Re: fs module test harness

Right. I can appreciate those issues. I was thinking more along the
common core things: reading n files in parallel, load testing, releasing
files, adding files. Those things at least should be common across the
board today.. If your sfs test suite does those + the non-generic sfs
stuff, perhaps pulling the non-generic code out would provide the basis
for a test harness. Then as the interface is cleaned up or bugs are
found we add them under the harness?

Anything that stops me completely reinventing the wheel test-wise will
be much appreciated :]

Rob

----- Original Message -----
From: "Kevin Littlejohn" <darius@bofh.net.au>
To: <squid-dev@squid-cache.org>
Sent: Wednesday, March 14, 2001 4:12 PM
Subject: Re: fs module test harness

>
> It proves difficult to do such at the moment, because of the slightly
> different interface for each module. I've got a sfs-specific test
program,
> that creates a drive, mounts, writes, reads, etc. unmounts. But it's
fairly
> sfs-specific atm, I think. Part of the work to clean up the fs
interface
> (which I'm slowly getting back to) would involve such. Most of the
testing
> Adrian and I did was using squid - although that ran us straight into
> aforementioned interface issues :(
>
> KevinL
>
> >>> "Robert Collins" wrote
> > Does such a thing exist? I've the beginnings of a win32 thread based
> > module, and I'm hoping to test it outside of squid...
> >
> > Rob
>
>
Received on Tue Mar 13 2001 - 22:17:59 MST