As Magento software gets more popular, we’re increasingly running into people who are experiencing errors with our setup. A majority of these are addon issues from Magento Connect and have been addressed via their forums / knowledge base – though we’ve had one that took us a moment to figure out.
Fatal error: Function __autoload(b514i6_xp_2pia3m5o191rxqn3) threw an exception of type ‘Exception’ in /home/<user>/public_html/magento/lib/Zend/Validate/Hostname.php on line 420
Long story short – this is because we include the Zend Framework on our include path for everyone to utilize freely – the issue is Magento also bundles Zend Framework locally since they rely on it as well. They likely do this to maintain a stable version across all platforms – and possibly slightly modified it (not confirmed). As a result people trying to install Magento on our machines were getting that obscure error.
To fix it it’s quite simple – change the default include path to your home directory (put this in the .htaccess file in your public_html/ folder):
php_value include_path /home/<yourusername>
Hopefully this solves some issues people have with a similar setup as ours.
-Cody
This is an ongoing issue, and it would be simple for the Magento folks to fix. I remember I had a similar issue in an early 0.x version of Magento and even reported it somewhere, but it looks like there is no fix.
Anyway, instead of appending the include_path, when the shop is initialized, they need to prepend to the current path. The include path is always search from left to right, e.g. .:/foo:/bar — then the current directory (aka “.”) is first, then /foo, before /bar.
So If Magento would be “first” there would be less issues with the environment.
Regardless of the above, the default include_path should be:
php_value include_path .:/home/
🙂 Though it probably doesn’t matter since Magento adds its own jist anyway. But some of your own scripts might fail if you forget the “.”.
Definitely good information – it is nice to see that you include the Zend Framework 🙂
@till – That was an oversight on my part – I’ll make sure to update the post shortly with relevant information. Typically the people running into this issue the above is fine, though you’re right depending on what scripts they’re running.