Re: Minutes of the 9th March 2017 Teleconference
Date:Fri, 10 Mar 2017 12:12:57 + From:Andrew JoseyMessage-ID: <90d2058a-5014-4f9c-95f9-51dba40af...@opengroup.org> | Bug 0001038: Specification relies upon unspecified behaviour OPEN | http://austingroupbugs.net/bug_view_page.php?bug_id=1038 | | The whole call was spent on this item. Actually, it seems that more time might have been spent on the issue of what the default PS1 should be (and what it means if it becomes unset) none of which is really that item. | We have a proposed change to the "$ " problem in the etherpad, Looks fine to me. I might have just added a new bullet to the list, but either way is fine. Some comments about text I see on the etherpad... Mark: ] Suggested change: ]At page 2352 line 74952 change: ]The default value shall be "$ ". ]to: ]The default value shall be '$ '. That actually changes nothing - the quotes in that are a typographical convention in the doc to show that the default value of PS1 is the two character sequence dollar space, they're not intended to be sh syntax quotes. The 2 character sequence is still subject to parameter expansion, whatever format the value is represented in the text as. It might make a difference of the text said something more like The default shall be as if the command PS1="$ " were executed. but it doesn't (and there is no real point rewording it that way.) Mark: ] How about: The default value of PS1 should be defined in a way that ] results in '$ ' to be displayed? That's not good enough, I need to know how to set it explicitly, not just that the shell can implement some internal magic that makes this happen. What's more I need to safely be able to do PS1="my-string${PS1}" and know that's safe - that nothing that happens to be in "my-string" can affect the (embedded) expansion of PS1 - for that I need to know what the initial value of PS1 will be. Mark: ] Each time an interactive shell is ready to read a command, the value ] of this variable shall be subjected to parameter expansion and written ] to standard error. ] ] I read this as the intent being: if it is unset, so can't be evaluated, ... That would be bizarre - it is perfectly OK (and done all the time) to evaluate unset variables - the result is just a null string, nothing special is needed to handle this case. It really is no different than unsetting PATH - the shell has to evaluate $PATH to perform command lookup for all commands not containing a / (and not being special builtins...) Yet no-one believes that "unset PATH" means that PATH gets magically reset to its default value when it is needed. As far as I'm aware, only IFS is somewhat special in this regard, and only sometimes (and when that happens it is made very clear.) I agree that some text could be added to explain the startup sequence better, where the initial values of the shell variables (of which PS1 is just one) and how they come about, is better defined. Probably this belongs in 2.5.3, possibly by changing line 74898 from The following variables shall affect the execution of the shell: to something more like The variables listed below shall affect the execution of the shell. Where an initial (or default) value is specified, that value shall be assigned to the variable during shell initialization, before the environment is examined, which may then supply different values. Subsequently (including while processing the script given by ENV, when such processing occurs) these variables can be updated, or unset, as with any other variable. Except as noted no other special processing takes place if that occurs (default values are not re-instated into unset variables.) It might also be worth noting with IFS that it is unspecified whether its value can be obtained from the environment (some shells do not allow that.) And most importantly (and back on issue 1038) ... Mark: ] Does anyone disagree with below, as our operating presumption: ]Re: 3606 ]Parameter and other Expansions happen on tokens, not text. I do. Or I do unless by this you mean that a double-quoted string, or a here document text, can be called a word token (which is what it would be - at least strings - when lexically scanned before parsing) but if that is what you mean then it doesn't really help anything. On the other hand, if you mean that strings, here docs, etc, should be tokenised (according to 2.3) as a first step as part of performing the various expansions permitted (or required) then that's simply wrong. "tokens" as a notion only make sense as input to the parser - they are the input symbols to the grammar, they have no meaning, or use, in other contexts. The test in processing step 5 of section 2.3 that relates to $ processing, is so that tokens will be correctly recognised, that is so that x=$(echo
Event [Austin Group teleconference +1-888-426-6840 PIN: 2115756] 2017-03-16 at 11:00 America/New_York created
