On Friday 02 September 2005 23:37, Sergey Plis wrote:
> Hi!
>
> I use huge amounts of data in my bigforth/xbigforth program. In order
> to do that I start xbigfroth with -m switch and tell it to allocate
> say 1G of memory. This makes whole idea of running .m files just as
> executable programs worthless. instead of doing
>
> ~# test .m
>
> on the command line I have to do
>
> ~# xbigforth -m 1G test.m

Or create a wrapper shell-script. I think 1G is too much; you can do 
something like 900M.

> Is there a way to allocate OS memory when xbigforth has already
> started? If not then may be you know how in bash pass parameters to
> the interpreter written in the first line of the script?

Probably the best way to do that is to create a special loader with 
different parameters built-in. The #! is limited to one extra 
parameter.

I think I'll write a program to patch the bigforth executable reliable, 
or to create a patched version on-demand without too much effort. 
Another possible way would be to create a special single argument to 
define all possible sizes (e.g. --sizes:m=1G,s=512k or so), to work 
around the limit of one argument.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

Attachment: pgpMYd59WBRbw.pgp
Description: PGP signature

Reply via email to