> On Sep 8, 2019, at 3:31 PM, Tony V E via Lib <lib_at_[hidden]> wrote:
>
> Do we have / could we have / should we have
> a clear long term (20 years) direction for text in C++?

I would like that very much, but we don’t control the ecosystem, and will have to, to some degree, roll with where the community takes us.

>
> ie the long term direction is unicode.
> and/or specifically the long term direction is UTF8.

I think we do have wide spread agreement on that, though UTF-16 is likely to remain strongly relevant in some niches.

> We expect everyone to use char8_t then? Or we expect char to become utf8 someday?

I think it is very unlikely that there will be a mass migration to char8_t. My expectation is that it will be used for the internal encoding within some percentage of new projects and components.

With regard to char, I expect it to remain the type used for text that may or may not be UTF-8.

I think Microsoft will eventually provide (non-experimental) means to use UTF-8 with Win32 and that this will likely come in three forms:

1) support for UTF-8 as the system wide Active Code Page (ACP). This is already available as an experimental option.

2) support for executables to opt-in to a per-process override of the system wide ACP. In this mode, stdio would presumably traffic in the system wide ACP and require transcoding (I don’t think implicit transcoding is realistic). This is already available as an experimental option.

3) support for a subset of Win32 interfaces that take char8_t. E.g., U8 variants of some existing A/W interfaces.

z/OS is a bit more interesting. Though EBCDIC based, ASCII interfaces that implicitly transcode to EBCDIC are available for a subset of C interfaces . As far as I am aware, there are no plans to extend this support to include UTF-8.

> What do we want the long term future to look like?

🎵You can’t always get what you want 🎶

> deprecate std::string?

Probably not.

>
> And then a list of short term stop-gap measures, like "we know we can't do X yet,so we do Y for now".
> Like we use char, but plan on switching to char8_t.
> Or QoI escape hatches. etc.

I think we need to plan to support use of both char and char8_t for UTF-8 text for the foreseeable future.

Tom.

