On Mon, 18 Feb 2013 08:23:59, Ruud Harmsen <firstname.lastname@example.org> wrote:
> Sun, 17 Feb 2013 10:16:12 -0500: Shmuel (Seymour J.) Metz > <email@example.com> schreef/wrote: > > >In <DmJ5SKFdRQph-pn2-Iu4uIsOT7cBR@209-142-179-216.dyn.centurytel.net>, > >on 02/14/2013 > > at 01:46 PM, firstname.lastname@example.org (Will Janoschka) said: > > > >>Then your thread safe code cannot call itself from within itself > > > >What gives you that idea? A procedure that is recursive and reentrant > >will generaly have local data on the stack, but that has nothing to do > >with making copies of other data. > > The call arguments are pushed onto the stack before the actual call, > and that is a copy action.
And the return info places are also reserved on the stack The caller is allowed to transfer the results globally not the procedure, else the called procedure is not reentrant, This is the definition.
Each instantiation must be distinct and independent. Most of this is not in the realm of Computer Science. It is in the realm of scraping bits from the floor and doing that correctly..
The difference between Recursion and Reentrant is both minor and major but always difficult. Recursion is a repetitive process, Reentrant is used where all is chaotic. In an air to air missile that must self destruct 12 seconds after launch chaotic is all there is. No time for properly structured software. > >