Engines are Go…

Sounds very cool
Tho a sidenote, it’s just as easy to maintain a local PEAR as it is to have a local version of any other lib.

PEAR::Service_ExchangeRates however is under the PHP License. That restricts distribution and honestly I can’t see many shared host users maintaining a local PEAR repo. At the end of the day, it’s more worthwhile to put a ZF component in place. I don’t mean to be seen as criticising PEAR, it just doesn’t quite meet my requirements for open source projects.

I would welcome this class. I was thinking if maybe it should contain a caching mechanism so it doesn’t hit the webservices more than needed, but then again, that is probably not the responsibility of this class. (But maybe should be noted in the docs)

Yeah I knew you didn’t like the license but mearly wanted to point out the fact that it’s no harder to include a PEAR lib in a project than it is to have ZF component or a eZ one (I’m not talking about a full fledge pear thing via command line or such)

The plan (given it’s early days in the proposal) is to use a Zend_Cache module being developed separately. This will let us cache data for a specified period possibly based on the update frequency of the related source.

There is also a usage for Zend_Locale and Zend_Currency to handle locale specific issues. We’re avoiding any bias towards the en_UK/en_US conventions.

If you follow the mailing list archive, André has made some comments on the draft, and his own is pending.

I created this snippet a while ago; I don’t think it is exactly what you are looking for but it is functional. Based on a csv market service from the Bank of Canada.

[geshi lang=php]define(‘FILE_EXCHANGE_RATES’, ‘http://www.bankofcanada.ca/en/markets/csv/exchange_eng.csv’);

$nameToCode = array(
‘USD’ => ‘united s’,
‘EUR’ => ‘european’,
‘ZAR’ => ‘south af’,
‘GBP’ => ‘pound st’,
‘GBP’ => ‘u.k. pou’
);

$lengthMatch = strlen($nameToCode['EUR']);

$csv = fopen(FILE_EXCHANGE_RATES, ‘r’);
if (!$csv) {
trigger_error(‘Could not open market values.’, E_USER_ERROR);
exit;
}

$values = array();
while ($line = fgets($csv)) {
$start = strtolower(substr($line, 0, $lengthMatch));
if (($pos = array_search($start, $nameToCode)) === false) {
continue;
}

// Assume the latest value is at the end of the line
$rates = explode(‘,’, $line);
$values[$pos] = (double)end($rates);
}

fclose($csv);

if (!isset($values['EUR'])) {
trigger_error(‘Failed to receive base currency EUR.’, E_USER_ERROR);
exit;
}

$baseCurrency = $values['EUR'];
foreach ($values as $index => $value) {
$result = mysql_query(‘UPDATE currency SET EUR = ‘ .
($value / $baseCurrency) . ‘ WHERE code = \” .
mysql_real_escape_string($index) . ‘\”);
if (!$result) {
trigger_error(‘Could not update currency ‘ . $index . ‘.’,
E_USER_WARNING);
}
}[/geshi]

MySQL schema

[geshi lang=sql]CREATE TABLE `currency` (
`currency_id` tinyint(3) unsigned NOT NULL,
`name` varchar(50) NOT NULL,
`code` char(3) NOT NULL,
`symbol` varchar(7) NOT NULL,
`EUR` float NOT NULL,
PRIMARY KEY (`currency_id`),
KEY `code` (`code`)
) TYPE=InnoDB AUTO_INCREMENT=5 ;

INSERT INTO `currency` (`currency_id`, `name`, `code`, `symbol`, `EUR`) VALUES (1, ‘Euro’, ‘EUR’, ‘€’, 1),
(2, ‘Pound sterling’, ‘GBP’, ‘£’, 1.46803),
(3, ‘United States dollar’, ‘USD’, ‘$’, 0.780612),
(4, ‘South African rand’, ‘ZAR’, ‘Z’, 0.111675);[/geshi]

Nothing wrong with it I can see. While Zend_Service_Currency will have a few preset sources, there will also be an interface provided so that others can append further sources. As I noted somewhere recently in the mailing list – just two sources is worthless since the rate most likely used by a business is pre-determined based on their Country, Locale, and Accounting Policies.

It’s a solid piece of code.

You are back! Thank god you are back!