P3655 R1 zstring_view

[wg21bot]

P3655R0 zstring_view (Peter Bindels, Hana Dusikova, Jeremy Rifkin)

[tahonermann]

SG16 reviewed P3655R0 during the 2025-04-23 SG16 meeting. The following polls were taken.

• Poll 1: P3655R0: No objection to use of std::char_traits for consistency and compatibility with std::string_view.
  • Attendees: 8 (no abstentions)

    SF	F	N	A	SA
    7	1	0	0	0

  • Strong consensus.
• Poll 2: P3655R0: Forward to LEWG with encouragement to add analysis of overload resolution and techniques to address ambiguity.
  • Attendees: 8 (no abstentions)

    SF	F	N	A	SA
    8	0	0	0	0

  • Strong consensus.

Poll 1 was taken in response to questions brought up regarding the utility of std::char_traits. SG16 has, on many occasions, discussed various deficiencies of std::char_traits as it applies to modern text. The general sentiment within SG16 is that std::char_traits is no longer fit for its original purpose and that new interfaces should avoid using it or taking a dependency on it. In this case, SG16 felt that an exception was warranted due to std::basic_string_view's existing use of std::char_traits and the explicit intent that specializations of the proposed std::basic_zstring_view be substitutable for, and convertible to, the corresponding specializations of std::basic_string_view.

Poll 2 demonstrates that SG16 sees significant value in std::basic_zstring_view. It was acknowledged that null terminated strings are not going away any time soon and that this proposal offers a more natural and performant way to interact with existing interfaces that consume such strings. However, SG16 also recognized that there are deployment and adoption concerns with respect to existing function overload sets that accept std::basic_string, std::basic_string_view, and const CharT* arguments. Such concerns fall more under LEWG's purview, so SG16 has encouraged the proposal authors to add additional analysis of the consequences of adding an additional overload that accepts a std::basic_zstring_view argument to an existing overload set, and the consequences of passing a std::basic_zstring_view argument (which will be convertible to std::basic_string_view) to such existing overload sets. In cases that lead to ambiguous overload resolution, SG16 encourages investigation into how such ambiguity can be mitigated; perhaps through the use of constrained function template.

[wg21bot]

P3655R1 zstring_view (Peter Bindels, Hana Dusikova, Jeremy Rifkin)

[tahonermann]

#2338 also includes a zstring_view proposal. My understanding is that the authors of that paper will work with the authors of this paper to unify their proposals. SG16 has already forwarded this proposal, so I tagged the other one for SG16, but will await word from the authors regarding their unification efforts. If a lack of consensus for a unified proposal emerges, then I will expect to see both papers back in SG16.

[dascandy]

We have aligned and are busy merging the two papers. We intend to use this paper number for the paper proposing zstring_view, and the other number will be used for a related subject that we found out during discussion (like P3566 and P3711 are too)