Minutes of March 2008 Meeting

Austin/418
Minutes of the 14th Plenary Meeting of the Austin Group
Mar 04 - 05, 2008, The Open Group, San Francisco, CA.

Attendees

Andrew Josey called the fourtheenth meeting (a.k.a. Austin/M15, since this counting includes a teleconference) of the Austin Group to order at 9:13 am Tuesday, March 4th at The Open Group facility, San Francsico, CA.

Meeting Goals

The goal of this meeting is to review draft 4 comments for the 200x revision, and continue processing inbound defect reports so that they can be addressed for the final draft.

The main part of the meeting will be spent doing issue resolution based on Aardvark comments submitted against Draft 4 and also those outstanding on the 2004 edition of the standard. Attendees are assumed to be familiar with the proposed dispositions for the aardvark prior to the attendance of the meeting - these will not be discussed unless there is some disagreement with the proposal.

1. Administrivia
   • Introductions

     All the participants introduced themselves.

   • Approval of Agenda

     The agenda was approved as published.

   • Approval of Minutes

     Minutes of the last plenary meeting (Austin/386, Sep 04-07, 2007) were reviewed. Approved with no objections.

   • Review of Action Items. See Austin 387.
     1. ACTION 2007-09-07: Ulrich and Nick to write a set of guidelines on how to shutdown a multi-threaded C++ program with global variables prefereably for the Kona mailing (Sep 9). CLOSED.
     2. ACTION 2007-09-08: Nick & Herb Sutter to develop a C++/POSIX liaison update for SC 22.
        CLOSED
     3. ACTION 2007-09-09: Mark Brown to review 2004 XSH ERN 147 and make a proposal. CLOSED.
     4. ACTION 2007-09-10: Larry to provide details of how HP/UX behaves w.r.t. POLLHUP, contrasting this with Solaris/Linux behavior. CLOSED OBE.
     5. ACTION 2007-09-11: Mark to verify Mark Harris's assertion that AIX behavior w.r.t. POLLHUP differs from Solaris and Linux. CLOSED OBE
     6. ACTION 2007-09-12: Andrew to update the roadmap and workplan with the new dates. CLOSED.
     7. ACTION 2007-09-13: Don to provide updates to ENAMETOOLONG aardvark (2004 XBD ERN 100) with D3 page and line numbers. CLOSED.
     8. ACTION 2007-09-14: Andrew to publish draft interpretations from 2004 aardvarks closed during M/14. CLOSED.
     9. ACTION 2007-09-01: Andrew to update AI-06 with Austin/382 and distribute for comments. CLOSED.
     10. ACTION 2007-09-01: Andrew to update the issues list with the resolution of issue 15 and 16. CLOSED.
   • Review of Issues List

     SD1 September 10 2007. All issues are currently resolved.

   • Review of Consent List

     SD2 September 10 2007 - no updates.

2. Procedures Update
   • JPC Procedures
     No updates to report. IEEE PASC have a new set of procedures but there are no operational changes from our perspective.
   • Committee Draft development procedures.

   No updates.

3. Identification of Organizational Reps

   See attendance list above.

4. Status Reports

   • Austin Group Status; 530 people on the reflector. 4 snapshots releases went out between d3 and d4. 19 aardvarks submitted against D4.

   • Organizational Reps Status

     • ISO: This is a meeting of the ISO editing group. No PAG meeting will occur this week. Both ISO C and C++ are in revision.
     • IEEE: New procedures (see above). No issues.
     • TOG: No new news.
   • IEEE PASC Status
     

     New WG, 1003.27, for POSIX-C++ Binding. Separate report later on agenda. New PASC procedures avaliable.

