Re: Dynamically generated objs

From: Henrik Nordstrom <>
Date: Tue, 17 Mar 1998 23:06:40 +0100

Yee Man Chan wrote:
> How does Squid bypass objs generated by a CGI or Java or JavaScript
> program? I want to know how it detects them and where the relevant codes
> are. (my squid is 1.1.20)


Bypass what? Squid can't bypass itself.

There is three different mechanism at play here

1. hierarchy stoplists. Forces URLs matchin the stoplist patterns to
always go direct (from the assumption that the resulting objects is

2. cache stoplists. To force squid to not cache URLs matching certain
patters, since we don't trust the servers to send correct expiry

3. Expiry information generated by the origin server. (Expires,
Cache-Control, Last-Modified, and some other headers)

Java & Javascript does not differ in any way from standard HTTP objects.
JavaScript is only text encoded into the HTML page, java is only one or
more binary files.. Squid does not handle these any different than other
HTTP objects.

Squid can't detect javascript or Java 1.0 .zip archives, as this
requires parsing of the HTML pages (Squid only parses the headers, not
the actual contents of the page).

Java .class files and Java 1.1 .jar files can be detected by squid,
since they use a unique extension.

Neither javascript or java is dynamically created objects in HTTP. The
dynamics is only inside the browser. On the server (and in Squid) the
objects are static.

Henrik Nordström
Sparetime Squid Hacker
Received on Tue Mar 17 1998 - 15:42:14 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:39:24 MST