ISO/IEC JTC1 SC22 WG21 N4556 - 2015-10-09
Marshall Clow, firstname.lastname@example.org
Time: 08:00am N.Am. Pacific Time
Duration: 2 hours
Agenda is http://open-std.org/JTC1/SC22/WG21/docs/papers/2015/n4552.pdf
1. Opening and Introductions
Sutter called the meeting to order 8:05 Pacific Time.
J. Daniel Garcia
Gabriel Dos Reis
1.2 Adopt the Agenda - by UC
1.3 & 1.4 deferred to face to face meeting
1.5. Review of project editor & Liaison assignments
Herb has a correction - we don't have a liason for WG23.
2.0 Status, Liason, & AI reports
Core (Mike): 3 or 4 telecoms since last meeting; dealt with concepts TR ballot resolutions. 2 issues processing telecom.
25 issues in ready status, which we anticipate moving. Will not be moving any Tentatively Ready, due to document handling changes. Going forward, we will have a document in pre-meeting mailings detailing Tentatively Ready status so that we can move them at a meeting.
No more core issues list in the mailing - saving on size at the mailings. Core will be producing a document at the meeting that will be moved at the meeting containing just the issues that are to be moved. This will (hopefully) encourage people to review the issues, and allow Core to pull back any issues that have had problems discovered at the meeting.
Plans for this meeting:
* Issues processing as time permits
* Five papers on the core wiki now.
* Papers that come from EWG.
Q: Mike - Is Andrew planning on bringing issues to Core at this meeting? A: I'm not sure.
Q: John - Maybe some notification when the Core issues list changes? A: There will be a revised Core Issues list on the Core Wiki, and sent to Keld for posting on his site.
Q: Gaby - I think it's valuable to make the issues list available to the larger C++ community. A: Yes, I plan on updating the Keld's site right after the meeting.
Q: Walter - Does this revised procedure violate what we generally call "the 2 week rule"? A: Herb: The 2 weeks rule is from INCITS, so that's not really germane; though we do like it. All the non-core issues lists are on GitHub. Mike: All the ready issues that we are bringing forward were voted to ready at the previous meeting. The Tentatively Ready will go into the pre-meeting mailing. Ville: There's a difference between processing issues and moving issues; the latter requires more time and notice.
* 33 papers in the pre-meeting mailing; most are revisions of earlier papers.
* Priority is for large changes that are shooting for C++17.
* Second priority is for revisions of older papers.
* Ville is looking for guidance from the larger committee about what will be going into C++17
Herb asked to defer that to Monday
Daniel: Ville and I have discussed having a joint session between SG1 and EWG about coroutines. There are several papers about coroutines.
* six telecoms since the last meeting; 3 on issues processing, and 3 on Eric's ranges proposal.
36 issues in ready status, which we anticipate moving. Like Core, we no longer publish issues list in the mailing, but will publish Tentatively Ready issues in future pre-meeting mailings. LWG's issues lists are available on github at http://cplusplus.github.io/LWG/lwg-active.html. Like Core, LWG will be producing a document at the meeting which contains just the issues to be moved.
Plans for this meeting:
* about 10 papers directly to LWG - a couple will move to SG1
* Issues processing
* Papers that come from LEWG.
Q: Herb: Is there an idea for adopting the Networking paper as a working draft, and then sending it out as a PDTS at the end of the meeting? A: That's a possibility, but it will depend on what people think during the week.
LEWG (Jeffrey - not present)
2.2.1 Liaison reports
* We have a full agenda
* Handling the Concurrency TS ballot resolution issues; plan is to vote it out as a TS. Run the changes past LWG.
* Large collection of papers.
* Papers P0075 and P0076 were assigned to LWG, but SG1 should look at them first
SG2/3/4 - completed work
SG5 (Michael by email)
We are still working on several new proposal such as on_commit handler, and retry for the next TM2 proposal based on real-world industry experience with TM from Wyatt Technologies. No meeting this week. We continue to hold official meetings every 2 weeks on telecon. We do not plan to have a meeting at Kona.
* Evaluate and outline the current proposals, then papers as time permits.
SG7 (Chandler not present):
SG8 now in Core
SG9 now in LWG
* Will be having an evening session Tuesday to talk about their charter.
* One Core issue from Lenexa
* No papers that directly address UB
* One morning or afternoon meeting this week
SG13 now in Core
SG14 (Michael by email):
We had 2 official telecon calls and 1 F2F meeting at CPPCon. All very successful. We plan to continue monthly telecon to conduct paper reviews. At Kona, we will have a Wednesday night evening session to update everyone.
Please see report here:
We have 8 papers
P0037R0 Fixed point real numbers John McFarlane LEWG SG14/SG6
P0038R0 Flat Containers Sean Middleditch LEWG SG14
P0039R0 Extending raw_storage_iterator Brent Friedman LEWG SG14
P0040R0 Extending memory management tools Brent Friedman LEWG SG14
P0041R0 Unstable remove algorithms Brent Friedman LEWG SG14
P0048R0 Games Dev/Low Latency/Financial Trading/Banking Meeting Minutes 2015/08/12-2015/09/23 Michael Wong SG14
P0059R0 Add rings to the Standard Library Guy Davidson LEWG SG14
P0130R0 Comparing virtual functions Scott Wardle, Roberto Parolin EWG SG14
We had a major brain storm on why games programmers tend to turn off exception handling, and a brain storm session to see what kind of options and ideas to address their concerns.
