ronald@UCDAVIS.EDU (06/04/89)
>>! if (rep->len > 0) >>! #if defined(i386) >>! memset(rep->s, 0, rep->sz * sizeof(short)); >>! #else >>! bzero(rep->s, rep->sz * sizeof(short)); >>! #endif > >This isn't the right way to do this, as i386 isn't the same thing as >USG. There are systems using the i386 processor which run BSD (at >least Sequent) and many many systems running USG which are not >i386-based. > >I don't know what #ifdefs should be used, since there doesn't seem to >be a standard USG define. Maybe one should just force users to >explicitly specify USG in their Makefiles and use #ifdef USG in the >g++-include files ? That doesn't seem a very good alternative. > >Anyone got any ideas ? That is exactly the problem with the include files. I was a little leery of doing it this way, but I expected other users to improve upon my donated work. I believe (but am not sure, therefore I didn't commit to anything) Sequent defines the symbol 'sequent'. It would be easy to change that line to read: #if defined(i386) && !defined(sequent) ... etc. should that assertion turn out to be true; and that is exactly what I expected users to do with my patches. The other alternative would be to provide two separate headers, one for BSD and one for System V; but that will most probably be an unacceptable solution. -- Ronald Cole | uucp: cvms!ronald voice: +1 916 895 8321 Senior Software Engineer | internet: csusac!cvms!ronald@ucdavis.edu CVM Systems +---------------------------------------------------- "No sex, please, we're software engineers." - Michael Swaine