[lsb-discuss] LONG, but README unofficial LSB conference call minutes for 25 Jun 2014

R P Herrold herrold at owlriver.com
Wed Jun 25 17:56:13 UTC 2014

Hash: SHA1

LSB teleconference

Wednesday, 25 Jun 2014 @11:00 to noon, US ET 

Where:	(605) 715-4920 Access Code: 512468

Gobby:	gobby -j lsb-temp.pmman.net:6522 with passwd: LSB

IRC channels are:
	#lsb-meeting at freenode.net during meetings
	#lsb at freenode.net most other times

in the form : name (organization)  [irc_nick]
Jeff Licquia (LF)  licquia
Robert Schweikert (SuSE)  robjo
Jeff Johnson (Open Mandriva) 
Russ Herrold (Owl River)  orc_emac
Kay Tate (LSB)  ktate
Mats Wichmann (LSB)  mwichmann  (partial - last half)

Mats Wichmann (LSB)  mwichmann  (partial)
Rick Troth (Velocity Software) rickt VMWorkshop

Posted Agenda 

 - GTK 3 changes.

 - Small update to LSB 4.1, and version-independence.

 - Other issues: LSB 5 status, new LSB.

New business:

Meeting opens at 11:00 with:

QM test fix applied to Test infra, and pushed for the Yocto 
tester.  Turns out that this is NOT really version 
independent.  A refactor is needed to make it truly version 
independent.  Russ thanks Jeff L yet again for this as it 
helps out only audible client at the moment

There is never just one roach, and similarly to the QM issue, 
Mats' efforts on gdk/gtk have exposed prior errors in our 
existing database which need to be exterminated as well

Jeff L mentions his desire to be able to delegate LANANA work; 
there is an open bug on this

Jeff L notes he is hte single 'gate' through which all 
certification requests have passed.  There are some privacy 
concerns pre-certification as an applicant works through 
issues, but there seems to be no reason this cannot be 
addressed with limited NDA's when performing LSB certification 
work for non-public certifications --- needs further 
discussion and study, per Russ

IRC nic: lsb_tester / Milan has been slogging through the test 
suite and a Yocto platform

Jeff Johnson's work has yielded a LSB 5 hookup and population 
have yielded'bubble diagram' at Launchpad has drill down 
capabilities -- click a bubble for details and then scroll 
back to the top right on the resultant page

P1 bugs included into blockers for LSB 5 -- count is 14

ID 	Summary
3986  make go and no-go decisions on new libraries for 5.0
3951  uplift FHS reference to 3.0
3931  gtk3 gtk.h/gdk.h not buildable
3754  missing interface definitions
3508  Implement modularization in LSB 5.0
3463  LSB 5.0 SDK Release
3168  LSB 5.0 Application Battery Release
3167  LSB 5.0 Application Checkers Release
3166  LSB 5.0 Distribution Tests Release
3165  LSB 5.0 Specification Release
3164  LSB 5.0 Complete Release
2877  App Cheblo cker fixup bugs for native apps
2199  RFE: Include libsane interfaces in LSB
1391  add async I/O interfaces to LSB 