We also briefly discussed the scope and coverage and decided that because of the common interest of
Real Time Graphics
Which is of interest to Games Dev, but they also intersect Financial Trading, Simulation and Embedded devices.
As such we propose to rename the group Games Dev/Financial Trading/Banking/Simulation and invite representatives from these industries to join.
As such we intend to bring to discussion in addition to Exception/RTTI costs, topics regarding the following, some of which crosses into other groups such as SG1 Concurrency, SG6 Numerics and SG7 Reflections.
There are a few papers in the mailing that are of interest to this SG14.
GPU Accelerator support
P0069R0 A C++ Compiler for Heterogeneous Computing
We actually intend to study the design of C++AMP, OpenMP Accelerator/OpenACC, OpenCL, OpenGL and Vulcan for a GPU accelerator design to support gamers. We would like to review this paper under a co-located SG14 session which can run with SG1, but we are interested in taking this work further.
P0089R0 Quantifying Memory-Allocation Strategies
Coroutines and games
P0054R0 Coroutines: reports from the fields
P0055R0 On Interactions Between Coroutines and Networking Library
P0070R0 Coroutines: Return Before Await
P0071R0 Coroutines: Keyword alternatives
P0073R0 On unifying the coroutines and resumable functions proposals
P0099R0 A low-level API for stackful context switching
low latency for financial/trading
P0076R0 Vector and Wavefront Policies
P0106R0 C++ Binary Fixed-Point Arithmetic
2.2.2. SG22 report
* Plenary was in Wash DC a couple weeks ago.
* There was a large discussion about how Linux was to be handled.
* Papers are available
2.2.3 SC22/WG14 (C) report
* Meeting adjacent to us in Kona (the week after)
Q: Herb: Is there anything to happen at the WG14 meeting that will affect us? A: Tom: Not that I can think of?
Q: Ville: What about the Transactional memory stuff that WG21 is talking about? A: Hans: It's a somewhat slimmed down version of what we are proposing.
3. New Business
3.1 Review priorities http://isocpp.org/std/status
* Marshall - we also want to (at least) start a Ranges TS working paper
Q: Clark: Should we just fold ranges into LFTSV2? A: Marshall: There's a lot of stuff in the ranges proposal, and it's tightly focused. Walter: Ranges depends on concepts. LFTS does not. Let's not mix that.
Walter: We should start thinking about the ranges proposal as "STL2". It has some fundamental improvements to the current standard library. Ville: Eric has stated (in his talk at CppCon) that he's working towards a new standard library. (post c++17)
Herb: That was the interest in Urbana and Lenexa.
Herb: Ville has been canvassing group chairs about what should/can go into C++17. We want to have a discussion in Kona; but we need to set out some ground rules so that the discussion will not spin out of control. Will bring this up on Monday, as it will affect what is done during the week.
Ville: Re: the Language features that people want to get into C++17, the current status of the proposals is concerning. There is lots of work still to be done on these proposals
Clark: Do we want to add something to the agenda to let people know that we'll be discussing this? Ville: I think it would be reasonable to do this. Herb: Clark published N4545, and it will fit under items 3 and 4. Clark: When the minutes for the meeting come out, I'm going to have an updated agenda. Herb: At least, have a bullet at the top "By the way, we're going to be talking about features for C++17" or add it to item #3.
Herb: We have at least one paper that proposes pulling part of a TS into the standard. Lawrence: The TS's are to gain experience; before pulling a TS into the standard, we should ask what experience has been gained.
3.2. Review of the mailings
* We've categorized the mailings
3.3 Other Business
4.1 Review and approve resolutions
4.2 Review and approve action items
* Herb: take the SG22 / WG14 report and defer to face to face
Mike: Previous guidance was that people who were interested should look at the minutes, and ask questions in plenary. Herb: Ok, then we'll continue to do that.
Ville: Wants to not have a major C++ conference the same week as the mailing deadline.
Herb: Yeah, we don't want to do that again.
5 Closing processing
* Add SG14 to the agenda for next time - unanimous consent.
Next fall meeting: Marshall says that it is currently unlikely that Qualcomm will host a meeting next fall; the ongoing reorganization makes it unclear who can make such a decision.
* Paper management.
** Herb: Would like to have a centralized place for papers. He suggested Github, but he's not insisting on it. More discussion in Kona.
Ville: Is there some preference when the information for the June meeting available?
Herb: Sooner would be better; in the post-Kona mailing?
Ville: That's what I'm hoping for.
Herb: People need to know basic logistics: exact dates, hotel, airport, etc. as soon as possible so that they can make plans.
Walter: Does that have to be an N-paper?
Herb: Yes. That's an WG21 administrative paper.
Clark: What about the JAX meeting announcement
Barry: That's just about ready.
Clark: We don't have to wait for the post-meeting deadline; livelink will just push it.
5.2/3 - deferred to face to face.
5.4 Meeting adjourned at 9:35