Re: Re: header

Wouldn't it be nice if in the $_SERVER array you could get the whole URL now that PHP has a parse_url function? Thank you, Micah Gersten onShore Networks Internal Developer http://www.onshore.com Jochem Maas wrote: > Robert Cummings schreef: >> On Tue, 2008-09-16 at 21:18 +0100, Nathan Rixham wrote: >>> Jim Lucas wrote: >>>> Robert Cummings wrote: >>>>> On Tue, 2008-09-16 at 18:13 +0100, Nathan Rixham wrote: >>>>>> I also meant: >>>>>> define('PUBLIC_BASE_HREF' , >>>>>> 'http://php.net/') >>>>>> <head> >>>>>> <base href="<?php echo PUBLIC_BASE_HREF; ?>" /> >>>>>> </head> >>>>>> >>>>>> and further thinking there has to be an easier / cleaner way of >>>>>> getting "http://" / >>>>>> "https://" in php.. >>>>> What's wrong with $_SERVER['HTTPS'] ? >>>>> >>>>> Cheers, >>>>> Rob. >>>> my way >>>> >>>> <?php >>>> >>>> function get_http_protocol() { >>>> $proto = 'http'; >>>> if ( isset($_SERVER['HTTPS']) && >>>> strtolower($_SERVER['HTTPS']) >>>> !== 'off' ) { >>>> $proto .= 's'; >>>> } >>>> return $proto . '://'; # Adding the :// is optionial >>>> } >>>> >>>> echo get_http_protocol(); >>>> >>>> ?> >>>> >>>> That is about as simple as you can get. >>>> >>> lol this one's got my brain going; here's mine: >>> >>> <?php >>> function get_http_protocol() { >>> return 'http' . ($_SERVER['SERVER_PORT'] == 443 ? 's' >>> : '') . >>> '://'; >>> } >>> echo get_http_protocol(); >>> ?> >> >> Just because 443 is the default HTTPS port, doesn't mean HTTPS is being >> served over 443, nor does it mean HTTPS isn't also being served over >> another port. This is less effective than using $_SERVER['HTTPS']. > > although I hazard to offer the following as more complete: > > function get_http_protocol() > { > $proto = 'http'; > if (isset($_SERVER['HTTPS']) && $_SERVER['HTTPS'] && > strtolower($_SERVER['HTTPS']) !== 'off') { > $proto .= 's'; > } > > return $proto . '://'; # Adding the :// is optionial > } > > because if $_SERVER['HTTPS'] happens to be false (which I don't think > it ever should be) it > would cause the return value to be 'https://' which I > don't think > would be correct in that case > (even thought the case is theoretical and likely to indicate an engine > problem if it did occur) > >> >> Cheers, >> Rob. > >