SHELL commands

[Robert E. Carneal]

Since SHELL loads a second copy of the Dos command processor
temporarily to execute a command/program from Dos, and then
return to the program that contained the SHELL command, I am
beginning to be concerned my programs will run out of memory
before finishing.

I write tiny program (easier to debug, and usually less than
35K), some of my "main" programs will have 100s of SHELL
commands. Is there a more efficient way to do this? I have not
found a method to determine how much memory my programs
eventually need.

Thank you.

Robert Carneal

------------------

[Mel Bishop]

Originally posted by Robert Carneal:
I have not found a method to determine how much memory my
programs eventually need.

FRE(-1) tells you how much memory you have available for data
and other stuff. Example:

row = csrlin : col = pos(0)
locate 25,1
print;fre(-1);
locate row,col

Check the help file for other uses of FRE().

In as far as "100's" of shells, it all depends on the size of
the files and how much data they generate. One possibility you
may want to try: Instead of one big MAILFILE.EXE, try chopping
it up and CHAIN each file on an as-needed basis.


------------------




[This message has been edited by Mel Bishop (edited April 20, 2003).]

[Lance Edmonds]

Another solution would be to stick with the SHELL methodology, and use UltraShell from www.infoms.com - you include a small INC file and after compiling to an EXE, run a utility that patches the executable. Once the app runs, the SHELL statement swaps out all but about 5kb of conventional memory to EMS/XMS or a disk file, leaving almost the entire conventional memory area available for the SHELLed app. On return, memory is reloaded, and the original app continues along where it left of.

Very smooth. It would have to be the best library/utility I ever invested in for PB/DOS development.

------------------
Lance
PowerBASIC Support
mailto:[email protected][email protected]</A>