P2093R1 Formatted output (Victor Zverovich)

2020-09-29 Library Evolution Telecon Minutes

Chair: Fabio Fracassi

Champion: Victor Zverovich

Minute Taker: Ben Craig

POLL: We prefer std::cout as the default output target.

POLL: Add a member function on ostream instead of a std::print(ostream&, ...) free function overload

Author's Position: WF

Attendance: 35

Outcome: Consensus against.

POLL: Remove std::println from the paper

Outcome: No strong consensus for change.

Author's Position: N

Attendance: 35

P2093R2 Formatted output (Victor Zverovich)

P2093R3 Formatted output (Victor Zverovich)

Polls

Poll: We want P2093r2 to revert moving the print(ostream) overloads to the header.

Attendance: 17

# of Authors: 1

Author Position: WA

Outcome: Keep the paper as written.

Summary

This was only a short discussion at the end of a session. As before we like the direction of the paper. No new issues came up and we affirmed the design decission of the author to put the ostream related overloads into the ostream header.

Outcome

No changes to the paper. Discussion on this paper will continue.

P2093R4 Formatted output (Victor Zverovich)

SG16 reviewed P2093R2 in our 2020-12-09 telecon and P2093R3 in our 2021-02-10 telecon. R3 implemented the feedback provided during the review of R2.

The following poll was conducted:

Forward P2093R3 to LEWG.

Consensus? Yes

Sustained opposition exists regarding progressing this paper without exposing low level interfaces to 1) detect whether output is directed to a console/terminal, 2) write directly to a console/terminal. The author has agreed to propose such facilities in a separate paper.

Concerns regarding encoding and how support for additional character types can be incorporated remain, but these concerns were not deemed sufficient to preclude progress on this paper. Participants with such concerns are encouraged to follow up in a separate paper.

Removing the SG16 label.

@tahonermann, there is some unrelated text in the poll statement above ("D1885R2: We want to support non-IANA registered encodings provided by the implementation").

there is some unrelated text in the poll statement above

Oops, thank you @vitaut, fixed!

@brycelelbach, I think LEWG related tags may not be correctly set for this paper. I suspect this should have the "ready-for-library-evolution-meeting-review" tag.

P2093R5 Formatted output (Victor Zverovich)

2021-04-06 Library Evolution Telecon

P2093R5: Formatted output

2021-04-06 Library Evolution Telecon Minutes

Poll: Block P2093 until we have a proposal for a lower level facility that can query tty/console and preform direct output to the console

Attendance: 26

# of Authors: 1

Author Position: SA

Outcome: Consensus Against

POLL: We want std::print et al. non ostream overloads to go to:

Attendance: 24

# of Authors: 1

Author Position: Against both of the favored options.

Outcome: Both <print> and <format> have consensus, author can choose on for the next revision.

Summary

We discussed several questions that were posed to us by SG16 (Unicode). We were able to make progress on several technical questions. The discussion about transcoding requirements and consequences however needs more information to progress.

Outcome

We decided not to block the paper to wait for a lower level facility to interact with consoles/terminals. We discussed naming and header placement as guidance. The paper has been sent back to SG16 (Unicode) for clarification. The issues that need to be decided for this Paper are in SG16's expert domain. Library Evolution needs more information about the design space.

P2093R9 Formatted output (Victor Zverovich)

We revisited P2093R8 during our SG16 call on September 8th, 2021. We took the following polls during our discussion:

• Poll: Use of UTF-8 as the literal encoding is sufficient for facilities to establish encoding expectations.

  • Attendance: 9

    SF	F	N	A	SA
    2	3	2	1	0

  • Consensus in favour.

• Poll: Correct the P2093R8 wording for [print.syn].31 to remove ambiguities, and forward P2093 as revised to LEWG with a recommended ship vehicle of C++23.

  • Attendance: 9

    SF	F	N	A	SA
    1	4	2	0	0

  • Consensus in favor.

Removing SG16 label.

P2093R10 Formatted output (Victor Zverovich)

2021-11-29 Library Evolution Telecon

P2093R10: Formatted Output

2021-11-29 Library Evolution Telecon Minutes

Chair: Bryce Adelstein Lelbach

Champion: Victor Zverovich

Minute Taker: Ben Craig

Start: 2021-11-29 11:09 Pacific

Does this proposal have:

• Examples?
  • Yes.
• Implementation experience?
  • Partially. We have experience in fmt, but not in a C++ Standard Library or reference implementation written from the spec.
  • No field experience in fmt with the mojibake prevention.
• Usage experience?
  • Partially. We have experience in fmt, but not in a C++ Standard Library or reference implementation written from the spec.
  • No field experience in fmt with the mojibake prevention.
• Deployment experience?
  • Partially. We have experience in fmt, but not in a C++ Standard Library or reference implementation written from the spec.
  • No field experience in fmt with the mojibake prevention.
• Performance considerations?
  • Yes.
• Discussion of prior art?
  • Yes (fmt, Python's print).
• Changes Library Evolution previously requested?
  • Yes.
• Wording?
  • Yes.
• Feature test macro?
  • Yes.

Open Questions:

• Should vprint_unicode always require Unicode?

POLL: Modify P2093R10 (Formatted Output) by updating the wording of the vprint_unicode stream overload to match the vprint_unicode FILE* overload and not reference basic_filebuf specifically, and send the revised paper to LWG for C++23 with priority B3 (to be confirmed with a Library Evolution electronic poll).

Attendance: 21

# of Authors: 1

Author Position: SF

Outcome: Strong consensus in favor.

End: 11:55

Summary

We discussed P2098 (formatted output), which we've seen a few times in the past. The paper has been iterating in the Text and Unicode study group, which recently approved it and sent it back to Library Evolution.

There have not been many significant revisions to the library API since the last time Library Evolution reviewed it. We did identify some mismatches between design intent and wording during our review, but these were all minor matters with a clear resolution.

Outcome

Modify P2093R10 (Formatted Output) by updating the wording of the vprint_unicode stream overload to match the vprint_unicode FILE* overload and not reference basic_filebuf specifically, and send the revised paper to LWG for C++23 with priority B3 (to be confirmed with a Library Evolution electronic poll).

P2093R11 Formatted output (Victor Zverovich)

2021-12 Library Evolution Electronic Poll Outcomes

POLL: Send [P2093R11] (Formatted Output) to Library Working Group for C++23, classified as an addition ([P0592R4] bucket 3 item).

Consensus in favor, forwarded to LWG

P2093R12 Formatted output (Victor Zverovich)

P2093R13 Formatted output (Victor Zverovich)

LWG did a final review on 2022-03-25 and approved for c++23

https://wiki.edg.com/bin/view/Wg21telecons2022/P2093-20220325

poll: include P2093r14 into C++23?

P2093R14 Formatted output (Victor Zverovich)