[Libreoffice-qa] minutes of ESC call ...
michael.meeks at suse.com
Thu Aug 8 08:42:19 PDT 2013
+ Hangout: Michael M, Thorsten, Kohei, Joel, Kendy, Andras
+ Phone: Lionel, Cloph, David, Norbert, Caolan
* Completed Action Items:
+ review date / time patches before Monday
[ in 4.1.1rc1 ]
+ improve configure to avoid versioning mistakes (Cloph)
[ punted to longer term ]
+ take a look at the gerrit migration (Norbert)
[ pending Florian's return from Brazil - but database name mismatch found & fixed ]
+ Poke andras wrt. translation updates / README (Cloph)
[ done, covered, translators double-checked strings ]
AI: create i18n tool to ensure style issues (also unique fn names) don't recur fdo#67786 (Andras)
+ Impress autoplay feature -> deferred to 4.1.2 (Thorsten)
[ merge done ]
* Pending Action Items:
+ come up with proposal for what to do about .debs (Rene, Bjoern, Petr)
[ still waiting, issue perhaps not as bad as thought ]
+ rotate easy hacks on https://wiki.documentfoundation.org/Development/Easy_Hacks (Joel)
- first demo up from Robinson:
- looking forward to having it integrated.
+ polish & push chrome performance bits to a feature branch (Pierre-Eric)
ongoing; heard there is a standalone viewer too (great)
- ongoing, waiting for migration to new vm
+ dig out the URL for GSOC travel funding (Cedric)
[ on vacation ]
+ upgrade 3.6 users to 4.0.5 once that is out (Thorsten/Kendy)
+ add 'download updates from inside LibO' to GSoC wiki page (Astron)
[ pending ]
* Release Engineering update (Cloph)
+ 4.0.5 rc1 status
+ already in public testing, announced, on website.
+ Linux build testing appreciated
+ testing appreciated - CentOS print header touch etc.
+ what / where is gtkprint used ?
+ 4.0.5 rc2 deadline next monday (12th)
+ 4.1.1 rc1
+ almost done with building, up-load/push coming, announce
due this evening
+ 4.1.1 rc2 deadline week after monday (19th)
* each patch should be committed with the name of it's author (Michael)
+ important to get the flavour of the individuals to interact with
+ can use signed-off to do that.
AI: + send details to Tsahi (Michael)
* QA update (Joel)
+ Contest result blog
+ prizes being ordered: 5-6 different countries, no problem
+ MAB handling:
+ by next major release, merge into single list,
re-visit them as a policy
+ saves a lot of time-consuming work, interacting on the issues.
+ thanks to Tommy for his great work triaging / updating existing.
+ North America QA Pub Chat
+ First ever North America QA call - just US present though, so
perhaps US chat is more appropriate
+ Regressions in 4.2
+ Already been mentioned but QA has noticed this - a few nasty ones
that have already been added to MAB
[ could it be we suddenly have a dbgutil build ? ]
+ A couple in 4.1 as well that are quite nasty (fdo#67665,
+ problem of not two downloads on the page (Michael)
+ still concern that download page should show only release (Astron)
+ quality - it's poor user-experience to have unexpected bugs (Joel)
+ and work too closing them.
+ Mozilla have optimised downloads, etc. (Astron)
+ Not completely against telling users of early RC's / Betas on
the download page (Astron)
+ mostly concerned wrt. the lack of hierarchy, and lack of
labelling: unclear blue vs. green icons etc.
+ yellow bar: help us test the latest version ?
+ user difficulty with two stable branches anyway (Lionel)
+ users willing to help, but don't read PR etc. (Joel)
+ option within the product: 'any updates' -> test latest builds ?
+ Regressions (philosophical, what is it?)
+ confusing for users & QA, iterations around regression/not/etc.
+ concern around twitter comments
+ if not sure - don't add the keyword (Kohei)
+ it decreases the usefulness.
+ no real confusion from QA,
+ please take the people working on the features seriously (Kohei)
+ if not 100% agreed on that it's a regression
+ some users put it in to get dev's attention.
+ BSA automatically adds it if users say it worked in a previous
version, can user (Joel)
+ some - multiple users said; we were used to it => regression
+ others - developers say that's a feature.
+ not a huge gap between QA and dev (Lionel)
+ sometimes - we decide to keep a regression, we're going to keep it
+ seems fine to me (Joel)
+ close as wontfix
+ binfilter - eg. (Lionel)
+ need a solution - wrt. regression vs. something else (Joel)
+ QA - focus is on user-experience: seems like a clear regression
if users think it is.
+ concern wrt. flow of regression (?)
+ could we move regressions -> enhancement ?
+ for purposeful regressions / ignored until a big patch to fix it comes in.
+ filter these out of the stats (Michael)
+ concerned wrt. limitations of bugzilla design (Kohei)
+ easily to list 'real regressions' - would be awesome - portal needed (Michael)
+ Warnings (Joel)
+ If we know something in particular that a user did is broken or
limited in the extension they are saving, is it possible to give
explicit warnings about that particular thing when they try to save?
