ranson@crcge1.UUCP (D. Ranson CNET) (08/05/87)
If you are a Aztec C user, and a MacSE/II owner, you may have noticed that
current release 1.06i is not really compatible with those new machines. The
assembler will generally go south when it finds errors in your code. The
linker will sometimes (I haven't found a pattern yet!) decide to go on strike.
I was told that the new 3.4 (??) version was on the way. I hope it's out soon.
Since I could use Azte for useful work, I spent some time looking at libraries
or NOSYfying some code. I have found some things that may cause problems to
all applications compiled with Aztec C that will run on new machines:
- The "newrom" library function tests EQUALITY of ROM85 with 0x7FFF. This
means new machines will be seen as 64K ROM machines. For all traps that were
not present in 64K ROM, you'll be executing glue instead of the trap.
- In the initialisation code of applications (sacroot and such), there are
several instances of illegal fiddling with handle bits, like masking the
master pointer with 0xFFFFFF to get the block address (why would they ever
need to do that?).
I have seen postings with warnings about LightSpeed C, so I think these Aztec
C problems should be known. I was a happy Aztec user until I got my SE. No
I'm using LsC, until I get an upgrade...
Daniel Ranson
...!seismo!mcvax!inria!{crcge1 or cnetlu}!ransonbradn@tekig4.TEK.COM (Bradford Needham) (08/06/87)
As an SE owner, I gave up on Aztec C when I found it wouldn't work under system 4.1. It was the last straw in a pile of minor incompatibilities. It seemed to me that Manx hadn't put any support into Aztec C for over a year, so I switched to a less-expensive, better-supported, more-productive compiler -- Lightspeed C. Now that I've used LSC for a while, I don't know why I didn't make the switch earlier! Brad Needham bradn@tekig4.TEK.COM