   • ISO SC22 Status
     Nothing to report.
   • IEEE/The Open Group Status
     See above.
   • ISO C Status.
     C is starting a revision. Possible changes to their library section. Attributes likely to be added. The auto keyword will likely be deprecated. Unicode characters are being considered, but have serious problems. ACTION: Larry and Nick to develop a position paper on char16 and char32 issues for the pre Delft WG 14 mailing (14 March 2008).
   • ISO C++ Status.
     C++ revision is complaining that the POSIX influence is causing too much instability in their standard.
     
   • POSIX C++.
     1003.27 is now officially under way. Reviewed the current C++ table of contents and looked at the places where there is overlap with POSIX. This led to a list of liaison points to send to C++, which were delivered. There will be a meeting in Reading in June, following the C++ meeting in Nice (Sophia Antipolis). This needs to get a document outline
   • POSIX Certification and test Status
     QNX have certified.
   • Editors Status
     A few minor whitespace issues with known workarounds, but otherwise OK. Nothing new to report.
   • Interpretations Status
     All interps are currently closed. There may be a few left to be started down the path from 2004 aardvarks, but these may go to TC1.
   • Networking Status.
     Nothing to report.
   • Joint Procedures Committee Status
     Nothing new.
   • LSB 3.2 now out. Perl and python added. Updated Qt library. Printing support. Bug fixes to core. 4.0 under way. There is an ongoing alignment with POSIX work item. A revision of the ISO standard is promised, and 3.2 will probably be the basis for this.
5. Old business
   Nick reminded everyone about the null pointer C liaison question. There is proposed response from Tom Plum (austin-sep2007 # 76). This is a complex rewording. We certainly need to drop the informative note from 3.243 (action completed by Cathy for Draft 5). Geoff noted (in #77):

   "It seems to me that the additional part about the value of NULL is only of relevance to the C Standard. Now that we require NULL to have type void *, it would not be needed in SUSv4 (TC1) as the value of NULL already meets the existing definition of a null pointer (it is a pointer obtained by converting an integer constant expression with the value 0 to void *)."

6. New business
   • Timeline / workplan:
     Review meeting this week. Sanity D5 April 1. Review for 2 weeks. D5.1 at end of April. Start DIS ballot in May. IEEE final ballot at the same time. Publish in July/August. Anything that might destabilize the draft should be targetted at TC1 from now on.
   • Open Interpretations
     All current interpretations are resolved. There are still a small handful of open 2004 aardvarks that might go down the interps path. What happened to interp 122? Still in "proposed" status, similarly interp 125, although the text says that is finalized. ACTION: Andrew to investigate the status of Interps 122 and 125. Wednesday update: see below
   • Open D4 Aardvarks
     
     • XBD
       1. Multi volume change from Don. Better editorial representation of character constants. Accept.
       2. At editor's discretion. Possibly use an empty book name for the front matter. No line numbers before ToC.
       3. Proposed resolution from teleconference. Use this.
       4. Affects readability, but after much deliberation: Accept.
       5. Proposed resolution from teleconference. Use this.
       6. This issue is linked to 2004 aardvark XSH 234. Aardvark report currently has this marked OPEN, but minutes from Feb 15 teleconference say ACCEPT, Interp standard is clear standard is wrong. Use proposed editors notes from 2004 XSH 234 as basis for accept as marked for this draft.

          Change to XBD ...
          
           At page 403 line 13554 section sys/wait.h, after WUNTRACED add
          
               "[XSI] WCONTINUED   Report status of continued child process.[/XSI]"
          
           At lines 13567-13568 delete WCONTINUED and WNOHANG (and their
           descriptions).
          
           After line 13569 add 
          
               "The [XSI]WCONTINUED[/XSI] and WNOHANG constants, described above for waitpid(),
               can also be used with waitid()."
          
          Cross Volume Change to XSH page 2182 line 68921, Change
          
               "Return immediately if there are no children to wait for."
          
           to
          
               "Do not hang if no status is available; return immediately."
          
           At line 68932 change
          
               "If WNOHANG was specified and there are no children to wait for, 0
               shall be returned."
          
           to
          
               "If WNOHANG was specified and status is not available for any
               process specified by idtype and id, 0 shall be returned."
          

