home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!spool.mu.edu!umn.edu!staff.tc.umn.edu!wright
- From: wright@staff.tc.umn.edu (Mark Wright)
- Newsgroups: comp.os.msdos.programmer
- Subject: Phar Lap 286-lite problem
- Message-ID: <1993Jan7.005122.28069@news2.cis.umn.edu>
- Date: 7 Jan 93 00:51:22 GMT
- Sender: news@news2.cis.umn.edu (Usenet News Administration)
- Organization: University of Minnesota
- Lines: 34
- Nntp-Posting-Host: staff.tc.umn.edu
-
- Because of some serious memory problems, I've been fourced to try out that
- stupid LITE286 thing that came with Borland 3.1, and I've been having the
- following problem.
-
- I have 3 files and a library I need to link together. The files are
- 1533 LSTSQRS\LSQ.C
- 554 LSTSQRS\LSQIO.C
- 572 LSTSQRS\MATRIX.C
- 2659 total
- where the number is the number of lines in each. The library is one I've
- built using bcc286 and tlib. When I do a plain old real mode
- "bcc -ml lsq.c lsqio.c matrix.c mwl.lib", everything is just hunky-dory.
- However, when I try to "bcc286 lsq.c lsqio.c matrix.c mwe.lib", I get a
- DGROUP >64K error. The lite286 docs say that I should adjust the stack size,
- but that doesn't help, no matter how small I make the stack.
-
- So what's the problem? Can I use tlib with protected mode .obj's? Do I have
- to spend $500 to get the real Phar Lap extender, or will I have the same
- problem with that? If I have to buy the sdk, should I go with Phar-Lap?
- All I want is to be able to malloc() > 640k worth of long doubles (total, not
- necessarily in one block), and not have to make any changes in my existing code,
- if possible. Oh, and it's gotta run under windows (so no DJGPP, sigh... of
- course, I'd have to re-write all the dos/borland specific stuff if I could use
- djgpp).
-
-
- --
- Mark Wright
- wright@epx.cis.umn.edu
-
-
- --
- Mark Wright
- wright@epx.cis.umn.edu
-