Re: TRACE with VS6 onVista

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

See below...
On Wed, 20 Jun 2007 09:24:11 +0900, "Norman Diamond" <ndiamond@xxxxxxxxxxxxxxxx> wrote:

"Les" <l.neilson@xxxxxxxxxxxxxxxxxxx> wrote in message
news:f58p51$1hrd$1@xxxxxxxxxxxxxxxxxxxx

"MrAsm" <mrasm@xxxxxxx> wrote in message
news:9smf735mp5gi141jsdkg37vop1jjsosdnh@xxxxxxxxxx
On Tue, 19 Jun 2007 06:08:10 -0700, "Tom Serface"
<tom.nospam@xxxxxxxxxxxxx> wrote:

I remember back in the day when mistakes cost way more time. One wrong
statement required a wait until the next day when a new deck of cards
could be compiled and run. Modern day compilers and IDE's are so
incredible...

I imagine that programming in the "old days" of punch cards, without
IDEs, without debuggers..., must have been a very *hard* job! The "old
days" programmers must be very smart people, IMHO.

Oh, we had debuggers. Dumps were printed on hex in about 200 pages of dead
trees (give or take a few orders of magnitude depending on the program) and
we had disassemblers sitting on comparatively more solid constructions from
dead trees.
*****
The day we added our 8MB of extended memory to our /360-65 mainframe, we had a program
fail. Nobody noticed the fundamental error in the OS until it started into the second box
of printer paper as it printed the dump.

The OS tried to print the entire 8MB of memory, 64 bytes per line; 2184 pages (there were
2000 pages in a box, as I recall)
*****

Hello, someone talking about me? :-)
I well remember IBM manuals for OC7 errors (in COBOL code) saying
something like "You have an error in yor code. Correct the error and
re-run the program"

Well, that compiler was particularly malicious. I tried to solve it by
comparing the field to blanks before comparing it to zeros. For the
comparison to blanks, the compiler generated an OR instruction to change the
non-numeric sign nibble to a plus sign so the last character would no longer
be a blank and the comparison would fail. Then the program would proceed to
compare to zeros, and the compiler would not generate the same OR
instruction, so the last character had a non-numeric sign nibble and
exception 0C7 crashed the program. Correct the compiler and re-run the
program? Nope, it was closed-source, written by a formerly famous former
monopoly which no one could fix, so we had to add workarounds to our
programs.

More fun was seeing IEBGENER abend with 0C4. IEBGENER was another
closed-source program written by a formerly famous former monopoly which no
one could fix.
****
My first compiler was a summer project in a course I took on programming the /360 in
assembly code. It compiled a rather simplistic Algol-like language. As a member of that
class, I got the unheard-of privilege of running TWO runs per day! All my error messages
started out as IFU, e.g., IFU00001, IFU00002, etc. My instructor asked me if IFU stood
for what he thought it meant, and I said, "Of course". He laughed, and I said "What's so
funny about "I Fail to Understand"? (Actually, it DID stand for what he thought it stood
for, but he wouldn't get me to admit that...) [This was in the summer of 1967; I first
started coding it 40 years ago this month...kind of scary!]

The original semester project was boring, so I asked the instructor if I could do a more
challenging project. He asked me what I wanted to do, and I said I wanted to write a
compiler. He seemed amused, but told me to go ahead and try, not expecting anything would
come of it. I wrote a compiler and a runtime system, all in assembly code, never having
seen the architecture of the machine, in eight weeks. He'd never seen anyone do that
before. I still have its punched cards somewhere...This was my first course in CS.

Given how much code I wrote for the /360, I'm always amazed that I only used it from
1967-1970. By 1970 we had the PDP-10.
joe
*****
Joseph M. Newcomer [MVP]
email: newcomer@xxxxxxxxxxxx
Web: http://www.flounder.com
MVP Tips: http://www.flounder.com/mvp_tips.htm
.

Relevant Pages

  • Re: Pedants
    ... However Jacob publishes only his binaries. ... I would be interested to see what you think you are comparing against ... were compiled by the same compiler program and with identical compiler ... PUSH ...
    (comp.lang.c)
  • Re: optimized code
    ... > All depends on what exactly you are measuring, how you are comparing and ... > take advantage of the back-end optimizer which can do a better job than ... > produced by the C++ compiler, ...
    (microsoft.public.dotnet.languages.csharp)
  • Re: 64-bit on the horizon? (Was Re: Vista Requirement Already)
    ... Win64 code is very often slower than win32 because the size of the ... by 1.3 thought the i386 compiler contains more code because ... you be comparing cross-compilation results? ... So I still don't see how you conclude that memory footprint ...
    (borland.public.delphi.non-technical)
  • Re: gcc vs. Sun Workshop/Forte / was: Re: Can someone do us a favour? Need core trace back.
    ... > up to him today by finding out why our AIX compiler was failing). ... > to the more recent versions of gcc. ... > Not much use for comparing to Forte but interesting that you can get ... The Mozilla.org default for Sun Workshop is "-xO4" but I reduce that to ...
    (comp.unix.solaris)
  • Re: TRACE with VS6 onVista
    ... I tried to solve it by comparing the field to blanks before comparing it to zeros. ... the compiler generated an OR instruction to change the non-numeric sign nibble to a plus sign so the last character would no longer be a blank and the comparison would fail. ... Nope, it was closed-source, written by a formerly famous former monopoly which no one could fix, so we had to add workarounds to our programs. ...
    (microsoft.public.vc.mfc)