P1 and P2 in critical path of LSB 5 -- count is 92

ID 	Summary
3999 	 devchk should not test lsb/ headers
3986 	 make go and no-go decisions on new libraries for 5.0
3978 	 should xml2 testsuite move out of desktop?
3958 	 can/should mkheader emit funcptrs separately?
3951 	 uplift FHS reference to 3.0
3949 	 Specification does not outline module composition
3945 	 lsb-build-qt does not build with 5.0
3944 	 inb/outb are not defined in <sys/io.h>
3943 	 gtk3 missing types
3941 	 Gtk 2/3 usages of "struct _Foo" instead of "Foo"
3940 	 LSB5 import lets glib usage of "struct _Foo" creep in 
where should be "Foo"
3935 	 spec .defs files need to handle arch-specific 
3932 	 gtk3: function pointers, enums, structs misdefined
3931 	 gtk3 gtk.h/gdk.h not buildable
3929 	 gdk3 gdk.h/gtk.h enums wrong values
3928 	 devchk build problem gtk2 vs gtk3 (native)
3912 	 LSB RPM format needs to support multiple Provides
3910 	 need updated snapshot of NSPR on spec site for 
changes to 5.0 spec
3909 	 need pkgconfig files for ncurses/ncursesw
3900 	 foomatic-rip is specified in the wrong place
3877 	 LSB specifications that are not ISO stds should not 
imply they are
3875 	 RFE: add ioperm to LSB
3874 	 convert libchk and other checkers to 
3857 	 RFC: spec production: appendix, index?
3838 	 resolve documentation links for 
3825 	 refactor loopback code in core and dist-checker
3818 	 add sendmail's sysexits.h return codes to the LSB
3812 	 Test needed for improper freopen() behavior in recent 
3811 	 uplift c++ tests
3810 	 improve devchk c++ support to help uplift
3794 	 dist-checker complains about missing qt 3 tests with 
5.0 desktop-test
3772 	 update reference to x86_64 ELF ABI
3764 	 RFE: add crypt_r
3760 	 printing-test uses deprecated httpConnect()
3754 	 missing interface definitions
3730 	 ncursesw/curses.h defining TRUE, not defining FALSE
3727 	 should lsbcc use -nostdinc or not?
3725 	 xmlParserInputRead deprecated upstream
3715 	 mkheader inserts #undef if *any* macro is found by 
the same name, even if in unrelated header
3706 	 RFE: spec cleanup: make Core/Base follow conventions
3701 	 Elfchk handling of dwarf sections requires 
3697 	 need mechanism for lsbcc to include proper headers 
for uplifted libs
3678 	 uplift and/or deprecate gtk2
3669 	 /tset/initd/misc/install/general-tc 2,4 FAIL - Ubuntu
3668 	 tcl clock34.{6,12..18} fail on Rawhide
3659 	 RFE: uplift png to version 1.5
3644 	 devchk native build problems
3595 	 Snapshot SDK fails to build lsb-build-qt3
3548 	 Add GTK+ 3 for LSB 5.0
3512 	 rewrite lsb_release sample to support new 
modularization scheme
3508 	 Implement modularization in LSB 5.0
3496 	 tcl tests segfault on filename test, never complete
3486 	 Rename architecture IA32 to x86
3485 	 Artifacts exist showing incomplete architecture 
3481 	 Spec references obsoleted RFCs for RPC/XDR
3469 	 RFE: uplift LSB to POSIX 2008
3468 	 RFE: add netlink.h to the standard
3465 	 missing verr/vwarn functions
3464 	 error() function is an orphan
3463 	 LSB 5.0 SDK Release
3459 	 Remove deprecated Perl modules from LSB, or ship 
compatability package
3264 	 RFE: add pam_get_user, pam_set_data
3259 	 Uplift appbat package versions for 5.0
3168 	 LSB 5.0 Application Battery Release
3167 	 LSB 5.0 Application Checkers Release
3166 	 LSB 5.0 Distribution Tests Release
3165 	 LSB 5.0 Specification Release
3164 	 LSB 5.0 Complete Release
3100 	 uplift C++ headers to more recent upstream
3051 	 irowbytes has vanished from png.h
2885 	 Uplift ghostscript base version to 8.63 for OPVP v1.0
2877 	 App Checker fixup bugs for native apps
2831 	 libutil section has no data definitions
2770 	 Should DEV_BSIZE be defined?
2593 	 provide pkg-config files to make sure gio is not 
2582 	 lsb_release/Distributor ID not well specified
2431 	 add proper c++ build support
2413 	 SDK may not be version-independent for C++
2407 	 Consider dropping sendmail as a required core 
2389 	 POSIX UPU command issues
2199 	 RFE: Include libsane interfaces in LSB
2189 	 Add InterfaceAttribute entries for some previleged 
1819 	 Update references to the DWARF standard
1800 	 RFE: appchk should issue a warning when DT_TEXTREL is 
1761 	 add libncursesw to the LSB
1716 	 lsbcc should not automatically add -lpthread
1391 	 add async I/O interfaces to LSB
1196 	 RPMTAG_GROUP unclear how to use
1033 	 VTT array contents not described
619 	 release lsbdynchk
154 	 name conflicts in /opt
147 	 package naming rules still confusing/restrictive 


General discussion on workflow going forward:  Russ suggests 
the modest addition of a connector from BZR to git, so that we 
still have access to the history in BZR, but we move to using 
git.  Authoritative commits would initially be in BZR and be 
picked up by that new process and moved to git

