Re: Sun compilers: acc vs. gcc?

harold@forsythe.stanford.edu (Harold Finkbeiner)
Mon, 3 Oct 1994 21:01:53 GMT

          From comp.compilers

Related articles
Sun compilers: acc vs. gcc? cooper@holly.imsi.com (1994-10-01)
Re: Sun compilers: acc vs. gcc? harold@forsythe.stanford.edu (1994-10-03)
Re: Sun compilers: acc vs. gcc? rfg@netcom.com (1994-10-05)
Re: Sun compilers: acc vs. gcc? J.C.Highfield@loughborough.ac.uk (1994-10-06)
Re: Sun compilers: acc vs. gcc? khb@Eng.Sun.COM (1994-10-07)
Re: Sun compilers: acc vs. gcc? cord2403@cslabs2c4.engr.ccny.cuny.edu (1994-10-20)
| List of all articles for this month |

Newsgroups: comp.compilers
From: harold@forsythe.stanford.edu (Harold Finkbeiner)
Keywords: C, sparc, GCC
Organization: Stanford University
References: 94-10-009
Date: Mon, 3 Oct 1994 21:01:53 GMT

>Can anyone offer a comparison of Gcc vs. Acc, sun's ansi compiler? I'd
>like to know if there is any reason to buy Acc, if I'm happy with Gcc.


I was using the acc compiler (SC1.0.1) and switched over to gcc for a
couple of reasons. The gcc compiler has better error messages and with
-Wall it covers a broader range of circumstances I like to be warned
about. With gcc you have better control over what the compiler does.
If you have to debug a program :-), I prefer gdb over dbx (was well as
xxgdb over dbxtool).


I found that acc was giving me messages because the header files as
defined by Sun do not follow the ANSI standard (memcpy and memcmp
should use void * but Sun defines them as char *).


I would assume that the acc compiler would be better at code
optimization (especially with floating point operations) but I have
never tried it out.


You might look into SPARCworks. With the new version it sounds like
you can debug/edit code on the fly similiar to an interpreted version
of C. I believe though that you require at least Solaris 2 for the new
version which has this feature.


Have a nice day,
            -Harold


Harold Finkbeiner Email: harold@forsythe.stanford.edu
Stanford University Phone: (415) 725-3353
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.