http calls, or 3, use frames/iframes. Option 3 is pretty much ruled out straight away, and I;ve gone for testing out option 2 (as I’m not sure if we’ll be able to run the buddy space service on the same machine as Moodle).

So my first task for this was to figure out how to make http requests in PHP, for this I needed pear installed. I’d read that pear was included with PHP 5 but jsut needed activating, but I didn’t have go-pear.bat in my PHP installation. This was because I’d just downloaded the PHP 5 installer rather than the PHP 5 zip package. So I downloaded the zip package and unpacked over the top of my existing PHP 5 install (overwriting any files) and that appears to have worked fine, as I now have the go-pear.bat and necessary extensions – just need to actually install them now!

Another thing we’ve been looking at at is how we’ll do the XML parsing in Moodle using php. Most Moodle installations run on PHP 4 which doesn’t have built in XML functions, whereas PHP 5 does. There are libraries available for PHP 4 to do the XML parsing, but these (I think) would mean rewriting the code if we (and/or Moodle) moved to PHP 5. The alternative (so code didn’t need to be rewritten for different versions of PHP) would be to use regular expressions on the XML string – but that seems like a bit of a kludge to me!