🚀 go-pugleaf

In article <DKKHCH.L6r@midway.uchicago.edu>
           mps@geodesic.com "Michael Spertu" writes:

>For this reason, the two possibilities are:
>
>1) All the texts (including K & R) are advocating code that is not reasonable
>on (the most common) compliant implementations of C.

It is not the responsibility of C language texts to pay lip-service to a
particular implmentation (otherwise we'd all be using near+far pointers,
inline assembly etc). Rather it is up to the implementation to

a) conform to the language specs., which is objective

b) provide a reasonable 'quality' (including features, performance etc.) a
   great deal of which is subjective. If you have a poorly performing
   implementation, the implementation needs to be fixed, not the language
   or the textbooks. An inefficient implementation of printf is (or can be)
   still conforming. You don't go and change all the textbooks just because
   an implementor was lazy.

>2) 32 bit alignment of malloc() on PCs is not compliant.

It always has been and still is.

>Either possibility is very serious, and violates assumptions that most 
>programmers take for granted

It is the implementation that violated the programmer's assumptions. This
has nothing to do with the C standard or C language textbooks. It may
have some relevance to PC/Windows programming textbooks.

-- 
-----------------------------------------
Lawrence Kirby | fred@genesis.demon.co.uk
Wilts, England | 70734.126@compuserve.com
-----------------------------------------