Trip report: Winter 2021 ISO C++ standards meeting (virtual)

Today, the ISO C++ committee held its second full-committee (plenary) meeting of the pandemic and adopted a few more features and improvements for draft C++23.

A record of 18 voting nations sent representatives to this meeting: Austria, Bulgaria, Canada, Czech Republic, Finland, France, Germany, Israel, Italy, Japan, Netherlands, Poland, Romania, Russia, Spain, Switzerland, United Kingdom, and United States. Japan had participated in person during C++98 and C++11, and has always given us good remote ballot feedback during C++14/17/20, and is attending again now; welcome back! Italy and Romania are our newest national bodies; welcome!

Our virtual 2021

We continue to have the same priorities and the same schedule we originally adopted for C++23. However, since the pandemic began, WG21 and its subgroups have had to meet all-virtually via Zoom, and we are not going to try to have a face-to-face meeting in 2021 (see What’s Next below). Some subgroups had already been having virtual meetings for years, but this was a major change for other groups including our two main design groups – the language and library evolution working groups (EWG and LEWG). In all, over the past year we have held approximately 200 virtual meetings.

Today: A few more C++23 features adopted

Today we formally adopted a second round of small features for C++23, as well as a number of bug fixes. Below, I’ll list some of the more user-noticeable changes and credit all those paper authors, but note that this is far from an exhaustive list of important contributors… even for these papers, nothing gets done without help from a lot of people and unsung heroes, so thank you first to all of the people not named here who helped the authors move their proposals forward! And thank you to everyone who worked on the adopted issue resolutions and smaller papers I didn’t include in this list.

P1102 by Alex Christensen and JF Bastien is the main noticeable change we adopted for the core language itself. It’s just a tiny bit of cleanup, but one that I’m personally fond of: In C++23 we will be able to omit empty ( ) lambda parameter lists even when we have to declare the lambda mutable. I’m the one who proposed the lambda syntax we have today (except for the mutable part which wasn’t mine and I never liked), including that it enabled making unused parts of the syntax optional so that we can write simple lambdas simply. For example, today we can already write

[x]{ return f(x); }

as a legal synonym for

[x] () -> auto { return f(x); }

and omit the empty parameter list and deduced return type. Even so, I’ve noticed a lot of people write the ( ) part anyway, which isn’t wrong or anything, it’s just that often they write it because they don’t know they can omit it too. And part of the problem was the oddity in pre-C++23 that if you need to write mutable, then you actually do have to also write the ( ) (but not the return type), which was just weird but was another reason for people to just write ( ) all the time, because sometimes they had to. With P1102, we don’t have to. That’s more consistent. Thanks, Alex and JF!

In the spirit of “completing C++20,” P2259 by Tim Song makes several fixes to iterator_category to make it work better with ranges and adaptors. Here is an example of code that does not compile today for arcane reasons (see the paper), but will be legal C++23 thanks to Tim:

std::vector<int> vec = {42};
auto r = vec | std::views::transform([](int c) { return std::views::single(c);})
             | std::views::join
             | std::views::filter([](int c) { return c > 0; });

Further in the “completing C++20” spirit, P2017 by Barry Revzin fixes some additional glitches in ranges to make them work better. Here is an example of safe and efficient code that does not compile today, where for arcane reasons the declaration of e isn’t supported and today’s workaround is to make the code more complex and less efficient. This will be legal C++23 thanks to Barry:

auto trim(std::string const& s) {
    auto isalpha = [](unsigned char c){ return std::isalpha(c); };
    auto b = ranges::find_if(s, isalpha);
    auto e = ranges::find_if(s | views::reverse, isalpha).base();
    return subrange(b, e);

P2212 by Alexey Dmitriev and Howard Hinnant generalizes time_point::clock to allow for greater flexibility in the kinds of clocks it supports, including stateful clocks, external system clocks that don’t really have time_points, representing “time of day” as a distinct time_point, and more.

