SS> Is this the 1.0 version you are referring to, or does this refer to the
 SS> recently-released 1.5 version?  It's disturbing to hear of poor support

My own experience is with 1.0.  I don't like to comment on things I haven't
experienced myself,  but I have noted that some participants of this
conference stated that 1.5 did not fix some (if any) of the code-generation
bugs they had with 1.x.

With this in mind,  I didn't think Borland deserved an upgrade fee,  so I
didn't get 1.5.  I kinda feel that, for the us$250 I spent on v1.0,Borland
owes me (without further remuneration) a working compiler  --and they still
haven't delivered it.

 SS> from Borland (I've used their DOS compilers for years), but the pricing
 SS> (at least, from Indelible Blue) is $199 for BC++ 1.5 and $425 for CSet++.
 SS> That difference in price is significant... does IBM offer competitive
 SS> upgrades, perchance?

Since I made my statement about the prices,  I too went to look at the
street values (recently posted here).  I was wrong to say they were close
in price.  I do feal, however, that $100 spent on Cset++ FirstStep would be
better than $200 spent on Bc++ 1.5.


 SS> Is CSet++ 100% ANSI C compatible?  I'm careful to use ANSI C functions as

Yes,  even more so than Borland in my experience.  I too am mainly a C programmer.


 SS> As long as IBM has the ANSI eqivalents (fopen(), findfirst()), I'll be
 SS> fine (I hope).

findfirst() is not an ANSI function -- the fact that you didn't realize
this is another reason why I didn't like Borland's compiler;  They have too
many runtime functions that are too easily mistaken for
"standard" functions.  Many of the other compilers supply
equivalent functions, but they prepend them with an underscore.  _read(),
_dos_findfirst(),_splitpath(), for example.

IBM's compiler does not supply an equivalent of the Borland findfirst() or
MSC _dos_findfirst(),  and many other such low-level functions.

However, it is very easy to directly call the operating system's
DosFindFirst() function (and this is probably why they decided they needn't
bother coding it in the C runtime).   When you do this,  you're writing
compiler independent code (but OS specific :-().



--- Maximus/2 2.01