Subj : Re: was: chess, changed to WHICH
To   : Rolf Jentsch
From : Jasen Betts
Date : Wed Nov 21 2001 10:44 pm

Hi Rolf.

20-Nov-01 10:47:00, Rolf Jentsch wrote to Jasen.Betts@xspamp42.f531.n640.z3.f


 RJ> From: email@rolfjentsch.de (Rolf Jentsch)

 RJ> Jasen.Betts@xspamp42.f531.n640.z3.fidonet.org meinte am 18.11.01
 RJ> zum Thema "Re: was: chess, changed to WHICH"

 RJ> Hello Jasen,

 JB>> Using %comspec% /c works but you lose any modifications done to
   >> variables (unless you pipe them back, or store them in a file)

 RJ> Sure! The same should go for a call, shouldn't it?

it depends what you mean by "a call",
4dos's CALL command doesn't isolate the variables
but it could be argues that that isn't really "a call"
using setlocal in the called batch will isolate the variables.

 >> if you want to use gosub recursively you have to do-without local
 >> variables relying instead on the parameters to hold that
 >> information. I think recursal depth is limited to about 16 levels.
 RJ> Local variables can be used locally. And the recursal depth is of
 RJ> course limited to the nesting level of gosubs - that is around 15
 RJ> or 31 (which I do not remember right now exactly

 >> if you maust have local variables they can be faked by using a
 >> variable variavbe name but you have to remember to clean up, or
 >> accept junk in the environment...
 RJ> Thats the way Rex implemented them. If you debug a file with local
 RJ> variables in gosubs you'll see that the local variable x is
 RJ> declared as chr(01)x for the first nesting level and chr(02) for
 RJ> the second a.s.f

they don't really seem to be local variables - they can't easily be
modified (I can't work out how to modify them)

 -=> Bye <=-

---
 * Origin: Those who can't write, write manuals. (3:640/531.42)

.