>
>
>
>> On Sun, Sep 8, 2019 at 2:46 PM Corentin via Lib <lib_at_[hidden]> wrote:
>>
>>
>>> On Sun, 8 Sep 2019 at 19:30, Tom Honermann <tom_at_[hidden]> wrote:
>>>> On 9/8/19 12:40 PM, Corentin wrote:
>>>>
>>>>
>>>>> On Sun, 8 Sep 2019 at 18:12, Tom Honermann <tom_at_[hidden]> wrote:
>>>>>> On 9/8/19 6:00 AM, Corentin via Lib wrote:
>>>>>>
>>>>>>
>>>>>>> On Sun, 8 Sep 2019 at 11:17, Corentin <corentin.jabot_at_[hidden]> wrote:
>>>>>>>
>>>>>>>
>>>>>>>> On Sun, 8 Sep 2019 at 09:52, Billy O'Neal (VC LIBS) <bion_at_[hidden]> wrote:
>>>>>>>> > I agree that EGCS is the best option. That doesn't drag locale
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Because we don’t get to assume that we’re talking about Unicode at all, it absolutely drags in locale.
>>>>>>>>
>>>>>>>
>>>>>>> Sorry, I should have been more specific.
>>>>>>> There is a non-tailored Unicode EGCS boundary algorithm (but it can be tailored)
>>>>>>> I didn't mean to imply that text manipulation can be done without knowing its encoding and never use "locale" to mean encoding.
>>>>>>>
>>>>>>> EGCS are only defined for text whose character repertoire is Unicode, other encodings deal with codepoints
>>>>>>
>>>>>>
>>>>>> To be clear, the difference of whether the EGC algorithm is required to be tailored or not is that tailoring for all intent and purposes requires
>>>>>> icu or something with CLDR, which restrict the platforms on which this can be implemented
>>>>> Tailoring is not relevant to this discussion.
>>>>>
>>>> It is - see https://unicode.org/reports/tr29/ "ch" is 2 EGCS in most locales but in Slovak it's 1. I don't make the rules :D
>>> It isn't relevant in determining how we resolve this issue. If the resolution is that field widths are measured in EGCs, then we've already decided that the width is locale dependent and tailoring becomes an implementation detail.
>>
>> No, format decided to be locale-independent (for good reason) and applying locale specific behavior implicitly would be against that.
>> I'n arguing for encoding specific behavior
>>
>>>>
>>>>> The locale dependency stems from the encoding itself being dependent on locale. Again, LANG=C vs LANG=C.UTF-8. If the specified behavior is encoding dependent (as it would have to be for field width to be a count of any of code points, scalar values, or EGCs), then it is also locale dependent (for char and wchar_t). Thus there is a trade off:
>>>>>
>>>>> Either the behavior is locale dependent in which case, field widths could be specified such that they count code points, scalar values, or EGCs when the locale selects a Unicode encoding (and something else for non-Unicode encodings), or
>>>>> The behavior is not locale dependent in which case, field widths can only be specified in terms of code units.
>>>>
>>>> Agreed, but let me rephrase:
>>>>
>>>> Either a string is text and therefore we need and to know its encoding, or it is a sequence of bytes (in the case of char)
>>>> I have an opinion about what we are dealing with in this context :D
>>> So your preference is for trade off #1 above and the cost is that std::format is no longer locale insensitive even in the cases where a std::locale argument is not provided.
>>>
>> It would be _encoding_ sensitive
>> It would not change for example the decimal separator.
>>
>> When Unicode is involved - and even when it is not, it is I think important not to conflate locale and encoding even if C kinda amalgamates the two and derives one from the other.
>>
>>
>>
>>> Since I don't think field width works for alignment, even if EGCs are used (see Henri's post - https://hsivonen.fi/string-length), I prefer trade off #2.
>>>
>>> Tom.
>>>
>>>>
>>>>
>>>>> Recall that, unless there is a call to std::setlocale, all C and C++ processes start with the locale set to "C"
>>>>>
>>>>> Tom.
>>>>>
>>>>>>
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Billy3
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> From: Lib <lib-bounces_at_[hidden]> on behalf of Corentin via Lib <lib_at_[hidden]>
>>>>>>>> Sent: Saturday, September 7, 2019 11:08:25 PM
>>>>>>>> To: Library Working Group <lib_at_[hidden]>
>>>>>>>> Cc: Corentin <corentin.jabot_at_[hidden]>; Victor Zverovich <victor.zverovich_at_[hidden]>; Tom Honermann <tom_at_[hidden]>; unicode_at_[hidden] <unicode_at_[hidden]>
>>>>>>>> Subject: Re: [isocpp-lib] New issue: Are std::format field widths code units, code points, or something else?
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>> On Sun, Sep 8, 2019, 5:30 AM Tom Honermann via Lib <lib_at_[hidden]> wrote:
>>>>>>>>>> On 9/7/19 10:44 PM, Victor Zverovich wrote:
>>>>>>>>>> > Is field width measured in code units, code points, or something else?
>>>>>>>>>>
>>>>>>>>>> I think the main consideration here is that width should be locale-independent by default for consistency with the rest of std::format's design.
>>>>>>>>> I agree with that goal, but...
