cjsv@ccadfa.adfa.oz.au (Christopher JS Vance) (10/11/90)
We want an ANSI C compiler for our Pyramids. Unfortunately, Pyramid don't seem to be in a hurry to ANSIfy cc, and last time I checked, GNU C (1.37.1) admitted that <stdarg.h> uses a different parameter passing mechanism than the Pyramid compiler and therefore the library. Programming without printf can get boring :-(. Questions: * Are Pyramid planning to convert cc to be an ANSI compiler? Which version of OSx will this come with? * Has anybody fixed gcc to do the right thing (in -ansi mode)? -- Christopher Vance
bgg@pta.oz.au (Ben Golding) (10/12/90)
In article <1958@ccadfa.adfa.oz.au> cjsv@ccadfa.adfa.oz.au (Christopher JS Vance) writes: >We want an ANSI C compiler for our Pyramids. The C compiler that comes with OSx 5.1 understands ANSI C. Following AT+T's usage, there are flags -Kt and -Ka to select between transitional and ANSI C. Transitional mode accepts function prototypes and other ANSI extensions but implements K&R semantics where ANSI and K&R differ, for example, type promotions involving signed and unsigned types are handled differently. OSx 5.1 is available now. Ben.
cjsv@ccadfa.adfa.oz.au (Christopher JS Vance) (10/15/90)
In article <1958@ccadfa.adfa.oz.au>, I wrote: > Unfortunately, Pyramid don't seem to be in a hurry to ANSIfy cc, and I have since been informed, and quite politely :-), that OSx 5.1 comes with an ANSI compiler. I assume this means ANSI headers, libraries, lint, and the works. > last time I checked, GNU C (1.37.1) admitted that <stdarg.h> uses a > different parameter passing mechanism than the Pyramid compiler and > therefore the library. Programming without printf can get boring :-(. I have also been informed that such functions work provided you don't have a declaration in scope which uses the ellipsis notation. I assume this is illegal, even if it is the pragmatic thing to do. #ifdef time... -- Christopher