From: Gary Leavens Newsgroups: comp.os.msdos.djgpp Subject: Boehm-Demers-Weiser garbage collector (gc) with C++ Date: Mon, 09 Jun 1997 09:49:24 -0500 Organization: Department of Computer Science, Iowa State Univ. Message-ID: <339C17F4.1125@cs.iastate.edu> NNTP-Posting-Host: larch.cs.iastate.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Lines: 31 To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Precedence: bulk This is in response to my earlier postings about version 4.12alpha1 of the Boehm-Demers-Weiser conservative garbage collector for C and C++. I finally got our C++ program to work under DJGPP with gc. (Hooray!) There were 3 problems. One was that my makefile generated a line that was too long for the PC to handle (sigh) and so the linker failed, but the djgpp make still kept going, and so I didn't know I wasn't getting the latest exectuable to test. There was also a problem with some pointers. But the main problem was getting rid of the uses of malloc and calloc in other parts of the code, by changing them to calls to GC_MALLOC. I did find that explicitly setting GC_stackbottom led to unreliable performance under DJGPP in this C++ program. Since the gc system sets GC_stackbottom itself to what seems to be the correct valvue, I'd advise other users *not* to explicitly set GC_stackbottom. However, these problems may be caused by my use of a call to system("shell-command"). I'm attempting to reproduce this on small examples. Gary Leavens -- 229 Atanasoff Hall, Department of Computer Science Iowa State Univ., Ames, Iowa 50011-1040 USA / leavens AT cs DOT iastate DOT edu phone: (515)294-1580 fax: (515)294-0258 ftp site: ftp.cs.iastate.edu URL: http://www.cs.iastate.edu/~leavens/homepage.html