Mailing List Review 2022-03-28: There was strong support for fmt::print to be synchronized. Victor offered to update the paper in that direction. Some request for a deeper analysis of performance trade offs. We will look at this on the mailing list when a new revision is published.

P2539R1 Should the output of std::print to a terminal be synchronized with the underlying stream? (Victor Zverovich)

Forwarded for C++23 through electronic polling in LEWG mailing list review started on 2022-03-28.
Some reservations about lack of implementation experience and performance overhead exist, author was advised to address them in the paper (+ benchmarks), before LEWG electronic poll.

2022-07 Library Evolution Electronic Poll Outcomes

Poll 1.9: Send [P2539R1] Should The Output Of print To A Terminal Be Synchronized With The Underlying Stream? to Library Working Group for C++23, classified as an addition ([P0592R4] bucket 3 item).

Outcome: Consensus in favor.

@brycelelbach I'm adding back LEWG, LWG saw that last week and there were design questions, it will come back to you

Presumably someone will file an NB comment for this, so I'll leave it tagged as C++23.

2022-09-27 Library Evolution Telecon

P2539R2: Should the output of print to a terminal be synchronized with the underlying stream?

2022-09-27 Library Evolution Telecon Minutes

Champion: Victor Zverovich

Chair: Bryce Adelstein Lelbach

Minute Taker: Steve Downey

Start: 2022-09-27 12:16 Eastern

Does this paper have:

• Examples?
  • Yes.
• Field experience?
  • Implementation experience?
    • No; fmt does not do this.
  • Usage experience?
    • No; fmt does not do this.
  • Deployment experience?
    • No; fmt does not do this.
• Performance considerations?
  • Yes.
• Discussion of prior art?
  • Yes, fmt and Rust.
• Changes Library Evolution previously requested?
  • Yes.
• Wording?
  • Yes.
• Breaking changes?
  • No - print hasn't shipped yet.
• Feature test macro?
  • Covered by print's feature test macro.
• Freestanding?
  • N/A.

POLL: print to a stream referring to a terminal that is buffered should always flush if the native Unicode API is used.

Attendance: 20

# of Authors: 1

Author Position: WA

Outcome: Consensus in favor.

POLL: Modify P2539R2 (Should the output of print to a terminal be synchronized with the underlying stream?) as per the previous poll, and then send it to Library for C++23 classified as B2 - Improvement, to be confirmed with a Library Evolution electronic poll.

Attendance: 20

# of Authors: 1

Author Position: SF

Outcome: Unanimous consensus in favor.

End: 12:39

Summary

P2539, which concerns whether print to a terminal should synchronize with the underlying stream, was sent back to Library Evolution by Library to confirm our design intent and consider a few issues. We discussed and had consensus to broaden the scope of this synchronization to cover all streams to a terminal that are buffered when the native Unicode API is used.

Next Steps

Revise P2539R2 and then take an electronic poll to send P2539R3 (Should the output of print to a terminal be synchronized with the underlying stream?) to Library for C++23 classified as B2 - Improvement, to be confirmed with a Library Evolution electronic poll.

P2539R2 Should the output of std::print to a terminal be synchronized with the underlying stream? (Victor Zverovich)

P2539R3 Should the output of std::print to a terminal be synchronized with the underlying stream? (Victor Zverovich)

2022-10 Library Evolution Electronic Poll Outcomes

Poll 1.14: Return [P2539R3] print To Terminal Synchronization to Library Working Group for C++23, classified as an improvement of an existing feature ([P0592R4] bucket 2 item) .

Outcome: Consensus in favor.

LWG approved with editorial wording changes in Kona for C++23 - pending check by Jeff and Tim (completed)

https://wiki.edg.com/bin/view/Wg21kona2022/LWG20221107-LA

poll: put P2539r4 into C++23?

P2539R4 Should the output of std::print to a terminal be synchronized with the underlying stream? (Victor Zverovich)