BEGIN:VCALENDAR VERSION:2.0 PRODID:-//opengroup.org//NONSGML kigkonsult.se iCalcreator 2.18// CALSCALE:GREGORIAN METHOD:REQUEST BEGIN:VTIMEZONE TZID:America/New_York X-LIC-LOCATION:America/New_York BEGIN:STANDARD DTSTART:20171105T02 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 TZNAME:EST END:STANDARD BEGIN:DAYLIGHT DTSTART:20170312T02 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RDATE:20180311T02 TZNAME:EDT END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT UID:58c298bda2...@opengroup.org DTSTAMP:20170310T121453Z ATTENDEE;ROLE=CHAIR:MAILTO:a.jo...@opengroup.org CREATED:20170310T00Z DESCRIPTION:Web/Project: Single UNIX Specification\nTitle: Austin Group tel econference +1-888-426-6840 PIN: 2115756\nDate/Time: 16-Mar-2017 at 11:00 America/New_York\nDuration: 1.50 hours\nURL: https://collaboration.opengro up.org/platform/single_unix_specification/events.php\n\n* All calls are an chored on US time **\n\nTopic: Austin Group teleconference\n-- -\nAudio conference information\n- --\nCall-in toll free number (US/Canada): +1-888-426-6840\nParticipant PIN: 2115756.\n\nAll Aust in Group participants are most welcome to join the call.\nThe call will la st for 1.5 hours .\nThis call is handling defect report processing.\n\nirc ://irc.freenode.org/austingroupbugs\nAn etherpad is usually up for a meeti ng\, with a URL using the date format as below:\n\nhttp://po...@posix.rhan sen.org:9001/p/201x-mm-dd\npassword=2115756#\nAdditional Call-in numbers: \nGermany Caller Paid0-69-2443-2290\nGermany Toll-Free 0800-000-1018\nUnited Kingdom Caller Paid 0-20-30596451\nUnited Kingdom Toll-Free 0800-368-0638\nUSA Caller Paid 2 15-861-6239\nUSAToll-Free 888-426-68 40\nDenmark Caller Paid32711870\nDenmark Toll-Free 80-717000\nCzech Republic Call er Paid 2-39016353\nCzech Republic Toll-Free 800-143-484\nCall-in numbers for other countries are available on re quest\n\nBug reports are available at:\nhttp://www.austingroupbugs.net\n DTSTART;TZID=America/New_York:20170316T11 DURATION:PT1H30M0S LAST-MODIFIED:20170310T071453Z ORGANIZER;CN=Single UNIX Specification:MAILTO:do-not-re...@opengroup.org SEQUENCE:0 STATUS:CONFIRMED SUMMARY:Austin Group teleconference +1-888-426-6840 PIN: 2115756 TRANSP:OPAQUE URL:https://collaboration.opengroup.org/platform/single_unix_specification/ events.php X-VISIBILITY:40 X-JOINBEFORE:5 X-CATEGORY:Teleconference X-PLATO-SITE:Single UNIX Specification X-PLATO-SITEID:136 END:VEVENT END:VCALENDAR meeting.ics Description: application/ics
Minutes of the 9th March 2017 Teleconference
All Enclosed are the minutes from this weeks call. The next call is on the 16th. The US clocks will have changed, so it will be an hour earlier in Europe. regards Andrew Minutes of the 9th March 2017 Teleconference Austin-805 Page 1 of 1 Submitted by Andrew Josey, The Open Group. 10th March 2017 Attendees: Don Cragun, IEEE PASC OR Mark Ziegast, SHware Systems Dev. Nick Stoughton, USENIX, ISO/IEC JTC 1/SC 22 OR Joerg Schilling, FOKUS Fraunhofer Geoff Clare, The Open Group David Clissold, IBM Martin Rehak, Oracle, The Open Group OR Eric Blake, Red Hat Apologies Richard Hansen, Google Andrew Josey, The Open Group * General news The draft PAR has not yet been submitted. Don and others had provided feedback by email. * Outstanding actions ( Please note that this section has been flushed to shorten the minutes - to locate the previous set of outstanding actions, look to the minutes from 28 Jan 2016) Bug 249: Add standard support for $'...' in shell Reopened http://austingroupbugs.net/bug_view_page.php?bug_id=249 We will return to bug 249 on a future call. Bug 953: Alias expansion is under-specified Was Accepted as Marked http://austingroupbugs.net/view.php?id=953 Richard has an action to propose new wording to discuss in a future telecon. * Current Business Bug 0001038: Specification relies upon unspecified behaviour OPEN http://austingroupbugs.net/bug_view_page.php?bug_id=1038 The whole call was spent on this item. We have a proposed change to the "$ " problem in the etherpad, and Don took an action to propose additional changes to the PS1 description for Issue 8 (to be applied after the changes in bug 1006). To be continued next week Next Steps -- The next call is on March 16th, 2017 (a Thursday) Calls are anchored on US time. (8am Pacific) Note that due to US daylight savings time, it will be one hour earlier in Europe. Apologies in advance from Eric Blake David Clissold This call will be for the regular 90 minutes. http://austingroupbugs.net An IRC channel will be available for the meeting irc://irc.freenode.net/austingroupbugs An etherpad is usually up for the meeting, with a URL using the date format as below: http://po...@posix.rhansen.org:9001/p/201x-mm-dd password=2115756# Andrew JoseyThe Open Group Austin Group Chair Email: a.jo...@opengroup.org Apex Plaza, Forbury Road,Reading,Berks.RG1 1AX,England Tel:+44 118 9023044