Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Definition of terms

2 views
Skip to first unread message

Johannes Schaub (litb)

unread,
Oct 26, 2009, 10:18:31 PM10/26/09
to
Hello all. Today i had a discussion of when diagnostic messages are required
to be given. It quickly became clear that syntax rule violations are deemed
to give a diagnostic.

I find it intuitive that rules in "Syntax" sections are "syntax rules", but
the one i discussed with wanted to get a normative guarantee of this.

I'm interested in such a thing too. Do anyone happen to know where one can
find this? Are there public drafts of such Standard papers that define these
terms?

Thanks in advance!

James Kuyper

unread,
Oct 26, 2009, 10:44:32 PM10/26/09
to
Johannes Schaub (litb) wrote:
> Hello all. Today i had a discussion of when diagnostic messages are required
> to be given. It quickly became clear that syntax rule violations are deemed
> to give a diagnostic.

That's not quite the correct working. A conforming implementation must
generate at least one diagnostic is required for any translation unit
which contains any violation of a syntax rule or a constraint violation.
(5.1.1.3p1)

> I find it intuitive that rules in "Syntax" sections are "syntax rules", but
> the one i discussed with wanted to get a normative guarantee of this.

It seems pretty clear to me that rules found in the syntax sections are
syntax rules; it's not obvious to me that any additional normative
guarantee is needed, and I'm not aware that any is given. The closest
thing I can think of is section 6.1, which describes the syntax used in
section 6 for defining "syntactic categories (non-terminals)" and
distinguishing them from "literal words and character set members
(terminals)".

Johannes Schaub (litb)

unread,
Oct 26, 2009, 11:02:16 PM10/26/09
to
James Kuyper wrote:

> Johannes Schaub (litb) wrote:
>> Hello all. Today i had a discussion of when diagnostic messages are
>> required to be given. It quickly became clear that syntax rule violations
>> are deemed to give a diagnostic.
>
> That's not quite the correct working. A conforming implementation must
> generate at least one diagnostic is required for any translation unit
> which contains any violation of a syntax rule or a constraint violation.
> (5.1.1.3p1)
>

Of course. I'm sorry for my unclear statements :(

>> I find it intuitive that rules in "Syntax" sections are "syntax rules",
>> but the one i discussed with wanted to get a normative guarantee of this.
>
> It seems pretty clear to me that rules found in the syntax sections are
> syntax rules; it's not obvious to me that any additional normative
> guarantee is needed

Thanks, this is what i thought too. But want to be sure that i've not
overlooked something.

0 new messages