rfg@MCC.COM (Ron Guilmette) (03/27/89)
Recently, grunwald@garcon.cso.uiuc.edu (Dirk Grunwald) writes: >In article <8903252156.AA15405@prep.ai.mit.edu> cvms!ronald@cvms.uucp writes: > > I need some feedback for improving my attempt at COFF support for g++. > >...one design goal should be to use as few features of COFF as needed. >E.g., I would rather have to use `collect' than use the COFF .section >codes for ctor/dtor munging. The less stuff you need to change, the >better. Please Dirk... if you are going to make bold pronouncements of this sort, please give the rest of us some idea of the rationale you are using to decide what constitutes proper and good "design goals". Obviously, you believe that it is obviously "good" to "use as few features of COFF as needed". Why? Is COFF such an awful scalding nasty broth that we cannot occasionally dip into it, even if its use ocassionally provides simple and straightforward solutions to minor implementation problems? Also, I'm not sure I see that it is obvious that "the less stuff you have to change, the better". But even if I agree with that notion, I believe that the COFF patches for G++ actually change only a couple of dozen lines at most (the rest being fixes for other things, like putting more of the 32000 crt0 code in C rather than assembly code, and fixing the problem of using $'s in identifiers for systems which don't like them). Thus, only a small amount of G++ is being changed by using .section directives! I dare say that you may not be following you own guidelines. I'll bet that the changes to collect were more substantial that the .section stuff. I'm willing to be proven wrong though. Another very important issue which has not been discussed with regard to G++/COFF is debugging. I'd like to know if ANYBODY has tried using gdb to debug a G++ program on a COFF based system. Without gdb support, G++/COFF is very much less useful no matter how it is implemented. // Ron Guilmette - MCC - Experimental (parallel) Systems Kit Project // 3500 West Balcones Center Drive, Austin, TX 78759 - (512)338-3740 // ARPA: rfg@mcc.com // UUCP: {rutgers,uunet,gatech,ames,pyramid}!cs.utexas.edu!pp!rfg