P2162 by Barry Revzin takes an important first step toward cleaning up std::visit and lay the groundwork for its further generalization. Even if you don’t yet love std::visit, it’s a useful tool that P2162 makes more useful by making it work more regularly. We expect to see further generalization in the future, which is much easier to do with a cleaner and more regular existing feature to build upon.

Finally, I saw cheers and celebratory emoji erupt in the Zoom chat window when we adopted P1682 by JeanHeyd Meneide. It’s very small, but very useful. When passing an enum to an API that uses the underlying type, today we have to write a static_cast to the std::underlying_type, which makes us repeat the enum’s name and so is cumbersome all the time and brittle for type-safety under maintenance if we change to use a different enum:

some_untyped_api( static_cast<std::underlying_type_t<ABCD>>(some_value) );

Thanks to JeanHeyd, in C++23 we will be able to write:

some_untyped_api( std::to_underlying(some_value) );

Note that of course standard library vendors don’t have to wait until 2023 to provide to_underlying or any of these other fixes and improvements. Just having a feature like this one voted into the draft standard is often enough for vendors to be proactive in providing it… these days, vendors are more closely tracking our draft standard meeting by meeting rather than waiting for the official release, in part because we are shipping regularly and predictably and we don’t vote features into the draft standard until we think they’re pretty well baked so that vendors have less risk in implementing them early.

We also adopted a number of other issue resolutions and small papers that made additional improvements.

Finally, we came close to adopting P0533 by Edward Rosten and Oliver Rosten, which is about adding constexpr to many of the functions in math.h that we share with C. This is clearly a Good Thing and therefore many voted in favor of adopting the paper. The only hesitation that stopped it from getting consensus this time were concerns that it needed more time to iron out how implementations would implement it, such as how to deal with errno in a constexpr context. This is the kind of question that often arises when we want to make improvements to entities declare in the C headers, because not only are they governed by the C standard rather than the C++ standard, but typically they are provided and controlled by the operating system vendor rather than by the C++ compiler/library writer, and those constraints always mean a bit of extra work when we want to make improvements for C++ programmers and remain compatible. As far as I know, everyone wants to see these functions made constexpr, so we expect to see this paper come to plenary again in the future. Thanks for your perseverance, Edward and Oliver!

What’s next

As long as we are meeting virtually, we will continue to have virtual plenaries like the one we had this week to formally adopt new features as they progress through subgroups. Our next two virtual plenaries to adopt features into the C++23 working draft will be held in June and November. Progress will be slower than when we can meet face-to-face, and we’ll doubtless defer some topics that really need in-person discussion until we can meet again safely, but in the meantime we’ll make what progress we can and we’ll ship C++23 on time.

The next tentatively planned face-to-face meeting is February 2022 in Portland, OR, USA; however, we likely won’t know until well into the autumn whether we’ll be able to confirm that or need to postpone it. You can find a list of our meeting plans on the Upcoming Meetings page.

Thank you again to the hundreds of people who are working tirelessly on C++, even in our current altered world. Your flexibility and willingness to adjust are much appreciated by all of us in the committee and by all the C++ communities! Thank you, and see you on Zoom.

Firsts in 2020 (or, A little dose of good news)

2020 has been mostly terrible. That includes for the C++ committee and many of our communities, where just this month we lost Beman Dawes. Beman was one of the most important and influential C++ experts in the world, and made his many contributions mostly behind the scenes. I and everyone else who has ever benefited from any of the standardized STL, Boost, C++Now, std::filesystem, C++98/11/14/17, and more — so, really, most people who have ever used C++ — all owe Beman a debt of gratitude. We miss him greatly.

To end the year with a little dose of good news, I thought I’d mention a just few positive C++ accomplishments that did happen for 2020, and were happier “first-ever” achievements.

First, the big one…

C++20 is the first ever “D&E-complete” release of C++. In February, we completed C++20, which is the first release of Standard C++ that includes every feature that Bjarne Stroustrup envisioned for C++’s evolution in his 1994 book The Design and Evolution of C++ (aka D&E), including concepts, coroutines, modules, and more, except only for one minor feature (unified function call syntax). Thank you to Bjarne for sticking with it until we got there, and personally doing the heavy lifting to drive important features like concepts into Standard C++!