       7. Proposed resolution from teleconference. Use this.
       Break for lunch
     • XSH
       1. Considered adding fcntl.h to synopsis for fstatat(). No, don't. Looked at the unistd.h pages marked in aardvark with "??". Could not see reason for having this see also there in the first place. Remove the question marks. Use proposed resolution from teleconference.
       2. Proposed resolution from teleconference. Use this. It is a change to the proposed change; Cathy believes she can handle this without us altering the text in the aardvark.
       3. Proposed resolution from teleconference. Use this.
       4. ACTION: Ulrich to file an interpretation on the use of const in the setstate() prototype. Done. One thought was that was added for C99 alignment. However Andrew traced it back to being in the original submission in XSH 4 Version 2 (i.e. SUSv1). Accept.
       5. Proposed resolution from teleconference. Use this.
       6. Proposed resolution from teleconference. Use this.
       7. Proposed resolution from teleconference. Use this.
       8. Proposed resolution from teleconference. Use this.
       9. Proposed resolution from teleconference. Use this.
          
       10. (note an additional ERN 10 was added during the meeting). Accept
           
     • XCU
       1. Proposed resolution from teleconference. Use this.
       2. Proposed resolution from teleconference. Use this.
     • XRAT
       1. Proposed resolution from teleconference. Use this.
   • 2004 Open Aardvark
     Several of these are marked as resolved in some way, but have not had any action taken.
     • XSH
       1. XSH ERN 235. Accept. Since this is just an example, move to D5 as editorial (via SD5)
          
       2. XSH ERN 236.  Accept.  Make an interp (Austin #216) . This is not targeted for TC1 as it is the same change as XSHd4 ERN 4, so will be done in D5 via that route.
       3. XSH ERN 230. Marked as accept, but not as yet targeted as an interp. Not applied to D4. Make a new interp, and target this for TC1 (Austin #212).
       4. XSH ERN 229. Not marked as implemented, but it is in D4. This is interp #210.
       5. XSH ERN 231. Make an interp, and target for TC1 (Austin #213).
       6. XSH ERN 232. Make an interp, and target for TC1 (Austin #214).
       7. XSH ERN 233. Editorial. Include in D5 (via SD5)
          
       8. XSH ERN 234.  Make an interp (Austin #215). This is not targeted for TC1 as the same change will be done in D5 via XBDd4 ERN 6 (see above).
     • XCU
       1. XCU ERN 177. Printf hexadecimal escape. Needs to go down interps track (Austin #211). Might be hard to do this in a TC. Also take change from Konrad Schwarz, sequence 11310 Push down interps track for D5. No...looks like some implementations do not support this yet. General feeling is that this is top of the list for the next revision (i.e. 2015 or so!).
          Email from Geoff Clare:

          A couple of thoughts on this one...
          
          The proposed APP USAGE addition explains that hex escapes are limited
          to two characters after the \x, unlike in C, because in C it is possible
          to use string literal concatenation to terminate the sequence whereas
          there is no equivalent in the shell.  However, there is an alternative
          method that can be used.  E.g. if you want to output a character using
          hex escape \x44 and then output the digit 1, you can use:
          
              printf '\x44%s1' ""
          
          The problem statement says that hexadecimal escapes are commonly
          implemented as extensions, but at least one major implementation
          (ksh93 builtin printf) does not implement them in the way proposed for
          standardisation.  It does not have the two character limit:
          
          $ printf '\x441\n'
          A
          $ printf '\x44%s1\n' ""
          D1
          
          In light of this, we should consider allowing both behaviours.
          