>>>>>>>>>> If we can say that width is measured in grapheme clusters or code points based on the execution encoding (or whatever the standardese term) without querying the locale then I suggest doing so.
>>>>>>>>> I don't know how to do that. From my response to Zach, if code units aren't used, then behavior should be different for LANG=C vs LANG=C.UTF-8.
>>>>>>>>>> I have slight preference for grapheme clusters since those correspond to user-perceived characters, but only have implementation experience with code points (this is what both the fmt library and Python do).
>>>>>>>>> I would definitely vote for EGCs over code points. I think code points are probably the worst of the options since it makes the results dependent on Unicode normalization form.
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> I disagree. Code Units is the worse option. For me anything involving code units is a big red flag. I agree that EGCS is the best option. That doesn't drag locale, might be a bit involved for implementers in 20.
>>>>>>>> I don't think specify EGCS for Unicode text and codepoints otherwise wouldn't be too difficult - implementation might be a bit challenging on some platforms in the 20 time frame but they could fallback to codepoints in the meantime. Not perfect but I think we need a good long term solution rather than a bad short term one
>>>>>>>>
>>>>>>>>> Tom.
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Cheers,
>>>>>>>>>> Victor
>>>>>>>>>>
>>>>>>>>>>> On Sat, Sep 7, 2019 at 5:13 PM Tom Honermann via Lib <lib_at_[hidden]> wrote:
>>>>>>>>>>> [format.string.std]p7 states:
>>>>>>>>>>>
>>>>>>>>>>>> The positive-integer in width is a decimal integer defining the minimum field width. If width is not specified, there is no minimum field width, and the field width is determined based on the content of the field.
>>>>>>>>>>>>
>>>>>>>>>>> Is field width measured in code units, code points, or something else?
>>>>>>>>>>>
>>>>>>>>>>> Consider the following example assuming a UTF-8 locale:
>>>>>>>>>>>
>>>>>>>>>>> std::format("{}", "\xC3\x81"); // U+00C1 { LATIN CAPITAL LETTER A WITH ACUTE }
>>>>>>>>>>> std::format("{}", "\x41\xCC\x81"); // U+0041 U+0301 { LATIN CAPITAL LETTER A } { COMBINING ACUTE ACCENT }
>>>>>>>>>>>
>>>>>>>>>>> In both cases, the arguments encode the same user-perceived character (Á). The first uses two UTF-8 code units to encode a single code point that represents a single glyph using a composed Unicode normalization form. The second uses three code units to encode two code points that represent the same glyph using a decomposed Unicode normalization form.
>>>>>>>>>>>
>>>>>>>>>>> How is the field width determined? If measured in code units, the first has a width of 2 and the second of 3. If measured in code points, the first has a width of 1 and the second of 2. If measured in grapheme clusters, both have a width of 1. Is the determination locale dependent?
>>>>>>>>>>>
>>>>>>>>>>> Proposed resolution:
>>>>>>>>>>>
>>>>>>>>>>> Field widths are measured in code units and are not locale dependent. Modify [format.string.std]p7 as follows:
>>>>>>>>>>>
>>>>>>>>>>>> The positive-integer in width is a decimal integer defining the minimum field width. If width is not specified, there is no minimum field width, and the field width is determined based on the content of the field. Field width is measured in code units. Each byte of a multibyte character contributes to the field width.
>>>>>>>>>>>>
>>>>>>>>>>> (code unit is not formally defined in the standard. Most uses occur in UTF-8 and UTF-16 specific contexts, but [lex.ext]p5 uses it in an encoding agnostic context.)
>>>>>>>>>>>
>>>>>>>>>>> Tom.
>>>>>>>>>>>
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Lib mailing list
>>>>>>>>>>> Lib_at_[hidden]
>>>>>>>>>>> Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/lib
>>>>>>>>>>> Link to this post: http://lists.isocpp.org/lib/2019/09/13440.php
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>> Lib mailing list
>>>>>>>>> Lib_at_[hidden]
>>>>>>>>> Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/lib
>>>>>>>>> Link to this post: http://lists.isocpp.org/lib/2019/09/13446.php
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Lib mailing list
>>>>>> Lib_at_[hidden]
>>>>>> Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/lib
>>>>>> Link to this post: http://lists.isocpp.org/lib/2019/09/13453.php
>>>>>
>>>
>> _______________________________________________
>> Lib mailing list
>> Lib_at_[hidden]
>> Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/lib
>> Link to this post: http://lists.isocpp.org/lib/2019/09/13458.php
>
>
> --
> Be seeing you,
> Tony
> _______________________________________________
> Lib mailing list
> Lib_at_[hidden]
> Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/lib
> Link to this post: http://lists.isocpp.org/lib/2019/09/13459.php