rab@murdu.oz (Richard Alan Brown) (11/13/89)
In article <17386@kuhub.cc.ukans.edu> BROWNRIGG@kuhub.cc.ukans.edu writes: >Anybody out there have any experiences with AOS/VS C? > We have had AOS/VS C running at the University of Melbourne for about 5 years. >In particular: > >-Is it a solid, stable product? > Yes. This, I might add, in in stark contrast to the AOS/VS F77 compiler, which has given us nothing but trouble from the start. In fact we have three versions on our system at present, because they ALL have bugs: it's just a question of using the ones that work for the piece of code you need to compile!! >-Is it closer to ANSI standard conforming than the typical Unix compiler? > Less conforming? > Any portability problems to note? Prototypes are not yet supported. The include files structures are not standard. A typical problem is that using, say, 'atof()' works fine without #include'ing <ctype.h> (I think that's the one), but of course this gives you rubbish on another machine (a Sun, for example). > >-Are the run-time libraries reasonably complete, and not wanting? > You don't say whether you have MV/UX or not. One of the biggest problems is porting UN*X code. To be fair, DG have done a good job here, making AOS/VS look like unix, but there can be some *real* headaches. Several include files do not exist, and of course there is no analog for many unix functions. Fork() does not handle file descriptors properly. If you do not have MV/UX or do not want to port unix code, you shouldn't have any problems. >-Efficient execution, good optimization? > Yes. Richard Brown (pbrown@munda.ph.unimelb.edu.au)