Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!samsung!caen!uflorida!mlb.semi.harris.com!rtpark.rtp.semi.harris.com!mcnc!rti!mozart!walker
From: walker@twix.unx.sas.com (Doug Walker)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: SAS C 5.10A has problems
Message-ID: <1991Jun21.152428.10283@unx.sas.com>
Date: 21 Jun 91 15:24:28 GMT
References: <mykes.3672@amiga0.SF-Bay.ORG>
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Distribution: sas
Organization: SAS Institute Inc.
Lines: 68
Nntp-Posting-Host: twix.unx.sas.com


Have you reported these problems to SAS/C technical support? 

I will summarize:

1. This is not a bug.  You are asking the compiler to
   use 16-bit offsets relative to an address register
   to access global data.  ("small data model")  You
   are attempting to refer to a data item that isn't
   accessible on this model - struct Custom is at a
   fixed location.  Therefore, you have to make an
   exception to the 'near data' rule - which you do
   with the 'far' keyword, exactly like you put it.
   If you really dislike 'far' for some reason, you
   can change the default with the -b0 option (You
   will have to link with different libraries.)

2. This is probably due to a known bug, although I
   can't tell without source code.  We elected not
   to fix it since the fix would probably generate
   other bugs and this one is relatively rare - this
   is about the fifth time anyone has reported it
   in the year that 5.10 and 5.10a have been out.
   Judicious use of the 'volatile' keyword is the
   workaround - contact tech support for details.

   BTW, the reason you get a CXERR 99 and not a
   more reasonable message is that CXERR's only
   come out when "impossible" situations

3. I don't know what this is, please contact tech
   support with additional details.  Is this the
   exact error message?  I've never seen it before.

In article <mykes.3672@amiga0.SF-Bay.ORG>, mykes@amiga0.SF-Bay.ORG (Mike Schwartz) writes:
|> Not to start a new round of language wars, but I have been working
|> with SAS 'C' for a few weeks (yes, I still write lots of assembler, too :)
|> 
|> Anyway, I've found that SAS is full of bugs and problems, many of which I
|> can find workarounds.  At this point, I would like to get (perhaps) some
|> answers from anyone on the net who might have solutions.  E-Mail might
|> be preferable, but if you think there is value to others in the solutions,
|> please post.
|> 
|> Problem #1:
|> 	BLINK complains about "extern struct Custom custom;"  It prints a
|> cryptic error message like "offset too far from A4".  I was able to fix
|> it by doing "extern struct Custom far custom;"  which may or may not be
|> preferable anyway (I say it is debatable :)  Many of the GfxMacros.h macros
|> rely on this custom variable.
|> 
|> Problem #2:
|> 	When compiling some sources, I get CXERR: 99, which is NOT documented
|> in the manual.  Nice messages, eh folks?  Real descriptive :)  The compiler
|> does not complain about anything else in the source file.  This error comes up
|> during the Optimizing phase.
|> 
|> Problem #3:
|> 	When compiling various sources, I get "Invalid section type in memory
|> option" error message.  I could find no documentation for this error message...
|>
--
  *****
=*|_o_o|\\=====Doug Walker, Software Distiller====== BBS: (919)460-7430 =
 *|. o.| ||                                          1200/2400/9600 Dual
  | o  |//     For all you do, this bug's for you!
  ====== 
usenet: walker@unx.sas.com       plink: dwalker       bix: djwalker