NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/lang/pascal/7635

Text File | 1992-12-22 | 1.7 KB | 34 lines

Newsgroups: comp.lang.pascal
Path: sparky!uunet!gatech!paladin.american.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!cs.utexas.edu!torn!news.ccs.queensu.ca!slip201.telnet1.QueensU.CA!dmurdoch
From: [e-mail] (Duncan Murdoch)
Subject: Re: 64K segments
Message-ID: <[e-mail]>
Lines: 22
Sender: [e-mail] (Netnews control)
Organization: Queen's University
References: <[e-mail]>
Date: Tue, 22 Dec 1992 22:52:16 GMT

In article <[e-mail]> [e-mail] (Mike S. Butterbrodt S92) writes:
>Probably a FAQ.   Was mentioned not long ago that there are way (painfull or
>otherwise) to get around the 64K segment limit in BP 7.0.  Ok, so how.  I would
>appreciate a few tips on this most annoying feature of an otherwise nifty
>language.

The same as in previous versions:  if you want a total allocation that's
bigger than 64K, break it up into pieces that are smaller than 64K.  For
example, for a 100 x 100 array of Doubles (80000 bytes, too big), use an
array of 100 pointers to arrays of 100 doubles (400 bytes for the pointers,
800 bytes for each array).

The advantage under BP 7 is that in protected mode the heap will be much
bigger; 1000 x 1000 arrays might be feasible.  This would be on a machine
with lots of ram, or under a DPMI server (like Windows or QDPMI, but not
the one bundled with BP 7) that offers virtual memory.  I haven't tried
pushing the allocation limits; you might well run out of some other resource
before you hit the RAM limit.  For instance, there's a system wide limit of
8192 selectors; 1000 large allocations would make a significant dent in that.

Duncan Murdoch
[e-mail]