On Sat, 18 Nov 2006 20:12:23 +0000, john gilmore <[EMAIL PROTECTED]> wrote:
>
> Moreover, it must be conceded that analogues of such PL/I constructs as
>
> declare phantom_cs character(32767) based(pcsp) ;
>
> pcsp = addr(<whatever>) ;
>
> and
>
> addr(<whatever>)->phantom_cs . . . ;
>
> which permit any sequence of [here at most 32767] storage locations to be
> viewed as a character string, are now grudgingly available in C too; but the
> much more commonly used C asterisk notation declares a pointer to an
> instance of a particular scalar or structure. C is more much strongly typed
> than PL/I; and Dennis Ritchie, its designer, has said more than once that he
> views this as desirable.
>
Unfamiliar with PL/I as I am, I see little practical difference between the
above and the equivalent (as far as I understand the PL/I):
typedef char[32767] phantom_cs;
...
... (phantom_cs) &whatever ...;
That said, I loathe C's reliance on null-terminated strings. (Some claim,
with some validity, that this is more a feature of the Standard Function
Library than of the language proper. But, practically, the two are
inseparable.) This makes it impossible to abstract a substring from
a static string without allocating storage, diminishing the usefulness
of the constructs in the example. The null-termination convention has
cross-infected the programming interfaces to most C kernels. In contrast, the
BPX1* callable interfaces have boldly and laudably eschewed null-terminated
strings.
So, in answer to Bob Shannon's question:
Linkname: IBM-MAIN archives -- August 2006 (#278)
http://bama.ua.edu/cgi-bin/wa?A2=ind0608&L=ibm-main&D=1&O=D&P=32264
... you may add one to the count.
-- gil
--
StorageTek
INFORMATION made POWERFUL
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
