From: fredex AT fcshome DOT stoneham DOT ma DOT us Message-Id: <199607190133.VAA02703@fcshome.stoneham.ma.us> Subject: Re: go-32 and Linux DOSEMU To: sandmann AT clio DOT rice DOT edu Date: Thu, 18 Jul 1996 21:33:38 -0400 (EDT) Cc: djgpp AT delorie DOT com In-Reply-To: <31eb2242.sandmann@clio.rice.edu> from "Charles Sandmann" at Jul 16, 96 00:01:54 am Content-Type: text Thinking furiously, Charles Sandmann wrote: > > > big enough). I had gotten it to where it appeared to work, then I > > constructed a 6-mb file for testing, and found that sometimes seemingly > > benign commands caused a segmentation violation. Couldn't figure it out. > > But when I upgraded to the new CWSDPMI the entire problem went away. > > Probably the insufficient internal heap bug in CWSDPMI. The default size > was increased in r2 so you can now allocate up to around 20Mb of data in > small pieces before badness. You can use CWSPARAM to bump it more if you > need to edit 15Mb binary files :-) > It must really be a bug--I would expect malloc to return me a NULL pointer when it can't allocate any more memory, rather than a pointer which when dereferenced causes an exception. But I don't think I'll be doing too much editing of 15mb files. The whole point was to verify that I COULD edit files bigger than RAM with the ported BEAV. (I use beav moderately often at work on several Unix platforms as well as DOS, and some of the files are significantly larger than DOS' lower memory, hence the desire for the port. An added benefit was that since I used the libc function ScreenPutString() to write all the output onto the screen, the program SCREAMS!) Fred -- ------------------------------------------------------------------------------- .---- Fred Smith / Office: fred AT computrition DOT com ( /__ ,__. __ __ / __ : / 508-663-2524 / / / /__) / / /__) .+' Home: fredex AT fcshome DOT stoneham DOT ma DOT us / / (__ (___ (__(_ (___ / :__ 617-438-5471 -------------------------------- Jude 1:24,25 ---------------------------------