> > The second reason is that to test our code we developed a tool
> > for testing squids behaviour and wonder if you would want to use it.
>
> Very interesting. Please eloberate on the capabilities of this tool.
The tool consists of a few perl/bash scripts, so it's not too sophisticated.
It tests squid's interactions with clients and servers (i.e. messages squid
sends).
A single test goes like this:
- a predefined request is send to squid
- dedicated server listens for squid's reaction
- the server logs squid's request and replies with predefined response
- the client receives squid's response and logs it
- squid's request and response are compared to expected request and response
to verify if squid behaved correctly
Currently the comparison tries to match each line of the expected
response/request file with squid's actual response/request.
There are two options of this "line matching":
- the first when verification succeeds when the line matches any line i
actual squid's reponse/request,
- the second when verification fails when the line matches any line i actual
squid's reponse/request
This allows tests to verify if squid sends proper headers, as well as
if squid doesn't send improper headers...
Besides the logging server and client the tool includes some scripts to
automatize the testing.
The first draft of the description of the tool and the scripts themselves
can be found at our page:
http://rainbow.mimuw.edu.pl/~ms189442/mmsquid/mmsquid.php?menu[0]=presentati
ons&menu[1]=squid&menu[2]=testtools&lang=en
I will have this page updated till the end of the next week so the link
above won't point to "a draft of description" but to the full-featured
description.
Any comments are welcomed :)
Regards,
Mati.
Received on Thu May 13 2004 - 04:35:40 MDT
This archive was generated by hypermail pre-2.1.9 : Mon May 31 2004 - 12:00:02 MDT