wampler@unmvax.unm.edu (Bruce Wampler) (10/15/88)
I've seen several postings that continue to repeat the common misconception that QuickC supports only the Medium Memory model. This simply is NOT TRUE. I think this comes from the fact that the integrated environment supports only the Medium model. But, that environment is a real hinderance to any real development. QuickC supports a command line sequence (qcl) that supports almost all the switches that 5.1 does (cl) (except the optimization switches and a couple of other small features.) I find I can build almost any size/model program using qcl for debugging, then build the final version in cl. This is a very useful combination for serious development work. Bruce Wampler Reference Software, Inc.
jlh@loral.UUCP (Physically Pffft) (10/21/88)
In article <2041@unmvax.unm.edu> wampler@unmvax.unm.edu (Bruce Wampler) writes: >I've seen several postings that continue to repeat the common >misconception that QuickC supports only the Medium Memory model. >This simply is NOT TRUE. > >QuickC supports a command line sequence (qcl) that supports almost >all the switches that 5.1 does (cl) (except the optimization switches >and a couple of other small features.) I find I can build almost >any size/model program using qcl for debugging, then build the >final version in cl. This is a very useful combination for serious >development work. > I hate to disagree, but after getting bit by 2 bugs in qcl's large model I deleted it from my disk. I also seem to remember reading in the manual at the time that Microsoft (do they object to the name usoft?) said it was for medium model programming only, integrated environment or not. Both bugs I found had to do with segmentation problems, if I wasn't so familiar with 8086 programming I would have never found the problem. Sorry I can't be more specific but it's been 6 months now. I should mention that this was version 5.0, if the large model works in 5.1 I'd love to hear about it. Please don't say "My sisters third cousin removed uses it for everything", I just want to hear "I've used quick C version 5.1 large model for months now with no problems". Jim -- Jim Harkins jlh@loral.cts.com may work. Loral Instrumentation, San Diego {ucbvax, ittvax!dcdwest, akgua, decvax, ihnp4}!ucsd!sdcc6!loral!jlh
wampler@unmvax.unm.edu (Bruce Wampler) (10/23/88)
> I should mention that this was version 5.0, if the large model works in > 5.1 I'd love to hear about it. Please don't say "My sisters third cousin > removed uses it for everything", I just want to hear "I've used quick C > version 5.1 large model for months now with no problems". > Well, a follow-up to my original article. I've developed a VERY large commercially sold program (Grammatik III, to be specific). It uses large model. We do almost ALL of the development in QuickC, including testing, etc., etc., and have not detected ANY difference in operation other than .EXE size and speed between QuickC and regular MSC - this is true for both 5.0 and 5.1. The ONLY problem we've found in with the QuickC /Zr switch (generates checks for pointer usage), which does not work for large model. I've also done this for a couple of other big programs. Not only is QuickC faster for development, it works with CodeView. We've been plenty happy with MSC since 3.0, so I may be out of touch with the competition these days, but MSC 5.1 seems to generate very good code as long as you don't use /Gs /Oait for optimization. Bruce Wampler Reference Software, Inc.