ISO/IEC JTC1 SC22 WG21 N4489 - 2015-04-29
Jonathan Wakely, cxx@kayari.org
Teleconference information: Date: 2015-04-24 Time: 8:00am N.Am. Pacific Time Duration: 2 hours
Sutter called the meeting to order 8:05 Pacific Time.
In attendance were:
The agenda in N4427 was adopted by unanimous consent.
The minutes in N4362 are from the Skillman meeting. The minutes from the previous telecon were approved in Skillman. Meredith and Miller commented that it seemed strange that telecon minutes are approved at a face-to-face meeting, and vice versa, when the different meetings could have very different participants.
The previous telecon minutes were N4349, although Meredith suggested that since that was a ballot resolution meeting that N4250 were the most appropriate minutes.
Sutter suggested deferring approval of all minutes to face-to-face meetings, to ensure the set of people at each meeting can be properly represented. Miller and Brown agreed with the proposal. Sutter questioned whether administrative telecon minutes should be approved at the next face-to-face meeting as well. Meredith and Brown pointed out that the minutes for the current telecon won't be in a mailing before Lenexa, but Sutter said they will be posted to isocpp.org sooner.
Let the minutes reflect adoption of a new procedure to adopt all recent meeting minutes at the start of each face-to-face meeting.
Sutter suggested that as the standard ISO agenda format requires an item to approve minutes that we can add a placeholder saying "defer to next meeting".
Action: Sutter to update telecon agenda template.
The previous action items were all for Urbana, and were all completed.
Networking: Sutter updated the records to reflect that Wakely is the project editor.
Library Fundamentals: Sutter confirmed that Library Fundamentals v2 has the same project number and editor as v1.
Arrays: Crowl said a meeting is needed, maybe in EWG, to decide what to do. The current content is mostly ready, bar editorial issues, but it's open to debate whether it will be published at all. Nelson confirmed that N3820 was adopted in Issaquah.
Clow suggested asking in Lenexa for volunteers to be WG14 liaison. In the interim Plum and Nelson were appointed as liaisons. WG14 liaisons will be discussed in plenary on Monday in Lenexa.
Sutter reiterated the ISO rule change that removes the DTS stage and explained that the Library Fundamentals TS and Parallelism TS are ready for publication following the PDTS ballot. If there are no objections in Lenexa they will go straight to publication. After ballot resolution the Concepts TS will be ready too publish too. Meredith pointed out that the Transactional Memory TS has also completed its PDTS stage. There will be a discussion on Monday in Lenexa to confirm there are no objections to those four projects going to publication.
Boehm said there was a ballot comment for the Parallelism TS which there was no time to review in LWG at the last meeting and asked whether it should be reviewed in Lenexa and possibly make changes. Sutter suggested bringing it up in the WG21 plenary. Boehm explained that the comment was about relaxing requirements on exception handling, not requiring all exceptions to be bundled.
Action: Boehm to attach the comment to the SG1 Lenexa wiki.
Brown noted that there may be a large number of straw polls on Monday so it would be useful to have a list of those polls on the wiki in advance. Nelson suggested adding them to the agenda.
Action: Nelson to update Lenexa agenda.
Meredith asked whether papers that got LWG approval in Cologne would have a straw poll on Monday, or would wait for Friday as usual. Sutter stated that we don't do technicall votes on Monday and they would be polled on Friday, so that if someone wants to object during the week they can come out of ready status before the poll. Yasskin pointed out that the latest WP for Library Fundamentals v2 (N4481) is a "tentative working paper" including proposals approved in Cologne. If those changes are not approved first then the previous WP will have to be approved in the Monday plenary instead. Meredith said that for a Library Fundamentals v2 PDTS to come out of Lenexa the updated WP including the Cologne proposals would have to be approved on Monday. Sutter requested that the WG meeting reports in Lenexa should mention this.
Meredith pointed out that the new Concurrency working paper (N4399) also includes "tentatively ready" work from Cologne that hasn't been approved in full committee. Wakely expressed concern that some wording in the WP was still not ready.
Action: Wakely to send comments to SG1 before Lenexa.
LEWG will have a session about Networking design changes, which won't have time to go through LWG and be ready for PDTS at the end of the Lenexa meeting.
Sutter asked Yasskin if there are plans to propose moving a completed TS such as Filesystem into the C++ WP. Sutter suggested that could be done with a simple "move everything from the TS into the WP" motion on Friday. Orr reported that the BSI would like to have an actual paper for anything that gets moved into the IS, to make it clear that there is an extra stage to go through. Yasskin agreed and added that having the time for a paper to be in a pre-meeting mailing for review is important. Voutilainen pointed out there would need to be changs to namespaces for all library TS components to move them out of namespace std::experimental. Orr said he also expected some wording would need to be added to Annex C, that isn't present in a TS. Nelson said that papers proposing moving a TS into the IS should justify it in terms of implementation experience.
