You are here: Foswiki>Development Web>ReleasePlan>ReleaseMeetings>ReleaseMeeting01x02_20150504>ReleaseMeeting01x02_20150615 (16 Jun 2015, GeorgeClark)Edit Attach

Release Meeting: 15 June 2015

Details

  • Date: 15 June 2015 1300Z
  • Location: IRC #foswiki-release
  • Log: (See inline at end of this topic)
  • Participants (Add/Delete)

1. 1.2.0 blocker review

As of now, no new blockers have popped up.
  • 13424 (Confirmed): Malformed UTF-8 character (unexpected end of string) fails for some languages on some perl versions. I downgraded to Normal. Only impacts perl 5.16, and is non disruptive although logs can be noisy.
  • 8966 (Confirmed): Wysiwyg editor breaks standard topics Also downgraded. Some small issues were fixed, other topics get corrupted. But editing in System web is discouraged. Leaving open as a catch-all.
  • 13139 (Closed): Backwards compatibility issues with Foswiki 1.2 default extensions Two extensions are not backwards compatible. TopicUserMappingContrib and PatternSkin. Both use API changes in 1.2. The rest of the 1.2 extensions appear to work fine. One issue found in SubscribePlugin was fixed by Crawford.
For the two incompatible extensions, I'm proposing a simple PREINSTALL exit that will die if install is attempted on Foswiki 1.1. This should avoid anyone breaking their 1.1.x system. 
sub preinstall {
die "Not compatible with Foswiki 1.1.x" if ($Foswiki::VERSION < 1.999);
return 'Installation continuing';
}

2. Proposed Schedule for Alpha / Beta / GA

  • Build 1.2.0 Beta 2 immediately
  • Install 1.2.0 Beta 2 extensions on Foswiki.org for more thorough compatibility testing.
  • Build 1.2.0 RC1 - 29 June 2015 *Install RC1 on Foswiki.org
  • If no new tasks, build 1.2.0 on 4 July.

Next actions

On the automatic unit tests
  • TWikiFuncTests don't fail for GeorgeClark. Can anyone else recreate the failure?
  • LoadedRevTests: Needs someone who understands the test to fix.

For GeorgeClark's tests: I've been running into a lot of inconsistencies.
  • One big issue is that Meta::removeFromStore() is not able to remove directories if they contain utf8 named attachments. It shows up in Fn_SEARCH and a few other tests
  • A lot of it is test ordering. Once a utf-8 attachment gets "stuck", various tests fail intermittently.
  • EngineTests fail if Taint::Runtime is installed 
                   '_content' => 'Insecure dependency in eval while running with -T switch at /var/www/foswiki/distro/core/lib/Foswiki/UI/Test.pm line 13.

Any other test results?

