Re: P1788 latest revision
John,
The text is Ok for me.
> As a result of them, I have also added "literal" to the Definitions.
It would be good to update P1788.out, because bookmarks are after "literal"
are shifted now.
-Dima
----- Original Message -----
From: PryceJD1@xxxxxxxxxxxxx
To: dmitry.nadezhin@xxxxxxxxxx
Cc: stds-1788@xxxxxxxxxxxxxxxxx
Sent: Tuesday, March 24, 2015 5:26:13 PM GMT +04:00 Abu Dhabi / Muscat
Subject: Re: P1788 latest revision
Dmitry
Thanks for these comments, you have been reading carefully.
As a result of them, I have also added "literal" to the Definitions.
On 24 Mar 2015, at 04:59, Dmitry Nadezhin <dmitry.nadezhin@xxxxxxxxxx> wrote:
> John,
>
> I like using "all-flavor literals" and the removal of the word "valid" with literals.
>
> There are still some small issues.
>
> 1) The (AF) definition of mathematical interval of constructor 3.2.41
> has a link to set-based chapter. Now it can link to all-flavor chapter
> <<<
> \ssrf{lvl2bintvlconstructors}
> ===
> \ssrf{ch1reqdconstructors}
>>>
Done, also ref'd 7.5.3
> 2) As bare literals are no more considered as decorated literals by 9.7.5,
> the grammars in Table 9.5 and Table need to be fixed:
>
> <<<
> \Trm{decoratedIntvlLiteral}& \trm{bareIntvlLiteral} \M| \trm{bareIntvlLiteral} \L{\_} \trm{decorationLit} \\[.8ex]
> ===
> \Trm{decoratedIntvlLiteral}& \trm{bareIntvlLiteral} \L{\_} \trm{decorationLit} \\[.8ex]
> >>>
> and
> <<<
> C &\Trm{intervalLiteral} & \trm{bareIntvlLiteral} \M| \trm{bareIntvlLiteral} \L{\_} \trm{decorationLit} \ul{\M| \trm{NaI}}
> ===
> C &\Trm{intervalLiteral} & \trm{bareIntvlLiteral} \L{\_} \trm{decorationLit} \ul{\M| \trm{NaI}}
> >>>
Done x2
!! But also I need to delete item a) of 12.11.4, right? (p58 lines 14-16.)
> 3) Text 12.12.7 about setbased level 2 constructor needs some rewording.
>
> Page 62 line 18.
> Say explicitly "is a bare interval literal" instead of "is an interval literal".
>
> Page 62 line 19.
> Remove "or [l,u]_dx".
Done x2
> Page 62 lines 27, 28 are not correct now:
> "The decorated constructor succeeds if and only if
> the bare interval constructor succeeds. The decorated constructor fails returning NaI if and only if the bare
> interval constructor fails."
>
> Page 62 line 31-34 should be formulated in another way.
In fact all of p62 "Decorated interval constructors" and "Exception behavior" need revising. Line 26 "taking the same input(s) as the bare constructor" is no longer true. Will the following do? Is the 2nd sentence OK or should I use "respectively"?
======
Decorated interval constructors.
Let the prefix b- or d- denote the bare or decorated version of a constructor. If b-numsToInterval(l,u) or b-textToInterval(s) succeeds with result y, then d-numsToInterval(l,u) or d-textToInterval(s) succeeds with result newDec(y), see 11.5.
If s is a decorated interval literal sx_sd with Level 1 value x_dx, see 12.11.4, and b-textToInterval(sx) succeeds with result y, then d-textToInterval(s) succeeds with result y_dy, where dy = dx except when dx = com and overflow occurred, that is, x is bounded and y is unbounded. Then dy shall equal dac.
Otherwise the call fails, and the result is NaI.
Exception behavior.
UndefinedOperation is signaled when a constructor call fails.
NOTE—When signaled by the decorated constructor it will normally be ignored since returning NaI gives sufficient information.
PossiblyUndefinedOperation is signaled when the implementation cannot determine whether a Level 1 value exists (the two cases (b) above).
======
On SVN as r441. Please check
John Pryce