Re: Hacker FAQ (please comment and help fix)

Article: 7666 of alt.hackers
From: spc@news.gate.net (Sean 'Captain Napalm' Conner)
Newsgroups: alt.folklore.computers,alt.hackers
Subject: Re: Hacker FAQ (please comment and help fix)
Date: 15 Apr 1995 02:22:33 -0400
Organization: Cybergate, Inc.
Lines: 63
Approved: Certified to comply with the limits for a Class B computing ...
Message-ID: 3mnon9$22im@hopi.gate.net
NNTP-Posting-Host: hopi.gate.net
Status: RO

In article <3mng38$e3k@solutions.solon.com> seebs@solutions.solon.com
(Peter Seebach) writes:
>
>Does anyone have any suggestions for how you tell Ken Thompson from Bill
>Gates, in the general case?
>
  Now, hey there ... Bill Gates wasn't that bad of a programmer (when he
actually did programming) and was able to fit quite a remarkable amount of
code into a 4K space (and still leave some free).  And one could say that he
is successfully able to hack management and deals as well.

  Still not sure if I like the guy myself, but he does keep you on your
toes.  But I'm not here to sing the praises of Mr. Gates (heavens no).

  But, to your question:  How to determine the Good Hacker from the EL33TE
D00DZ ... unfortunately, it requires programing ability on the part of the
manager, but asking the hacker to provide sample code he's written and
explain why he wrote it and how he wrote it might be a good idea.  I don't
know of a single Good Hacker that won't jump at the chance to show off his
code.

Oh, well, for alt.hackers ...

ObDebuggingUnixProgramFromHellHack:

  At a recent job, I had to port a Unix editor (which shall remain unamed,
since otherwise it's a nice editor - but it's not Emacs) to QNX.  Well,
Watcom C is very very fussy about ANSI C conformity, which this editor falls
way short of.

  So, I ended up running the editor under the full screen debugger, finding
out it's trashing the frame pointer register spradically (but I was able to
restore the correct value in the register and keep the editor going).  I
finally tracked it down to several functions defined as:

int foo(int p1,int p2,int p3,int p4,int p5);

  and called (very liberally) as:

	foo(x,y);
	foo(x);
	foo(x,y,a,b);

  which Watcom C accepted and compiled, but the code was less than optimum.
Unfortunately, one of the first things foo() did was:

int foo(p1,p2,p3,p4,p5)
{
  int a1 = p1;
  int a2 = p2;
  int a3 = p3;
  int a4 = p4;
  int a5 = p5;
 ...
}

  And then it got cryptic (it didn't help this was a state machine to handle
terminal emulation).  So, I punted and recoded all calls to foo() with the
correct number of paramters.

  God, I hate Unix code.

  -spc (But that's besides the point ... )

Child

Back to index