Mail Archives: djgpp/2000/03/21/02:32:17
Eli Zaretskii wrote:
> On Sun, 19 Mar 2000, Kalum Somaratna aka Grendel wrote:
> > So it follows that on a p-166 your 10000 line program would take about 10
> > secs to compile. a long time indeed.
> You are kidding, right? Because 10 seconds is not even enough time to
> reflect upon the bugs you might have left lurking in the code you're
> compiling...
>
> Large projects take tens of minutes, sometimes hours, to build. Now
> *that* is a long time indeed. Whereas 10 seconds--well, it's just a
> blink ;-).
Some of use are spoiled though. What I usually do is compile, get a list of x
compiliation errors, fix the first one, then compile again. I have found that
in about 10% of cases, the initial error in the code confuses gcc so it 'make
up' some of the other errors. This means that I sometimes compile code every
10 seconds. In general, I believe my effort is more valuable than my CPU's,
so even if there is only a small chance that the errors will dissapear, I
still recompile.
The point being, if my code doesn't compile in < 2 seconds, I look for
ways to speed it up. I realize that, beyond some limit, there is nothing I
can do, and if most of the time is spent in the linker stage, there's not much
I can do, but 10 seconds is a long time for most of us.
--
(\/) Rolf Campbell (\/)
- Raw text -