[Libreoffice-qa] minutes of ESC call ...
michael.meeks at collabora.com
Thu Dec 11 08:07:33 PST 2014
+ Kendy, Michael M, Muthu, Andras, Noel, Robinson, Kohei, Cloph,
Miklos, Markus, Stephan, David, Jan-Marek, floeff, Eike, Bjoern,
Caolan, Michael S, Adam
* Completed Action Items:
+ poke wrt. FOSDEM / BetaCoworking - ask him (Kendy)
[ Cor poked, seems they're happy to have us for 2x days,
confirmed - and hoping to have some board rooms ]
+ get stats on bugzilla load / HW strength from Tollef (Robinson)
[ queued up - HW is 2x machines, 2x VM's, each 8GB of RAM,
front-end is couple of GB of RAM, but DB backend is 8GB,
12 virtual CPUs to each ]
+ reach out to Tollef to organise bz migration date/time (Robinson)
[ tentative date/time: Sat Jan 24th, 15:00 UTC - any problems ?
+ no mumering - go for it ].
+ FOSDEM paper submission ... [!] ... deadline tomorrow (Several)
submitted: Thorsten, Michael x2, Matus (not), Michael S, Miklos.
not yet: Markus, Eike, Bjoern
unknown: Jacobo, Adam ]
+ ask Cedric what's up with libcmis (Miklos)
[ mentioned on IRC: released a new up-stream version ...
needs to be integrated into master & merge the feature-branch ...
so getting there - no blockers, just work ].
AI: + review libcmis branch & merge to master & 4.4 (Miklos)
+ move PDF signing out of 'experimental' in 4.4 (Tor)
+ encourage students to come to FOSDEM; travel funding for talkers (Jmux)
[ sadly exams a week after, so 2x will come but
not all & not to the hack-fest ].
* Pending Action Items:
+ fix BSA login / XML-RPC issues (Cloph)
[ pending, Cloph workingon it ]
+ check Matus' startup performance measurement for layered images.zip (Kendy)
* Bugzilla migration update (Robinson)
+ BugZilla migration => Saturday, January 24, ~1600 UTC (~1700 Germany, Italy, France, ~1000 East Coast)
+ going well
+ please test out your tools vs. that.
+ whining feature
+ sends you a list of bugs by E-mail periodically
+ does anyone use it ?
+ Thorsten had one setup - but not unhappy if it dies.
+ all going smoothly; all code in gerrit.
+ thinking about next steps etc.
* Release Engineering update (Cloph)
+ 4.2.8 release status
+ released as final tomorrow - on ML / TDF blogs
+ a rather quiet release
+ 4.4.0 Beta 2
+ released last week.
+ 4.4.0 RC1 - due to freeze next week - on Monday
+ string & UI freeze date, but not branch-off.
+ single review - from someone other than the committer
to push to RC1 - cf. 4.3 branch etc.
+ Horribly late feature collection:
+ OneDrive / SharePoint bits
+ no objections to get it in for RC1 ?
+ OpenGL/CL JPEG de-compression to texture (jpeg-turbo patch)
[ patch still pending, kendy working on it ].
+ New Templates ...
+ 4.3.5 RC2 status
+ will be tagged later today; 5x patches in gerrit for the branch
+ cloph hunting for reviews later on IRC =)
+ Android / iOS Remote
+ Android on the Google Play
+ iOS - update translations, will be pushed too
Windows Signatures: invalid after certificate expires
(i.e. builds signed before April'14 won't validate)
+ how important it is to have old builds still valid?
+ sounds useful not to have that expiring quickly (Thorsten)
+ we have builds every ~month, so is there merit in the old builds? (Michael)
+ half a year old build at the end of life? (Thorsten)
+ 4.5.x will have a new certificate, but the previous 4.5.x-1 would be invalid (Cloph)
+ needless problems with the users? (Thorsten)
+ $200 does not sound that much for a longer term certificate (Thorsten)
+ let's just do it (Michael)
+ not a 'longer term', actually just does not expire
+ good => do it.
* Crashtest update (Markus)
+ expecting new hardware before Christmas.
+ got some new builds & a few new files, some new leads from Caolan & David
on files for inclusion
+ validation results for ODS are >50% better
+ otherwise mostly constant.
* FOSDEM bits
+ when is the meeting ? (Norbert)
+ currently planned Mon / Tues (board meetings)
* Hackfests (Bjoern)
+ initial scheduling for 2015:
+ FOSDEM, Bruessels, Belgium -- February 2015 contact: Kendy/Cor
+ Las Palmas de Gran Canaria, Spain -- March, 2015 contact: Bjoern
+ setup for GSOC etc.
+ contacted them - and plan to go ahead with that.
+ Cambridge, UK -- May 2015 contact: Michael
* UX Update (Kendy)
+ some nice improvements
+ 2x new people submitting their first patch
+ Mattias Põldaru, V Stuart Foote
+ lots of HIG fixes (Adolfo Jayme Barrientos)
+ switching some dialogs back to re-sizable
+ otherwise getting things straightened out
+ otherwise happy with 4.4
+ though a few smaller improvements there.
+ wanted to run a templates competition
+ can we still include them into 4.4 ?
