RE: Collaborative Flaming

Bob Denny (
Mon, 3 Oct 1994 21:00:53 +0100

I can speak with some authority on the subject of NCSA HTTPD 1.3 as well.
I "adapted" it to the Windows 16-bit and 32-bit environments. It took
a lot of work, and Rob was there for me when I needed it. I did spend
a lot of time working through the code before embarking on what turned out
to be a mostly rewrite. I wouldn't call that code "maintainable", but then
I wouldn't say too much about it being "bad" either. My only real complaint
was that it depended on the Unix semantics of forking children and letting
the process termination clean up all the memory allocations. Because I
had the goal of making the code multi-threaded (even on 16-bit Windows),
I also had to collect all of the global per-request data into a big struct
and pass its pointer around all over the place.

Anyway, it is what it is, and it does the job well for loads of people.

For my part: Rob, thanks.

-- Bob