🚀 go-pugleaf

In article <4jv7es$d2i@solutions.solon.com>, rdwells@mmm.com wrote:
>VC++ rejects 300F as a float constant, but my reading of
>the syntax part of section 6.1.3.1 of the standard indicates
>that it should be valid:
>
>    300 is a valid digit-sequence
>    a digit-sequence alone is a valid fractional-constant

You must have one of the flawed copies of the ANSI/ISO standard.  In
the definition of fractional-constant there is a period after
digit-sequence (at least in the ANSI original).

Thad

In article <4jv7es$d2i@solutions.solon.com> rdwells@mmm.com  writes:

>VC++ rejects 300F as a float constant, but my reading of
>the syntax part of section 6.1.3.1 of the standard indicates
>that it should be valid:
>
>    300 is a valid digit-sequence
>    a digit-sequence alone is a valid fractional-constant

No, it isn't. In some copies of the standard (such as that in Schildt) there
is an error in that section. It should read:

fractional-constant
          digit-sequence    . digit-sequence
                        opt
          digit-sequence .

Note the final . which is correctly stated in appendix B.1.4

...

>However, the description part of section 6.1.3.1 says:
>
>    "...either the period or the exponent part shall be
>    present"
>
>which would indicate to me that 300F is not valid.

300 can't be both a valid floating-constant and integer-constant or the
grammar would be ambiguous.

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

Thad Smith wrote:
> 
> In article <4jv7es$d2i@solutions.solon.com>, rdwells@mmm.com wrote:
> >VC++ rejects 300F as a float constant, but my reading of
> >the syntax part of section 6.1.3.1 of the standard indicates
> >that it should be valid:
> >
> >    300 is a valid digit-sequence
> >    a digit-sequence alone is a valid fractional-constant
> 
> You must have one of the flawed copies of the ANSI/ISO standard.  In
> the definition of fractional-constant there is a period after
> digit-sequence (at least in the ANSI original).

Thanks much for the clarification.  Is this one of those examples
of what I get for trusting the text of the ANSI standard in
Schildt's book?

	[No.  This bug existed in the ISO standard when Schildt got
	 a copy.  Note: ISO, not ANSI.  ISO's fault. -mod]

(Or do I lose my c.l.c.moderated posting privileges for even
admitting that I own (and use (the left-hand pages of))
a copy?)

	[I do too.  The annotations can be sort of amusing, actually;
	 it's like reading those "best of student bloopers" things. -mod]

rdwells@mmm.com wrote:
: VC++ rejects 300F as a float constant, but my reading of
: the syntax part of section 6.1.3.1 of the standard indicates
: that it should be valid:

[snip]

: Am I reading the syntax wrong?  Should I take this to
: comp.std.c where someone may actually care (:-)?

You are the victim of a typo in the ANSI version of the
standard.  (You'd think, as much as we have to pay to get
an ANSI standard, that the quality of the reproduction would
be better!)  In the second form of the syntax for a
fractional-constant in 6.1.3.1, digit-sequence should be
followed by a period.  Then the comment that you quoted
makes sense.

If you have doubts, check the syntax summary in Annex B
(Subclause B.1.4), which does not have the typo.

Perhaps you are using Schildt's annotated version of the
standard.  If so, don't blame this on him.  This error
occurs in the ANSI document itself.  I'd like to know
whether anyone has reported it through official channels.

Walter Murray
Hewlett-Packard
Support Technology Lab

walterm@hprctbs3.rose.hp.com (Walter Murray) wrote:

>You are the victim of a typo in the ANSI version of the
>standard. [..]
>
>Perhaps you are using Schildt's annotated version of the
>standard.  If so, don't blame this on him.  This error
>occurs in the ANSI document itself.

Not Schildt's fault maybe, but should an annotated edition not have
mentioned the fact that there is a typo in the standard?  Doing that
would at least give the reader the feeling that the annotator has
actually read what he's annotating.


==
Miguel Carrasquer Vidal                     ~ ~
Amsterdam                   _____________  ~ ~
mcv@pi.net                 |_____________|||

========================== Ce .sig n'est pas une .cig

It was written:

: > You must have one of the flawed copies of the ANSI/ISO standard.  In
: > the definition of fractional-constant there is a period after
: > digit-sequence (at least in the ANSI original).

: 	[This bug existed in the ISO standard when Schildt got
: 	 a copy.  Note: ISO, not ANSI.  ISO's fault. -mod]

What about copies of the ISO standard distributed by the standards
bodies in countries other than the U.S.?  When this came up before,
I got the impression that only the ANSI edition of the ISO standard
had this typo.

Walter Murray