1. RE: EU2.3, sanity.ex, and NT4.0
Rob:
Neither DJGPP nor CYGWIN programs have any problems starting up new
shells on this system, regardless of whether the shell is COMMAND.COM,
CMD.EXE, or BASH.EXE. In fact, if I replace
system( "del sanityio.tst", 2)
with
system( "find sanityio.tst -type f -exec rm '{}' ;", 2),
the file is successfully deleted, probably because 'rm,' the executable
that DJGPP 'find' spawns to delete the file, does it differently than
the builtin DEL of CMD.EXE and COMMAND.COM. In fact, if I replace 'rm'
above with 'cmd /c del' or 'command.com /c del,' the error re-appears.
The DPMI (which is the one that comes with NT) doesn't have any swap
space available, probably because the machine has 256MB, and the most
DPMI memory that can be made available to a console is 64MB (65535KB),
which probably (mistakenly) leads CMD.EXE and COMMAND.COM to think that
none is necessary. I'm guessing that the builtin DEL command checks for
swap memory before running the command and balks, while DJGPP programs
take memory any way they can get it, and don't die until they can no
longer do so.
I'll have to keep this in mind when spawning subprocesses on my NT
machine, at least until I figure out how to change the DPMI layout.
Once again, the third-party code (EU in this case) is exemplary, and the
underlying system is weak. Oh, well ...
Finally, thanks a _million_ for the help, Rob!
Thx, Phil Long
Robert Craig wrote:
> Phil Long writes:
> > Euphoria SANITY TEST ... Insufficient memory
> > PASSED (100%)
>
> It sounds to me like maybe the call to system()
> in sanity.ex:
> system("del sanityio.tst", 2)
> is giving you that error message, due to some problem that you
> have in starting up a new DOS shell.
> Just a guess. You might try commenting out that line.
>
> Regards,
> Rob Craig
> Rapid Deployment Software
> http://www.RapidEuphoria.com
>
>