C++20 is the first release of C++ that added a feature that made the standard smaller. When I talk about the importance of simplifying C++ by judiciously adding features that let programmers express their intent directly, some people legitimately object that adding a feature makes C++ bigger and more complex. I reply “but it makes C++ code simpler” and “if it replaces something more complex then we can teach a simpler C++ for new code,” but those effects have been hard to measure concretely. Now in C++20 for the first time we added a new feature that made the standard smaller: We added the C++20 spaceship operator to the language, but we also applied it throughout the C++ standard library and that made the library specification nearly 20 pages shorter — a net reduction. So for the first time we can measure that, yes, adding a feature to C++ can make C++ smaller. Thank you to everyone who helped me with that proposal and who are listed in the Acknowledgements in the link, and especially to Walter E. Brown, Jens Maurer, Barry Revzin, and David Stone!

First year for all-virtual standards meetings, including EWG and LEWG. Since March, for the first time all major subgroups including the two main design subgroups of EWG (language) and LEWG (library) have been having virtual meetings by telecon or Zoom and making progress in between face-to-face meetings. We’ve also had a record number of nearly 20 virtual subgroup meetings on average per month. It’s great to see that, despite the pandemic, the committee has continued work on C++23 and other long-pole features, and in November we were able to formally adopt the first C++23 features into our brand-new C++23 working draft. Thank you once again to JF Bastien (EWG), Bryce Lelbach (LEWG), and their assistants, and all the other subgroup chairs and participants for patiently supporting these changes that we had to invent and transition to at short notice, and as we continue to work out the kinks as we go!

Many first virtual conferences. And of course 2020 saw many of our C++ conferences hold their first virtual events (and create new ones like Pure Virtual C++), in the face of huge uncertainties and technical challenges with bleeding-edge technologies, to make it all work far more smoothly than we really would have had any right to expect on such short notice. Thank you to the organizers for working so busily behind the scenes to make it possible to have a facsimile of our face-to-face conferences until we can meet again in person!

Here’s hoping that by this time next year we will all be doing better in every way, and have a happier 2021 to reflect upon. Thank you again, everyone, for your interest in C++ and support for our many C++ events, forums, and other communities large and small, and best wishes for a great 2021.

Trip report: Autumn ISO C++ standards meeting (virtual)

On Monday, the ISO C++ committee completed its final full-committee (plenary) meeting of 2020 and adopted the first changes to the C++23 working draft, including a few new features.

This was a first in several ways: It was our first-ever virtual plenary, held online via Zoom. It was also our first-ever plenary meeting that wasn’t held at the end of a long around-the-clock week of intensive subgroup meetings; instead, it was held at the end of nearly nine months of virtual subgroup meetings.

Our virtual 2020

The pandemic was just getting started when we held our February meeting in Prague, Czech Republic. Since then it has of course been impossible to meet in person; as I mentioned before, our ISO C++ meetings are virtual until further notice, but we continue to have the same priorities and the same schedule for C++23.

So since the pandemic began, WG21 subgroups have been meeting virtually via Zoom. Some subgroups had already been having virtual meetings for years, but this was a major change for other groups including our two main design groups – the language and library evolution working groups (EWG and LEWG).

In all, since Prague we have held about 150 virtual meetings. When lots of subgroups are meeting, some of them weekly, those meetings add up!

This week: First C++23 features adopted

On Monday we formally adopted the first features of C++23, including the first C++23 language feature, as well as a number of bug fixes.

First up was P0330 by JeanHeyd Meneide, which adds a literal suffix for (signed) size_t, so in C++23 we will be able to write literals like 100uz. (I wonder whether uz will be pronounced Uzi.) See the many excellent side-by-side examples in JeanHeyd’s paper for how this helps make uses of size_t safer and more convenient especially in naked for loops iterating over containers. Congratulations to JeanHeyd for C++23’s first language extension, and also for his persistence with this paper – the adopted version is revision 8, and that number and the paper’s change history indicates the level of rigor that can be required to get a feature into C++. Many thanks!