          Looks like ksh93 just takes the last two digits. ACTION: Nick to propose new wording for hexadecimal escapes in printf that makes long strings of digits unspecified. App Usage/Rationale should highlight the difference in behavior between the common implementations.
   • Other Known Issues:
     Eric Blake submitted an issue against D3.3 for a problem with mkdtemp(). See 11255. Make this into a new D4 XSH aardvark, which can be accepted as editorial. GWC filed it as a new editorial aardvark. Mark as accept for D5. Break for the day at 15:57
   • Interp 122
     This has been out for a long time, and just got missed in the last round. Revised proposal was sent out last September. Some technical discussion on the merits or otherwise of the proposal. Privilege should not be required for using the monotonic clock. CLOCK_HIGHRES is not part of POSIX. The issue of resource consumption is a concern. If a very high resolution is required, the test for permission should be on the timer_settime() call if the requested resolution is too high and the caller does not have appropriate privilege. Leaning toward reject, with rationale suggesting that EPERM on timer_settime is permitted. Leave this OPEN for further discussion. ACTION: Nick and Ulrich to submit arguments against interpretation 122 to the interps group.
   • Interp 125. The interps site has this as "proposed", but the actual records show it is approved, and the action has been taken. Andrew has updated the plato site.
   • New Interps from 2004 Aardvark.
     Interps arising from XSH ERN 230, 231, 232, 234, 236 and XCU ERN 177.
     XSH 230 is Standard is ambiguous (interp #212)
     XSH 231 is Standard is silent (interp #213)
     XSH 232 is Standard is clear, std is wrong (interp #214)
     XSH 234 is Standard is clear, std is wrong (interp #215)
     XSH 236 is Standard is clear, std is wrong (interp #216)
     XCU 177 is Standard is silent (interp #211)
     
   • SD-5 changes.
     XSH ERN 233 and 235 and XRAT ERN 9 have been added to SD5 (editorial issues) and will be merged into draft 5.
     
7. Consent List:
   No changes.
8. Issues List
   No changes.
9. Action Item Review
   1. 2008-03-01 ACTION: Larry and Nick to develop a position paper on char16 and char32 issues for the pre Delft WG 14 mailing (14 March 2008). This will be discussed by email and also at the March 13th teleconference.
      
   2. 2008-03-02 ACTION: Nick to propose new wording for hexadecimal escapes in printf that makes long strings of digits unspecified. App Usage/Rationale should highlight the difference in behavior between the common implementations.
   3. 2008-03-03 ACTION: Nick and Ulrich to submit arguments against interpretation 122 to the interps group.
   4. 2008-03-04 ACTION: Andrew to update the roadmap as described in the Draft & Meeting schedule outlined below.
   5. 2008-03-05 ACTION: Andrew to create interpretation requests for XSH ERNS 230, 231, 232, 234, 236 and XCU ERN 177.
10. Draft & Meeting Schedule

    Draft 5 will fold in the 20 aardvark from the D4 review (19 prior to the meeting +1 added at the meeting), plus 3 editorial items added to SD5.  There will be a 3 week review  commencing 2008-04-04 until 2008-04-25. There will be a teleconference in week commencing 2008-04-28 to develop final instructions. Use normal Thursday slot 2008-05-01. D5.1 incorporates any changes from that (should be editorial only) and that goes for IEEE final recirculation and ISO/IEC DIS ballot. DIS ballot is 2 months (plus lead and trail time). IEEE ballot is minimum ten day; probably ask for at least 14 days to allow for travel plans. Expected timetable for D5.1 is 2008-05-15. Aim for RevCom submittal 2008-08-15, with IEEE Standard approved in September.

11. TC1 PAR should probably start after the document is approved.

    No face to face meetings planned for the immediate future.

    Next teleconference 2008-03-27 to discuss interp 122 and hexadecimal escapes, and possibly 2008-03-13 to discuss char16_t and char32_t paper.

12. AOB
    1. All members of the Austin Group wish to thank The Open Group for hosting this meeting, and generally taking care of our needs so dutifully.

    The meeting adjourned at 11.20am PST, Wednesday March 5, 2008.