+ yes - but a big chunk of development work (Michael)
+ patches welcome; need to aggregate issues though.
+ encouraging initial USA QA call (Joel)
+ five people from US & two from Europe.
+203 -145 (+58 overall)
many thanks to the top five bug squashers:
Cor Nouws 12
Lionel Elie Mamane 10
+ Prioritization of Bugs (Joel)
+ Currently 150 bugs or so marked as Blocker or Critical
+ During QA Call Friday going to request we comb through these and
add to MAB as we see fit - might result in spike on MAB list
+ Prioritization is getting substantially better - perhaps time to
consider using these fields a bit more?
+ don't use priorities at all (Kohei)
+ users randomly change it.
+ understand limitations of bugzilla (Joel)
+ the MAB list is/was a way to overcome these limitations (Michael)
+ difference dev vs. QA processes (Thorsten)
+ but whatever the QA team want to do is of course fine around priority
+ like more detailed choices than most annoying vs. not (Lionel)
+ if users abuse the severity field - can, warn, take measures etc.
+ like wiki vandalism - social proceedures
+ not necessarily a one-size fits all (Norbert)
+ who intervenes in the bug history etc.
+ would be nice to lock a priority / severity once a bugzilla user
has changed it (Michael)
+ in general not a huge problem: most users don't tweak severity (Joel)
+ really prefer to get a summary of bug reports (Kohei)
+ avoiding misc. user interaction etc.
+ if QA continues to grow & we can prioritise them correctly (Joel)
+ can get the priority & severity accurate & obsolete MAB's etc.
+ everyone wants to get UX best, prioritise bugs right, and not annoy developers
+ several Samba issues being reported again
* Suggest make sidebar non-experimental on 4.2 branch (Caolan)
+ should make an effort here; and make it non-experimental
+ what needs doing first ? (Michael)
+ if non-experimental anyway, try to use it ? (Astron)
+ could switch over to new navigator / style window
+ leave the properties dock disabled by default.
+ happy to bin task-pane in favour of sidebar for impress
AI: + switch to pure sidebar mode in master - but not enabled by default (Caolan)
+ except in impress.
+ get feedback on that (Astron)
+ shouldn't duplicate the functionality by default (Astron)
+ should we keep toolbars, and/or side-pane
AI: + come up with a proposal on what to do (Astron)
+ value in keeping that as an option for a while (Thorsten)
+ should try to make sidebar
+ can do more crazy stuff when sidebar is enabled if we retain
the old code (Michael)
+ instead remove the floating versions of gallery / styles / nav (Astron)
+ properties pane, duplicates toolbar eg. ?
+ we need a well thought-through design - cf. AI.
+ reminder - need feedback for the student (Kendy)
+ lots of easy things to fix - appreciated.
* UX input (Astron)
* 4.2 MAB jump (Caolan)
+ programming asserts showing up as crashes, but are not really regressions
asserts normally compiled out for corner cases
+ did a tinderbox get switched to a dbgutil build ?
+ if running a debug build - can we make a regression ? (Joel)
+ need to compare like with like - a dbgutil vs. a dbgutil or not
+ performance regression bits from that - noticeable (Michael)
+ crashes - should get full debugging info / stack trace (Caolan)
+ with that crash info we can do a great job bug fixing.
* Open 4.2 MAB
+ 7/7 2/2 2/2 1/1
* Open 4.1 MAB
+ 13/77 8/68 11/64 3/55 7/52 7/48 2/42 4/39 5/37 7/32 3/26 3/23 4/17
17% 12% 17% 5% 14% 15%
* Open 4.0 MAB
+ 32/157 14/138 13/137 12/137 12/135 12/134 12/133 12/132 16/132 14/130 16/129
20% 10% 9% 9% 9% 9% 9% 9% 12% 11% 13%
* Open 3.6 MAB
+ 38/228 60/245 62/244 60/243 62/243 60/241 57/238 63/237 65/236 68/235 67/232
16% 25% 25% 25% 25% 25% 24% 27% 28% 29% 30%
* Bibisected bugs open: whiteboard 'bibsected'
+ 40/134 33/128 33/125 34/124 34/123 36/122 38/119 37/114 39/114
* all bugs tagged with 'regression'
+ 324(+23) bugs open of 1849(+48) total
* ~Component count net *
Writer - 98 (+4)
Spreadsheet - 38 (+8)
Libreoffice - 36 (+4)
Crashes - 25 (+1)
Database - 21 (-2)
Drawing - 15 (+1)
Borders - 13 (+0)
Migration - 4 (+0)
BASIC - 3 (+1)
+ Migration: https://bugs.freedesktop.org/showdependencytree.cgi?id=43489&hide_resolved=1
michael.meeks at suse.com <><, Pseudo Engineer, itinerant idiot
More information about the Libreoffice-qa