P1679 by Wim Leflere and Paul Fee add a basic_string::contains function so we can write code like if (str.contains(substr)) std::cout << “found!\n”; … I can already hear the chorus of “finally!”

P0881 by Alexey Gorgurov and Antony Polukhin add a stacktrace library to C++23. This is a much-anticipated extension based on Boost.Stacktrace that will enable much easier-to-debug portable diagnostic messages.

The charmingly numbered P1048 by Juan Alday gives us an is_scoped_enum type trait to detect when an enumeration is defined using the new-style (C++11, but well it’s still “new”!) enum class. As the paper points out, this is particularly useful as a migration aid, including to write code that detects and measures the adoption of “enum class” over plain old “enum.”

Finally, P0943 by Hans Boehm supports C atomics (spelled _Atomic) in C++ where the two did not already overlap, which helps write headers that work in both C and C++. (The adopted version is R6 which should be published in the next few weeks.) This is one example of the ongoing extra coordination we’ve had lately between the C and C++ committees, which leads to the next thing we did…

New SG22: C/C++ liaison

We appointed a new study group, SG22, for C/C++ liaison. This is a unique study group, because it is shared jointly by both the C and C++ committees, and it continues the tradition of closer coordination between the two committees. Thank you to WG14 (C) and its chair David Keaton for their continued interest in coordinating the two languages, to Aaron Ballman for agreeing to chair this new group, and for our WG14 and WG21 project editors Thomas Köppe, JeanHeyd Meneide, and Richard Smith to serve as assistant chairs. Thank you all for being willing to step up!

Other updates

Thank you to Richard Smith for his work for many years as project editor for the C++ standard, and completing C++20 this month! Thank you also to the many of you who have helped Richard and shared the editing workload by providing PRs and proofreading to apply plenary resolutions; that has been very much appreciated by Richard and by all of us especially given that C++20 is a “big” release with many new features, all of which have created an unusually high amount of editing work for this release.

Starting now, as we begin C++23, Thomas Köppe has graciously agreed to step up to be our primary project editor for the standard, with Richard as backup project editor. Thank you Thomas, and thank you again Richard and to all who have helped with the editing for the C++ IS!

Next steps

While we are meeting virtually until further notice, we will continue to have virtual plenaries like the one we had this week to formally adopt new features as they progress through subgroups. Our next virtual plenary will be in February, on the Monday of what would have been the Kona meeting.

Progress during this time will be slower than when we can meet face-to-face, and we’ll doubtless defer some topics that really need in-person discussion until we can meet again safely, but in the meantime we’ll make what progress we can and we’ll ship C++23 on time.

Thank you again to the hundreds of people who are working tirelessly on C++, even in our current altered world. Your flexibility and willingness to adjust are much appreciated by all of us in the committee and by all the C++ communities! Thank you, and see you on Zoom.

My plans at CppCon

It’s hard to believe CppCon 2020 is nearly here… in fact, pre-conference tutorials are already in progress.

I’ll be at the conference throughout the week in the hallways and session rooms. Here are some of the times I’ll be participating on the actual program:

  • Sunday 1300 MDT: Organizer’s Panel. In the middle of the Welcome Reception, we’re holding an Organizer’s Panel where several of us organizers will be talking about what to expect in the week ahead and available for extensive Q&A to answer any questions you may have. (We will have live music during the whole Welcome Reception, so be sure to watch the chat window for how to join that stream if you want to listen to our CppCon house band leader, Jim Basnight, perform for us.)
  • Monday 1500 MDT: My AMA. I’ll be available for an “ask me anything” session. Attendees can ask questions across the board from C++20 features, to how the committee is working during the pandemic on C++23, to specific C++ features I’ve designed or contributed to like concurrency or structured bindings or the spaceship operator, or other things you may think of.
  • Tuesday 0730 MDT: Committee Fireside Chat Panel (moderator). This year, I won’t be a panelist myself, so I don’t plan to answer any questions. Instead, I’ll be the panel moderator, and we have a great slate of panelists again this year: Bjarne Stroustrup (of course), Bryce Adelstein Lelbach (library evolution subgroup chair), Hana Dusíková (compile-time programming subgroup chair), Inbal Levi (Israel national chair), JC Van Winkel (Netherlands national chair and teaching subgroup chair), JF Bastien (language evolution subgroup chair), Michael Wong (low-latency/gaming/embedded subgroup chair and AI subgroup chair), and Tony Van Eerd (expert in many subgroups and popular speaker). I can’t wait!
  • Wednesday 0730 MDT: Bjarne Stroustrup AMA (moderator). This is Bjarne’s AMA, I’m just the moderator.
  • (will be recorded) Friday 1330 MDT: Empirically Measuring, and Reducing, C++’s Accidental Complexity. This is my one actual talk, and it’s the last talk of the conference. It will be a major update of the talk I’ve given publicly one time before in Prague earlier this year, which after a broader intro focused specifically on parameter passing as an example of where we could dramatically simplify C++. This time I’ll include lots of updates, including that I hope to demo a working compiler implementation of the proposal.


  • Denver time zone (MDT) is the default, which is where the physical CppCon usually happens. The view lets you show the schedule in your own time zone.
  • The conference starts Sunday. You don’t have to wait for Bjarne’s opening keynote on Monday… if you’re attending, be sure to make use of the Open House from 0900-1200 MDT where you can wander around, and especially the Welcome Reception from 1200-1430 MDT which includes the first panel (see above).

I look forward to seeing many of you there!

C++20 approved, C++23 meetings and schedule update

A couple of interesting things happened in the ISO C++ world this week…

C++20 passed unanimously, on track to publish later this year

On Friday September 4, C++20’s DIS (Draft International Standard) ballot ended, and it passed unanimously. This means that C++20 has now received final technical approval and is done with ISO balloting, and we expect it to be formally published toward the end of 2020 after we finish a final round of ISO editorial work.

As always, we are not counting on ISO’s publication speed to call it C++20, it’s C++20 because WG21 completed technical work in February. If for some reason ISO needs until January to get it out the door and assigns it a 2021 publication date, the standard will still be referred to as C++20. That is already its industry name, and 300,000+ search hits can’t be (retroactively made) wrong!

ISO C++ meetings are virtual until further notice, Kona postponed

A month ago, I notified the committee that our face-to-face meetings will be postponed until further notice. We still need to plan for face-to-face meetings so that we’re ready to resume when that’s possible and safe, but for now all currently planned meetings should be viewed as “tentative.”

Among other constraints such as national and corporate travel restrictions, we are subject to face-to-face meeting bans from several parent organizations. Two of those extended or enacted face-to-face meeting bans this week, on Tuesday September 1:

  • INCITS, the U.S. standards body, extended its face-to-face meeting ban through March 31, 2021. This means that our Kona meeting planned for February is now formally postponed to an unspecified future date.
  • ISO SC22, our corner of the international organization for standardization that handles programming languages, resolved to ban face-to-face meetings of more than 100 people until further notice. Since our meetings lately have regularly seen over 200 attendees, we’re currently evaluating how this affects future post-Kona tentative meeting plans.

All of these bans are subject to further extension, and we won’t meet in person again until it’s safe to do so. As of this writing, our next tentative face-to-face meeting would be the rescheduled Varna meeting, in the first week of June 2021, but that should be viewed as the earliest possible resumption of meetings. As the pandemic develops and INCITS and ISO meeting bans and other restrictions are extended, it’s certainly possible that we may not be able to meet again in 2021 at all. We’ll see.

