C language book recommendations

[rpjday at ccu.umanitoba.ca]

In article <741 at taumet.com> steve at taumet.com (Stephen Clamage) writes:
>rpjday at ccu.umanitoba.ca writes:
>
>>>>  C: A Reference Manual 3rd Edition
>>>>  Samuel P Harbison & Guy L Steele Jr
>>>>  Prentice Hall
>
>>Even the occasional annoying typo doesn't detract much.
>>What typo?  Glad you asked.
>
>>On [page 94], in a table showing the legalities of
>>various declarations in various models, we read that
>
>>extern int x = 0 ;
>
>>in the omitted storage class model is treated as a reference.
>>However, according to the explanation on p. 93, this should
>>be illegal.  Comments?
>
>Look over the material again.  In their book, H&S are not presenting only
>ANSI C.  They are trying to cover the range of C implementations which have
>been at all widely used.  If you have to write code for a variety of C
>implementations, or understand some moldy old code written for some
>system you do not have access to, this book will help.  It explains all
>the common things which have been done, and how you can get your code
>to work.
>
>As to your specific example, it is clearly marked in the table as
>"illegal" for typical C implementations, and "Don't use" under
>the "recommendation" column.
>-- 
>
>Steve Clamage, TauMetric Corp, steve at taumet.com

  I don't have my 3rd ed. of H&S here, so I may very well 
embarrass myself, but I think I checked out that section
thoroughly enough to remember what it says.  I am aware that
H&S are showing various compiler models.  They also quite
CLEARLY say that one of them, the omitted storage class
model, is what ANSI C uses.  Their definition of this model
is what I have given above.  On the next page, they also 
CLEARLY say that "extern int x = 0;" is not, as you claim,
illegal, but that it is to be treated as a reference --
that is, a referencing declaration.

  What H&S recommend is irrelevant.  My concern is with the
entry in the table.  Comments?

R. Day