C++ Logo


Advanced search

Re: [SG16] [isocpp-core] Draft proposed resolution for CWG issues 411, 1656, and 2333; numeric and universal character escapes in character and string literals

From: Tom Honermann <tom_at_[hidden]>
Date: Thu, 9 Jan 2020 17:35:24 -0500
On 1/8/20 3:17 PM, Steve Downey wrote:
> I don't believe that "execution character set" is quite the right
> term, however I also believe that there isn't quite the right term in
> the standard to describe the encoding used at compile time for the
> values of members of the execution character set.

I agree it isn't the right term and I have also been under the
impression that the right term doesn't yet exist in the standard.

I spent a little time searching and came across

    "A universal-character-name is translated to the *encoding, in the
    appropriate execution character set*, of the character named."

This suggests that "the encoding of the execution character set" might
be better (for now). If you agree, I'll make that change.

> In [lex.charset] the standard defers to 'locale' for that:
> http://eel.is/c++draft/lex.charset#3.sentence-5
> "The values of the members of the execution character sets and the
> sets of additional members are locale-specific.
> <http://eel.is/c++draft/lex.charset#3.sentence-5> "

My interpretation of that sentence (and thank you, I wasn't aware that
links existed for individual sentences!) is that the execution character
set is locale dependent in the sense that the Visual C++ compiler
defaults the execution character set to the current locale at
compile-time, but can be overridden with the /execution-charset option.

> See P1859R0 <https://isocpp.org/files/papers/P1859R0.html> for early
> draft on terminology for execution character set encoding.

Indeed, I'm looking forward to that paper landing in a near future
standard and would expect the wording I proposed to be impacted by it! :)


> On Wed, Jan 8, 2020 at 9:19 AM Tom Honermann via SG16
> <sg16_at_[hidden] <mailto:sg16_at_[hidden]>> wrote:
> On 1/8/20 2:48 AM, Jens Maurer wrote:
> > I'm a bit confused.
> >
> > There is
> >
> > On 08/01/2020 07.53, Tom Honermann via Core wrote:
> >> Change in 5.13.5 [lex.string] paragraph 3:
> > but no red/green-marked changes follow for that paragraph.
> > (This appears a few times.)
> Thanks, Jens. There are no intended changes to those paragraphs,
> but I
> forgot to update their introduction to make that clear. I retained
> these paragraphs for ease of review given how pervasive the proposed
> changes are to [lex.ccon] and [lex.string]. I've attached an updated
> draft that states that no changes are made to those paragraphs and to
> highlight them with a blue background. I hope that is helpful.
> Tom.
> >
> > Jens

Received on 2020-01-09 16:37:56