In the meantime, though, we’re still making progress on our work: For several years, we have already been holding regular virtual meetings for some of our subgroups, including study groups (SGs) and CWG and LWG (language and library specification wording). Since the pandemic started, EWG and LEWG (language and library evolution, our primary design subgroups) have also begun meeting virtually, and we are continuing to adjust our process for how to approve design changes to progress proposals while not meeting in person. And starting in November, we will begin having virtual plenary (whole-group) meetings to formally approve changes, including potentially new features, to the C++23 working paper…

C++23 schedule and priorities

The C++23 schedule (P1000R4) and C++23 priorities (P0592R4) are unaffected by the pandemic. You may find this surprising, but that’s because the committee is on a “train model” that focuses on schedule and priorities for each release, instead of a specific feature set. One of the benefits of the train model is that it is very resilient, and can handle even major disruptions without change. We have already been in the mode of working on features all the time, including long-pole features that take many years, and each regular release train includes “whatever’s ready” with the next train opening up as soon as the previous one ships. So, that is unchanged.

What has changed, of course, is the speed at which we can work on features during the coming period. The pandemic disruptions have impacted all our lives, and reduced the time and energy WG21 participants have for standards work as well as our capacity to make progress face to face three times a year, and this has slowed down development of features we’re working on now that will land in { C++23, C++26, C++29 } . No virtual process will fully compensate for the lack of intense week-long face-to-face meetings, but as usual we’ll continue to make progress on baking features according to the P0592R4 priorities, including issue resolutions and an emphasis on completing C++20, and as usual we’ll load each feature into the currently loading train as the feature becomes ready. So progress continues, and the trains will continue to run on time to ship everything that’s ready.

Of course, the ISO C++ committee isn’t the only part of the C++ world that has “gone virtual” this year. We’ve been enjoying many virtual conferences, and just a week from now we’ll start the biggest C++ conference of the year: CppCon 2020, all online. I look forward to seeing many of you there, including literally seeing you at the video chat tables and in my AMA Q&A session early in the week, and the Committee Fireside Chat panel on Tuesday.

Thanks for your interest in C++ and C++ standardization! Be safe, everyone.

C++ on Sea video posted: Bridge to NewThingia (extended)

Two weeks ago, I had the privilege of speaking at the C++ on Sea 2020 virtual conference. The video of my talk has now been posted — it’s an extended version of the talk I gave at DevAroundTheSun in April. You can find it here:

Thanks very much to Phil Nash and all the other organizers and volunteers who made the virtual event run so smoothly! And thanks also to the post-processing team who did all the videos, including that they did a smooth job of fixing my video hiccup in the middle when I lost my link and had to reconnect to resume (entirely my pilot error).

A software note: I really enjoyed the Remo platform that C++ on Sea used this year. It allowed for lots of attendee interaction that’s surprisingly like mingling at a physical conference… you are in what looks and feels like a room with lots of tables, just like at a conference, and can easily talk face to face (video and audio) with all the folks who are at your table, and move from table to table to mix and mingle. It really felt seamless. I enjoyed spending time talking with some of you there that way — it was great to meet new people as well as see some familiar faces, and I look forward to doing it again soon this September because the current plan is for the also-all-virtual CppCon 2020 to use the same software. So, I hope to literally see many of you then! In the meantime, I hope you enjoy this talk.

AMA @cpp_russia on July 2

undefinedC++ Russia is an online event this year, and I’m happy to be one of many C++ folks to be invited to participate. On July 2 I’ll be doing a Q&A session, which is the first time I’m doing an “AMA” — no talk, just Q&A and discussion. I’m looking forward to it, and to see what kinds of questions are on people’s minds.

And don’t miss Bjarne Stroustrup’s similar AMA-style session at the same event, two days earlier on June 30!

Talk video available: Bridge to NewThingia @DevAroundTheSun

Last month, I gave a new talk “Bridge to NewThingia” at DevAroundTheSun. Using examples from the evolution of programming languages and a few other tech products, it analyzes some key design factors that let you confidently answer the question, “why will your NewThing succeed, when a lot of things that look like it have failed in the past?”

The talk video is now available on YouTube, linked below. I hope you enjoy it, and thanks again to the organizers of DevAroundTheSun for putting together this global online event.