Then dial in some test we are presently doing with the 
buildbots, and replicate Robert's test harness from SUSE to 
get familiar with it

JBJ notes: really there are Five products: Spec, App check, 
App battery, SDK, and the Test which are tracked.  Spec is a 
documentation task and has the most open bogs, and have the 
most policy decisions to be made. App check, App battery, and 
the Tests rot, and have incorrect assumptions exposed, and so 
need less maintenance

Russ asks Mats, based on his experience, for a WAG as to hours 
of work to issue a LSB 5 as matters now stand.  As that 'wild 
ass guess', Mats estimates there is 600 hours of work to get 
LSB 5 out the door.  Functionally Jeff is the only compensated 
staff and yields about 8 FTE hours a week.  By that metric, we 
have about a year and a half until 5 releases.  A 2016 release 
is just not credible.  The broader community needs to step up, 
or it never happens

Robert synthesizes and recaps.  The 'Ship LSB 5 first 
decision' from the 2014 Face to Face was based on incomplete 
information, and needs to be re-considered as we have firmed 
up 'burndown rate'.  He notes that re-jiggering infrastructure 
is not in the critical path of deciding: ship an LSB 5 or not.  
Jeff L notes that perhaps we need to cut more.  Russ points 
out that we have gutted the LSB 5 visioning wiki page to the 
bone already as to minimum viable featureset


The question of 'LSB.next' was raised, but making decisions 
there seems premature in light of Robert's observations.  The 
conclusions by the participants at the Collab Summit 2014: 

do not have explicit or formal buy-in beyond those present 

Jeff L will see about cutting and checking that P1 for 
children of blockers are correct

Call ends at local Noon

Weekly Bug triage:
The bug tracker is at: 
(A former bug tracker was used, but is not longer active at:
for people doing research.  Please do not use it for new 

Bug triage occurs weekly: irc-only #lsb-meeting at 
*** FRIDAY* ** at 10:00 Eastern

Bugzilla (lsbbugs) stats as of 25 Jun 2014:  
618 open, 614 open last week 
46 assigned, 5 new, 1 reopened, 1 needinfo, 4 pleasetest

New last 90 days: 46, last 30: 16; closed last 90: 29, last 
30: 11
 **** these close rates also foretell the sad tale of a LSB 5 
release **** 

Changed this week: 16, last week: 40
Bugs that are assigned but have no milestone: 245
Bugs (non infrastructure) that are assigned but have no 
milestone: 154
Rollup bugs: 26

Bug (bugs) stats:
62 open FHS
62 assigned, 14 new, 0 reopened, 0 needinfo, 0 pleasetest

last week's bug triage minutes:
Minutes (text): 

How to get involved:
NEW: LaunchPad visualization

 - see the last minutes of the month for a full list of ways 
to participate.    These may be found in the mailing list 
archive at:

 - get started ... there is a workflow to 'learn' and joining 
the IRC channel (details below) is easy and well suited to 
incrementally going

 - join and lurk in the IRC channels (free)
   on irc.freenode.net, channels #lsb and #lsb-meeting

 - join and participate in the mailing list

 - 'star', 'watch' and join with merge proposals our Github 
presence (free)

 - register for a bugzilla account (free)

 - watch, review, test and comment on the commits crossing the 
RO mailing list: 

 - dial into the weekly call and participate

 - join in the edit of the weekly call minutes in the gobby
   Free gobby clients available at: 

 - join bug triage and resolution

 - file, triage,and fix bugs ... this entails doing BZR CO, 
git CO, and setting up a MySQL dataase, and so is complex, but 
documented.  Improvements are ALWAYS in order, so ...

 - document matter in the wiki
   (requires LF web account to edit, free)

 - set up and run a puppet build master and build slave 
continuous integration tester

 - Read about editting the specification

Next meeting:
2 Jul 2014 11:00 to noon, US ET 

gobby unofficial minutes are open for edit for at least 15 min 
after each call, and then issued into the mailing list

This document: LSB_minutes_20140625.txt 

- -- 
- --
 .-- -... ---.. ... -.- -.--
Copyright (C) 2014 R P Herrold
      herrold at owlriver.com
   My words are not deathless prose,
      but they are mine.

Version: GnuPG v2.0.14 (GNU/Linux)


More information about the lsb-discuss mailing list