setjmp/lonjmp (was: vinum warning)

Joerg Sonnenberger joerg at britannica.bec.de
Thu Feb 3 15:01:49 PST 2005


On Fri, Feb 04, 2005 at 09:14:25AM +1030, Greg 'groggy' Lehey wrote:
> On Thursday,  3 February 2005 at 23:31:22 +0100, Joerg Sonnenberger wrote:
> >
> > The longjmp debugging is not needed soon, I'm working on removing
> > the use of longjmp completely.
> 
> I'm interested in the general aversion to setjmp()/longjmp().  In my
> opinion they're superior to the alternatives.  What objections do
> people have to them (NIH does count :-).

They destroy the normal flow of code. Even worse, they allow jumping
out of the current flow to a different stack.

Joerg





More information about the Kernel mailing list