Next meeting - - Monday Jun 29, 2015 1300Z — ReleaseMeeting01x02_20150629

IRC Log 

(08:59:37 AM) CDot  entered the room.
(09:01:33 AM) CDot: Je ist here
(09:01:44 AM) gac410: Hey everyone.    Welcome to the release meeting.  I'm on a tight schedule today, so if we run more than 1/2 hour, someone please take over :)
(09:02:05 AM) CDot: we won't, unless you make it overrun
(09:02:09 AM) CDot: let's get to business
(09:02:34 AM) CDot: lots on holiday, it would appear
(09:03:01 AM) gac410: we have no open blockers of any significance.    I've downgraded a couple to normals.
(09:03:12 AM) gac410: http://foswiki.org/Development/ReleaseMeeting01x02_20150615
(09:04:08 AM) CDot: I reviewed the open Normals today. Don't see anything panic-worthy
(09:04:10 AM) gac410: CDot  I added my suggested "pre-install" exit to that meeting topic.   A couple of lines will prevent someone from breaking a 1.1.x system installing an incompatible 1.2. extension.
(09:04:44 AM) gac410: Anyone else?   New blockers,  Issues with downgrades?     
(09:05:11 AM) gac410: MichaelDaum,  have you had any chance to try out the utf-8 changes?
(09:05:25 AM) MichaelDaum: nope
(09:05:37 AM) MichaelDaum: still in berserker mode wrt projects
(09:06:07 AM) CDot: gac410: do we have a picture of which *core* extensions are not backwards compatible?
(09:06:13 AM) gac410: I think we are ready then to move forward with Beta 2?    
(09:06:29 AM) gac410: Just 2 that I could find.    TopicUserMapping    and PatternSkin
(09:06:42 AM) JulianLevens1  entered the room.
(09:06:42 AM) gac410: Both are using some of the macro extensions / api changes in 1.2
(09:07:28 AM) gac410: You fixed SubscribePlugin,   and a quick check ... everything else appeared functional 
(09:08:01 AM) CDot: ok
(09:08:01 AM) gac410: JulianLevens1:  Any new blockers from you?   
(09:08:48 AM) JulianLevens1: No
(09:09:21 AM) gac410: Okay ...    So Release schedule.  ...  Beta 2 immediately.   (I'm close to ready to build - working on release notes.)
(09:09:35 AM) CDot: sounds good
(09:09:56 AM) ***CDot will review the lows and enhancements today as well, as a final sanity check
(09:10:16 AM) gac410: I'm proposing a 2-week beta,  and if nothing major pops up,  Build 1.2.0 RC1 on 6/29.     
(09:10:52 AM) gac410: And lets shoot for 7/4  release.   :)
(09:11:11 AM) MichaelDaum: excellent. finally. party.
(09:11:15 AM) CDot: sounds good. MichaelDaum will just have to catch up when he can :-(
(09:11:24 AM) MichaelDaum: CDot, such is life
(09:11:51 AM) ***CDot had hoped you would be able to review the unicode. But I think you will be pleasantly surprised.
(09:11:52 AM) gac410: Hurdles    if any checkins after RC1,  we need to go to RC2 instead of release. 
(09:12:11 AM) MichaelDaum: CDot, I will have even more work to do.
(09:12:11 AM) CDot: checkins? including doc?
(09:12:48 AM) CDot: [off] MichaelDaum: paying work? Ah yes, I have a vague memory of that... ;-)
(09:12:51 AM) gac410: hm.    I suppose Doc, and Translations are an exception. 
(09:13:11 AM) MichaelDaum: [off] woops
(09:13:21 AM) CDot: I'd give an exception for minor string changes (spelling etc) as well
(09:14:04 AM) gac410: CDot.   My fix to PlainFile   removeFromStore   encoding issues.   ... seemed to fix several of the failing tests.   Stuff being left behind.  I left that task for your review.
(09:14:29 AM) gac410: never mind.  You closed it  :)
(09:14:34 AM) CDot: I reviewed it this morning, and closed it.
(09:14:48 AM) gac410: yes  thanks.  
(09:14:59 AM) CDot: http://foswiki.org/Tasks/Item12042 is bothering me
(09:15:49 AM) CDot: n.m. i juts read it carefully, and I will be closing it no-action
(09:16:24 AM) gac410: okay
(09:17:28 AM) gac410: I think with the fabulous utf8 work by CDot and i18n testing by Jomo ... we are in great shape!
(09:17:44 AM) CDot: gac410: http://foswiki.org/Tasks/Item11115
(09:18:10 AM) CDot: and http://foswiki.org/Tasks/Item10428
(09:18:39 AM) gac410: I'm hoping that those are fixed by MAJOR tmpfile changes in CGI.   
(09:19:27 AM) CDot: I recall you were looking at that; just a heads-up.
(09:19:51 AM) gac410: yeah   I've had a hard time recreating it.   I've not noticed the issue. 
(09:20:30 AM) CDot: BTW I have Taint::Runtime installed (as always) and the EngineTests pass clean for me.
(09:21:07 AM) CDot: I fixed the PlainFile LoadedrevTests this morning
(09:21:26 AM) CDot: the TWikiFuncTests don't fail for me either.
(09:21:32 AM) gac410: I have no idea why that one test was failing.   I went back several perlbrew versions.   I could not get it to pass, and finally turned off taint checking for just that one
(09:21:45 AM) CDot: odd
(09:22:09 AM) gac410: The Func failures were due to the plainfile cleanup I suspect.    It was only happening with the full suite,  and was unpredictable.
(09:22:34 AM) gac410: But fsfs's  tests were consistently failing and it went away after the removeFromStore fix
(09:22:40 AM) CDot: k
(09:23:35 AM) gac410: thanks for the loaded rev fixes.   I did look at them and decided I didn't trust myself with the store complexity to get it right.
(09:26:14 AM) CDot: sure. The tests were actually testing that the RCS store was "correctly broken"
(09:26:37 AM) CDot: and I ported them to PFS without thinking that PFS is *not* broken (correctly of otherise) ;-)
(09:26:49 AM) CDot: *or otherwise
(09:27:40 AM) gac410: The /tmp directory on f.o has around 700 files.   But looking at the list, they go back to 2009,   so we appear to never clean /tmp and certainly are not accumulating uploads.
(09:29:26 AM) CDot: that's good
(09:29:45 AM) gac410: none from 2015 or 2014.    So I can't see how uploads are being left behind.
(09:29:51 AM) CDot: assume /tmp is where they actually go, and not some odd $CGITEMP dir
(09:31:00 AM) gac410: I don't think that f.o has any special configuration done for that.
(09:31:32 AM) CDot: well, I trust you to have looked in the right places ;-)
(09:34:08 AM) JulianLevens1: gac410: when do you have to leave?
(09:34:41 AM) gac410: anyway ... on that note,  I need to get to dentist.   I'll get 1.2 Beta2 built today.  Please review release notes, upgrade guides,  etc.  I've tried to catch up all the extension change logs last night.
(09:35:04 AM) gac410: I have about 10 minutes before have to be IN my car,  but things to do.  
(09:35:42 AM) JulianLevens1: Bye and good luck
(09:35:48 AM) CDot: have a fun time at the dentist, and remember, they toys in the waiting room are for the little kids. You'll get told off if you play with them.
(09:36:25 AM) ***CDot heads back out to the garden
(09:40:40 AM) JulianLevens1: Did anybody capture the log?
(09:41:35 AM) gac410: I've got them   and I'll leave window open
(09:41:56 AM) gac410: bye.   see yall later.
(09:43:45 AM) MichaelDaum: thanks

BasicForm edit

TopicClassification Select one...
TopicSummary
InterestedParties
Edit  | Attach | Print version | History: r2 < r1 | Backlinks | View wiki text | Edit wiki text | More topic actions
Topic revision: r2 - 16 Jun 2015, GeorgeClark
The copyright of the content on this website is held by the contributing authors, except where stated elsewhere. See Copyright Statement. Creative Commons License    Legal Imprint    Privacy Policy