Brown said N4437 might be a good model for proposals to move existing work into the IS. Sutter assumed that LEWG would be reviewing N4437 in Lenexa and LWG would get a chance to review it if needed, but pointed out that the wording is already in an International Standard. Meredith suggested LEWG should review N4437 for use of newer C++ features that weren't present when it was published. Sutter suggested such changes would be easier to make against wording already in the C++ WP. Meredith said he would rather see fixes made before adding wording into the WP that is known to need changes. Clow would like a paper saying "adopt with the following changes". LEWG can review the Special Math TS during Lenexa, and as no changes are expected, have a proposal to adopt it ready by the end of the week.
Action: Sutter to call attention on Monday to N4437 as an exemplar for moving published work into the IS.
Nelson suggested it would be nice to see papers in the mailing categorized into broad themes. Others respondedthat this was not possible as no-one sees all the papers until the mailing. Working group chairs will group related papers and schedule things logically. Yaskin invited people to create LEWG wiki pages. Voutilainen said he would be preparing the EWG wiki and requested that others not make changes to it before the meeting. Clow requested that people email him with any suggestions of groupings for LWG papers.
Miller said not all Concepts PDTS ballot comments have been received, so it won't be possible to address them all at Lenexa. Comments that have been received can be addressed, but any later ones will need to be discussed in a telecon after Lenexa, or in Kona. If there are more comments expected Miller requested informal comments be attached to the core wiki page or emailed to him.
Some US comments deal with design level issues, so it may be necessary or at least desirable to have EWG discuss some of them. In particular N4434. Any changes approved would need discussion whether they should be applied immediately, and whether another PDTS ballot would be needed then.
Clow relayed that Chandler Carruth had requested an evening session for reflection early in the week. Bjarne Stroustrup had also suggested one on the big themes of C++17. Sutter suggested C++17 on Monday, as it might influence work during the week. Clow points out Carruth is also hoping the evening session will guide work during the week. Clow would like an evening session for issue processing.
Voutilainen asked whether we should have any working group status reports. Sutter noticed he had cloned the wrong agenda for the telecon, so we had missed the status reports and suggested doing them under AOB.
Miller reported that there are 5 issues in ready status from Urbana, and 22 tentatively ready status from an intervening telecon. There are some new issues, number unknown so far. Core will review the TM ballot resolutions comments, and Jens Maurer has a paper with responses which will probably result in a motion to change the TS. Miller expects to spend time reviewing Concepts ballot comments, which include Brown's N4434 paper, several editorial comments and a few technical ones. At the end of the week it will be decided if the changes are minimal enough to go straight to publication after applying the changes or whether to have another PDTS. Some core resolution issues that are fairly substantial which will be looked at. He expects some papers to get sent from EWG during the meeting.
Clow reported LWG has 16 Tentatively Ready and 18 Ready issues following Cologne. There are 106 new issues to prioritise and work through. Seven papers were reviewed in Cologne and should be ready for straw polls in Lenexa. There are at least 20 new papers to prioritise and work through.
Voutilainen reported three dozen papers in the pipeline, TM PDTS comments are higher priority but only one targets evolution. EWG will also be looking at Concepts ballot resolution comments.
Yaskin reported there are 15-20 new papers. Two large-ish issues that will have long sessions, including networking design questions. He will try to insist on good wording coming out of LEWG instead of relying on LWG to do it.
Boehm has 17 or 18 papers in SG1. There has been a suggestion of a joint session with LEWG about vector extensions.
SG2 modules, not meeting in Lenexa, work now moved to EWG.
SG3 filesystem also not currently active.
SG4 networking, work has moved to LEWG.
SG5 TM, no status report, PDTS done.
SG6 numerics. planning to meet in Lenexa to pull together everything wanted in a TS.
SG7 reflection, there will be an evening session in Lenexa.
SG8 concepts, work has moved out of the SG.
SG9 ranges, work has moved to LEWG, no separate meetings are planned.
SG10 feature tests, the SG is hoping to meet. There has been lots of work done to keep up with the C++17 changes in Urbana. Nelson will want to meet with core and library to shift responsibility for coming up with recommendations away from SG10.
SG11 DB, inactive.
SG12 UB, no status report.
SG13, I/O, there is no updated paper ready for Lenexa, but one will be ready for Kona.
Nothing to be reviewed (this item should not have been on the agenda).
The action items were summarised.
Deferred to face-to-face meeting.
Deferred to face-to-face meeting.
Deferred to face-to-face meeting.
Sutter adjourned the meeting.