On Mon, 08 Jan 2007 22:17:12 -0600 (CST)
Al Hopper <[EMAIL PROTECTED]> wrote:
> On Tue, 9 Jan 2007, Roland Mainz wrote:
[...]
> But if you're trying to find a race condition and you recompile the code
> using different compiler options - you're no longer debugging the original
> code! This is the exact problem that DTrace is designed to solve.
>[...]
But if you recompile and the bug still exists then you have increased
the amount of information you have to solve a problem (and in mosts
cases I've run into "the" problem). If it doesn't reproduce then you've
cost yourself a systemwide recompile. In some cases this is a
reasonable cost/reward tradeoff.
I'm assuming this is done after somebody hasn't been able to figure out
the problem with dtrace or in parallel in an attempt to speed up
debugging. If its just a lack of familiarity with dtrace then I
suspect its more worthwhile to burn the recompile time on learning
dtrace as that will payback nicely in the future.
mph
_______________________________________________
opensolaris-code mailing list
opensolaris-code@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
