I am curious to know what you made of my suggested improvements, but
not curious enough to email you about it. If you think Seebs is a
troll, you don't know Seebs very well.
To be fair, I certainly have had a great deal of fun in epic newsgroup
trolling festivals, such as my attempt to convince people that a solar
eclipse is when the sun passes between the earth and the moon.
In this case, though, no trolling. I just think the guy's ineducable.
The corrections about topicality were quite correct, and Kelly's
excuses are implausible. If he were really picking groups based on
implementation language, it could be taken as an honest mistake -- but
when a casual search turns up announcements posted to groups like
comp.lang.ruby, it becomes clear that there is no such honest mistake
involved, but garden variety spam.
His responses to criticism are precisely those which show someone who
is willing to take only certain kinds of advice -- that which fundamentally
supports his positive self-image, which is usually that least useful
to him.
The fact is, the program in question is... well, frankly, pointless. He's
very actively fixing bugs that probably shouldn't have existed -- and for
that matter, which might even not be bugs, but design choices. The thing
is, he can't talk about that, because there's nothing to show that he's
got any kind of sense of how to design a program. It's full of cargo-cult
madness. Some of it's fairly carefully planned; some of it looks like it
was thrown together by someone totally uninterested in understanding
the environment.
In short, it's exactly what you'd expect from someone who can't handle
being told about newsgroup topics, and who insists on making up patently
false excuses for why he's posting in a particular group.
There's a lot of stuff here which could, conceivably, be interesting.
Very little of it has anything to do with C; if we were in a Unix group,
and someone posted this, I'd probably devote a longish post to discussion
of some of the design decisions and tradeoffs. For C, there's not nearly
as much; nearly all the functionality of this program is purely specific
to Unix-like environments. What little I can say in a C context is that
there are too many variables which are used in too many ways, and there's
some very repetitive blocks of code which are probably incorrect.
But, fundamentally, this is not a program that can be discussed coherently
in terms of the C language. The C style is poor but not unbelievably so,
but the substantive flaws are in the Unix-specific stuff, which just plain
isn't topical here. Ninety percent of what I think this program does wrong,
you could get wrong in exactly the same way using perl on Unix but couldn't
do in C on Windows.
-s
