> > > If readfile is indeed that "broken", maybe we can talk the
> > > powers that be into dumping full HTTP headers through the
> > > readfile call in a future version.... I mean, the main
> > > script has it already from when it was called... so it
> > > should be able to sub-call pages in that same context.
> >
> > readfile() is not broken at all.
>
> "broken" was in quotes for a reason... yes, it works the way the docs say -
> but it doesnt work well enough in light of the fact that virtual gives me
> garbage.
Well, shouldn't you be complaining about virtual() then?
> If I do a readfile() request to that SHTML file through PHP3.0.15, I ALWAYS
> get a Netscape based response, even when the calling PHP3 knows that it was
> called from an IE browser. This is no doubt due to the HTTP 1.0 connection
> used (To quote the online PHP manual:)
>
> If filename begins with "http://" (not case sensitive),
> an HTTP 1.0 connection is opened to the specified server
> and the text of the response is written to standard output.
>
> This is nice, and things like .CGI's and other scripts can be used this
> way - but not completely transparently as can be seen in my case (and this
> will of course fail when referencing a name-based host... something thats
> becoming more and more prevalent these days).
>
> But obviously it's not a FULL HTTP 1.0 connection - other wise it would have
> a
>
> User-Agent:
>
> line in the request header, wouldnt it?
It is a full connection and a User-Agent header is sent. The user agent
is reported accurately as "PHP" followed by a version number. It is not
Netscape nor IE making this HTTP request, it is PHP.
-Rasmus
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]