home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!usc!cs.utexas.edu!rutgers!cbmvax!jesup
- From: jesup@cbmvax.commodore.com (Randell Jesup)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: SAS/C 6.0 Details (LONG!)
- Message-ID: <34583@cbmvax.commodore.com>
- Date: 23 Aug 92 22:26:42 GMT
- References: <Bt8H7F.BL8@unx.sas.com> <BsxCu2.5A7@unx.sas.com> <rwm.714413035@atronx.OCUnix.On.Ca> <mwm.1lm3@contessa.palo-alto.ca.us>
- Reply-To: jesup@cbmvax.commodore.com (Randell Jesup)
- Organization: Commodore, West Chester, PA
- Lines: 41
-
- In <rwm.714413035@atronx.OCUnix.On.Ca>, rwm@atronx.OCUnix.On.Ca (Russell McOrmond) wrote:
- > I got my upgrade notice last night. I need to be convinced that this
- > upgrade is actually worth another $109+shipping for myself. The
- > 'tagcall' pragma is about the only thing that hit me as interesting, and
- > I can't think that it's worth $109 ;-)
-
- It's faster. It produces smaller, faster code (noticably, sometimes
- very noticably). The function inlining is nice (and it's flexible, too).
- The compiler no longer ever needs to generate link instructions (it always
- knows the offset off a7 for arguments, etc), so you get an extra pointer
- register. You can tell it all your code will be -b0 (data absolute-addressed)
- and you get a4 back as a pointer register also. CPR is much improved, and
- xCPR is handy when you have two machines (it's even possible to debug across
- a network). It can produce ASM output that can be fed directly to an
- assembler. The GST stuff can make compiles far faster, and can be helpful
- when debugging (since you can query it). It can add "coverage" code to help
- in making sure you tested all cases.
-
- I'm no slouch at tight assembler, but I can only rarely find
- places where I could improve on the compiler's output (given the contraints
- mandated by C).
-
- Totally new (and better) documentation, including extensive AmigaGuide
- hypertext docs (really handy in CPR).
-
- The new error messages are a vast improvement. If something is
- redefined, it tells you where, and where the original definition was. On
- a type mismatch, it tells you what it got and what it expected. Those
- are really helpful. In general, the error messages are no longer cryptic
- at all; they're quite readable.
-
- The stack checking now knows exactly how many bytes the routine will
- take, instead of generic value as in previous versions.
-
- --
- "Rev on the redline, you're on your own; seems like a lifetime, but soon it's
- gone..." Foreigner
- -
- Randell Jesup, Jack-of-quite-a-few-trades, Commodore Engineering.
- {uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.cbm.commodore.com BIX: rjesup
- Disclaimer: Nothing I say is anything other than my personal opinion.
-