Thanks - that makes me feel a bit better. ouple of clarifications... On 14/01/2008, Douglas E. Warner <[EMAIL PROTECTED]> wrote: > > On Saturday 12 January 2008, Mycroft Project wrote: > > I hope I've got something wrong... > > > > I tried the php snippet above in a testing directory which seems to be > > entirely destructive. Adding: > > echo $$__v, " | ", $__v, " | ", $__sg, "\n"; > > before the unset() line seems to suggest that a lot of the Mozdev > internals > > like $page are also scrubbed - I didn't get any output at all from the > page > > I was trying to load. > > The snippet I provided probably is too destructive due to mozdev using > global > variables a little too liberally. I'll include an updated version below. > It would probably be better to use Mozdev's defines for accessing the > current > page, which would be PAGE (without a dollar sign) (yes, I know this is > probably undocumented).
I was about to write that I wasn't using $page anywhere then did a search... hmm... I thought it was Mozdev's use that was stopping the page from loading at all but no longer entirely sure. > More generally, if I read > > http://uk.php.net/manual/en/language.variables.predefined.php correctly > > then $_GET['foo'] is okay (this wasn't what I understood from the > original > > message) but $REMOTE_ADDR has to be replaced by $_SERVER['REMOTE_ADDR'] > It > > also seems to suggest $HTTP_*_VARS is acceptable (though discouraged) > given > > the default value of register_long_arrays. > > $HTTP_*_VARS is also deprecated and should be avoided; it has been removed > in > PHP 6. In general, only the $_* variables should be used for getting > pre-set > data from PHP. Sure, can you confirm how register_long_arrays will be set initially. Just a matter of prioritising really. Thanks, Charles -- Charles Caygill Mycroft Project Owner http://mycroft.mozdev.org
_______________________________________________ Project_owners mailing list [email protected] https://www.mozdev.org/mailman/listinfo/project_owners