+ request for late feature
+ should not be translated (Moggi)
* Certification (Stephan/Bjoern/Kendy)
* Jenkins update (Norbert)
* QA (Robinson)
+ UNCONFIRMED under 600
+ quiet as testing out 4.4
+ near duplicate of shambolic image loss bug (Michael)
+ would be nice to stop these hyper-long comments (Markus)
+ 50x lines of comments without new information is not useful
reading that is very wasteful of dev. time.
+ ability to prioritise / hide comments perhaps in new BZ (Robinson)
* Bibisect update (Robinson)
+ was going to push yesterday, but wiki slow
+ it is up-loaded here:
+ will mail the QA list shortly.
* Big drop in regressions: v. positive
+ Markus spent his last 2x weekends fixing calc regressions
* VNC logins on Hackfest image (Bjoern)
+ have awesome images for hackfests right now
+ if devs can work there; then can we get QA there ?
+ can we get VNC logins there ?
+ how hard/trivial is this ?
+ easy to do (Cloph)
+ X2go - available for Win/Mac/Linux
+ could we even have that for the Bug Hunting Session?
+ can Robinson test VM X2go flow ? (Bjoern)
AI: + test this out & get good at setting this up (Robinson)
* pImpl vs. abstract base-classes (ABC) (Bjoern)
+ compiler test of ABC vs pImpl - no perf. difference
+ a matter of taste (Kohei/Bjoern)
+ also well worth splitting out constants / typedefs etc. (Michael)
+ lots of other cases to use.
* Performance (Bjoern)
+ SwNode / BigPtrArray foo ...
+ looked at how many SwNodeIndexes we register
+ ran with the unit test: 40k iterators registered in 1 document.
+ sounds really ... odd (but with debug builds)
+ looked at Kendy's branch too
+ best to solve the indexes thing first
+ indexes of whole array, so can't move anything around
without changing it.
+ move SwNodeIndexes from the whole thing, to the
index they're registered at.
+ is there a profile that shows we spend significant time
in these node indexes (Michael S)
+ or is this a theoretical problem ?
+ valid question (Bjoern)
+ try to have some other data structure & see.
+ surely should profile before optimizing (Michael)
+ seen lots of silliness around, but not this yet.
+ purpose of Kendy's branch is to also move to something
similar to a piece-table approach (Kendy)
+ to get somewhere closer.
+ which test-case is it ? (Miklos)
+ the one you mentioned had a lot of documents, in
real-world docs - wouldn't be surprised.
+ bookmark: 2x indexes, red-lines 2x per red-line
+ other unit-tests had 1k -> 5k range (Bjoern)
+ may be a bug, or perhaps normal (Miklos)
* ODF plugfest update (Michael)
+ ~50 attendees, lots of SMEs & ~10x ODF implementations
+ Google - committed to improved ODF: 1.2
+ tested & latest Google Docs ODF is awesome:
ODS has charts, styles, formats etc.
+ Lots of automated tests written (a lot by Andras)
+ Change tracking progress
+ Next plug-fest expected in June in the Netherlands
* QA stats:
+134 -130 (+4 overall)
many thanks to the top bug squashers:
Caolán McNamara 24
Michael Stahl 10
Jean-Baptiste Faure 9
Eike Rathke 8
Markus Mohrhard 7
Adolfo Jayme 6
Miklos Vajna 5
Julien Nabet 5
* Open 4.5 MAB
* Open 4.4 MAB
+ 7/24 7/22 5/19 4/18 3/17 3/16 3/16 4/16 5/15 4/14 3/12 3/12 3/12 7/12
29% 31% 26% 22% 17%
* Open 4.3 MAB
+ 84/146 58/117 33/90 19/73 17/65 16/64 16/64 15/63 16/63 16/60 17/61 16/60
57% 49% 36% 26% 26% 25% 18% 23% 25% 26% 27% 26%
* Open 4.2 MAB
+ 5/208 38/232 63/255 74/261 83/266 81/263 84/262 82/260 80/260 82/260 86/257
2% 16% 24% 28% 31% 30% 32% 31% 30% 31% 33%
* Bibisected bugs open: whiteboard 'bibsected'
+ 208/526 182/488 190/481 189/475 190/470 173/446 170/441 161/427 153/413
* all bugs tagged with 'regression'
+ 454(-19) bugs open of 3257(+9) total
* ~Component count net *
Writer - 137 (-0)
Spreadsheet - 93 (-10)
Presentation - 30 (+0)
Database - 27 (+0)
Crashes - 26 (-5)
Libreoffice - 25 (-1)
UI - 24 (+0)
Drawing - 21 (-1)
Graphics - 20 (-1)
Borders - 20 (+0)
Chart - 14 (-1)
Filters - 13 (-2)
Print / PDF - 10 (-3)
BASIC - 5 (+0)
Linguistic - 3 (+0)
Installation - 3 (+0)
Extensions - 3 (+0)
framework - 2 (+0)
Formula - 2 (+0)
sdk - 1 (+0)
Impress Remote - 0 (+0)
michael.meeks at collabora.com <><, Pseudo Engineer, itinerant idiot
More information about the Libreoffice-qa