Re: Code fails with Segmentation Fault

David T. Ashley wrote:
As the edit-compile-run cycle gets faster, people get more and more
careless.

No offense, but programs are also a metric fuckton more complicated
now.

I don't remember cc from 1982 having as many optimizations as GCC 4.1.

(of course I don't remember cc from 1982 at all as I was just born, but
I have used compilers from the mid 80s when I was ~11 or so)

While I agree that proper *design* and *auditing/verification* skills
should dominate the skillset of a competent developer, I don't think a
debugger is a hinderance. In short, when you're writing tens of
thousands of lines of code (often for free in my case...) *** happens.
A debugger can help you find a simple overflow/bad free/etc amongst
your 50,695 lines of code that make up, say, LibTomCrypt :-)

Similarly tools like valgrind help curb typos and what not by finding
uninitialized reads, memory corruptions, etc...

What I think you're observing as "careless" is simply people who are
not trained. All too often people who can merely "code" are considered
developers when in reality developers do a lot more than just code.
Design, test, and verification are as much as a developers job as
writing the lines of C [or whatever...]

Tom

.

Quantcast