Re: Parse error?

Thread View: gnu.gcc.bug

2 messages

[T] Back to Threads [V] Tree View [G] Group Articles

2 total messages Started by Richard Kenner Wed, 20 Nov 1996 00:00

#6166

Author: Richard Kenner
Date: Wed, 20 Nov 1996 00:00

15 lines
1115 bytes

    The point I think he was trying to make is that it wasn't a floating point
    number? 

    It was the hex integer negative 0x2E, plus the integer value zero. 

    The compiler is interpreting it as a floating point number in exponential
    form. It shouldn't be doing so since the 0x should indicate that the number
    is in hex. Basically it looks like the "0x" is having lower precedence than
    the "E" for exponential numbers.

*You* may be interpreting it as three tokens, and that's how a C
compiler would have interpreted it (and how GCC will with
-traditional), but that's *not* the way an ANSI C compiler is required
to interpret it.

Re: Parse error?

#6169

Author: Chris Torek
Date: Thu, 21 Nov 1996 00:00

50 lines
2162 bytes

(on `0x2e+0')

In article <9611210213.AA09203@vlsi1.ultra.nyu.edu>,
Richard Kenner <kenner@vlsi1.ultra.nyu.edu> wrote:
>*You* may be interpreting it as three tokens, and that's how a C
>compiler would have interpreted it (and how GCC will with
>-traditional), but that's *not* the way an ANSI C compiler is required
>to interpret it.

Richard Kenner is right.  The relevant sections of the ANSI standard
(to get ISO sections, add 3 to the first number, and if the result
seems wrong, look forward and back a bit just in case) are:

	2.1.1.2, Translation Phases
		(pay particular attention to phases 3 and 7)
	3.1, Lexical Elements
	3.1.8, Preprocessing Numbers

When 0x2e+0 is converted from a pp-token to a token, it violates
the `shall' in the constraint in section 3.1 (p. 19, ll 28--29);
a diagnostic is required.

A compiler that fails to emit any diagnostic for a program that
includes the pp-number `0x2e+0' where a number is required fails
to conform to the Standard.

The syntax for a pp-number is remarkably inclusive:

	pp-number:
		digit
		. digit
		pp-number digit
		pp-number nondigit
		pp-number e sign
		pp-number E sign
		pp-number .

Thus, 0_e+1r7E-alphabet_soup_42 is a single `pp-number'.  An X3J11
committee member (I forget which one) told me in 1987 or 1988 or
so that this was intentional; in particular, it allows future C
standards to add new numeric notations.  For instance, if someone
were particularly hot for binary:

	0b101

could be added as a new way to specify the number `5'.
-- 
In-Real-Life: Chris Torek, Berkeley Software Design Inc
El Cerrito, CA	Domain:	torek@bsdi.com	+1 510 234 3167

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads