Re: gzip-based html compression

On Wed, 28 Jun 2000, Teodor Cimpoesu wrote: > Hi Sebastian! > On Wed, 28 Jun 2000, Sebastian Bergmann wrote: > > > "GZip-Encoding". Packing speed is about 200-500 Kb/sec on > > Pentium-200. The unpacking takes virtually no time. It works on both > > Apache and MS IIS." > > > > I understand that the "major" browsers (Netscape Navigator (Mozilla?), > > MS Internet Explorer, ...) support the transport of gzip-compressed HTML > > pages through HTTP. > > Sadly I was not able to find information on this standard on w3.org. > There was once a post on php3 mailing list about it, ans since then I am > using it on my server. For static pages it's just a simple mod_rewrite trick. > Check for accepted gzip encoding and rewrite the link to the .html.gz file. > > This is highly efficient for big html files (w/ lots of tables and formatting.) > > > I'd like to write, if possible, a small library in PHP that sends the > > output of PHP scripts gzip-compressed, if the visitor's browser supports > > this. > Gzip compressing support exists for some time (just add --with-zlib when configuring), > but for string compressing just from a recent date (iirc). > So you may be able to use the trick for dynamic content too. > > You may enable output buffering from php.ini for all files and add > an append file like: > > [gzip.php] > $html = gzcompress(ob_get_contents()); > ob_end_clean(); > header ("Content-Type: text/html"); > header ("Content-Encoding: x-gzip"); > print ($html); > ob_end_flush(); > > *BUT*, unfortunately, PHP's content header is stubborned enough to > get in our way every time. First time, you cannot get rid of it > (I did filed a bug-rep) second, it insists to be the last one sent, > so the "encoding" won't be seen :( > e.g. > > HTTP/1.1 200 OK > Date: Wed, 28 Jun 2000 19:06:19 GMT > Server: Apache/1.3.12 (Unix) PHP/4.0.1 > X-Powered-By: PHP/4.0.1 > Content-Encoding: x-gzip > Connection: close > Content-Type: text/html > <gzip-content-here> > > I'm just staring myself at mod_php4.c but I dunno what to change > anyway :) You mean you can't change the Content-Type? I'll try to see if we can make the Content-Type header appear before the rest of the header lines. Zeev -- Zeev Suraski <zeev@zend.com> http://www.zend.com/