minutes of ESC call ...

* Present:
    + Kendy, Xisco, Heiko, Stephan, Christian, Michael M,
      Sophie, Michael S, Caolán, Mike Saunders, Olivier, Thorsten,
      Miklos, Eike, Shinnok

* Completed Action Items:
    + ship the horror win / VC runtime DLLs with the installer (Cloph)
        [ patch appears to work, but we already have them in the installer
          if you do an admin-install they are already installed.
          we ship them twice – once as LibreOffice, but once as a run-time module
          not entirely elegant.
          for 5.4 – merge it as-is, for 6.0 do something clever.
http://dev-builds.libreoffice.org/daily/libreoffice-5-4/Win-x86@62-TDF/tdf_108580/
        ]
    + research the best way to build HTML help / documentation (Olivier)
        [ working actively on this (Olivier)
              expect to have a solution quite fast.
          WIP patch pushed to gerrit for core/ (Kendy)
              checks if HTML help is installed or not,
              will add opening of pages later
              https://gerrit.libreoffice.org/#/c/43963/
        ]

* Pending Action Items:
    + run dev certification script (Kendy/Thorsten)
        [ have one name without running the script ]
    + provide / check web-page on dev certification process (Kendy)
        [ poke Italo to find out what to check (Kendy) ]
    + find data for x86 / Linux downloads with view to removing for 6.0 (Cloph)
        [ have the data, not broken down yet: termbin.com/2jc2 ]
    + fwd build details on Mac / CVE build / test failure to the list (Shinnok)
        [ Bartoz removed obsolete, non LODE details from the wiki
          removed quick-setup and mac-ports setup, recommends to
          --disable-cve-tests on the Mac.
          It works for me; what’s the issue (Stephan)
          re-send the mail to the list / Stephan (Shinnok) ]

* Release Engineering update (Christian)
    + 5.4.3 RC2
      https://gerrit.libreoffice.org/#/q/status:open+branch:libreoffice-5-4-3
      10 patches waiting
        + plan to tag this evening
        + please help review things.
    + 5.3.7 RC2 released as final today
      Resolves: tdf#111428 implement CellStoreEvent::swap() in ScColumn::SwapCol()
      possible ugly bug – 5.3.8 ?
        + Eike found a nasty bug, is it enough to do a 5.3.8
        + can lead to non-re-calculated values; would be nice to fix in 5.3
        + do we need a new release ? (Michael)
            + fixed in released 5.4.3 bits
            + merge it to the -5-3 branch in case of new release ?
            + quite a specific scenario (Eike)
               + would like to get it out if we can.
            + fine to encourage people to use 5.4 (Cloph)
               + move people to 5.4 a bit earlier too perhaps.
    + 6.0 Alpha1
        + 6.0 feature freeze: week 47 (Nov 20th)
        + late features
            + HTML help pieces (Kendy)
           + GPG / Windows bits (Thorsten)
           + Calc threading review with Eike (Michael)
           + KDE5 – merge but perhaps not ready for 6.0 (Thorsten)
               + can’t build on baseline anyway (Michael S)
           + bundling GPG-4-Win – a 25Mb package for GPG
             encrypt/sign for Win? (Thorsten)
               + pointer for an extra download ?
               + License ? (Michael)
                   + mix of GPL/LGPL (Thorsten)
               + want access to encryption reasonably easily
                   + maintenance / security updates etc. an issue.
               + 10% increase – for perhaps minority benefit.
AI: + mail the list (Thorsten)
           + Calc column span update from Noel (Michael)
        + Updating service – expected to miss 6.0 (Cloph)
    + Android
    + Online

* Documentation (Olivier)
    + Broke large script for new help into smaller pieces
        + aim to make easy for build and packaging integration
        + looking into how to hack a makefile
            + https://goo.gl/A8dGVf
    + resuming book editing with doc community for 6.0
        + cover of the book – is under discussion.
    + looking at XML editor for help + workflow (Mike S)
        + a bunch of technical things to investigate here.

* UX Update (Heiko)
    + Bugzilla (topicUI) statistics
        252(252) (topicUI) bugs open, 395(395) (needsUXEval) needs to be evaluated by the UXteam
    + Updates:
        BZ changes 1 week 1 month 3 months 12 months
             added 2(-7) 20(-3) 43(-1) 115(-2)
         commented 45(-32) 230(17) 565(17) 2368(-14)
           removed 0(0) 0(0) 6(0) 21(-1)
          resolved 6(-1) 23(4) 55(-6) 238(-20)
    + top 10 contributors:
          Philips, Yousuf made 134 changes in 1 month, and 792 changes in 1 year
          Tietze, Heiko made 113 changes in 1 month, and 1203 changes in 1 year
          Thomas Lendo made 57 changes in 1 month, and 475 changes in 1 year
          Cor Nouws made 40 changes in 1 month, and 201 changes in 1 year
          Foote, V Stuart made 35 changes in 1 month, and 409 changes in 1 year
          Faulí Tarazona, Xisco made 24 changes in 1 month, and 351 changes in 1 y
          Dieter Praas made 14 changes in 1 month, and 65 changes in 1 year
          Henschel, Regina made 11 changes in 1 month, and 121 changes in 1 year
          Timur made 11 changes in 1 month, and 75 changes in 1 year
          Monastirsky, Maxim made 9 changes in 1 month, and 33 changes in 1 year

    + new table styles tdf#101349
      (works under Linux but not at Windows – build issue?)
        + replacement of the binary file → XML ideal.
        + TB42 builds – apparently no styles (?)
           + choose default table size carefully ? (Miklos)
           + one item that is different on 32bit / 64bit platforms (Michael S)
              + we have two of these files.
        + not a big deal to setup styles locally,
          switch local files → shared (Heiko)
AI: + build new binary style file for 32bits (Heiko)
           + tested table styles on Linux (Xisco)
              + couldn’t see it – but Heiko does.
    + new templates tdf#103317 – 16x9 the new default here.
    + Discussion about 115% line spacing for Default
       + request to move to this – under discussion.

* clang-format-ness (Miklos)
    + Came up with a prototype at https://gerrit.libreoffice.org/#/c/43736/
    + got more input since last week: gerrit comments, ML posts
       + some are happy with git pre-commit approach
       + only way to enforce something
       + some people want to push it into gerrit/jenkins not locally
       + plan from DavidO on how this could be done in gerrit
       + style enforcement in different steps from Stephan
           + first - git-hooks, when clang-format is installed
           + second – jenkins-like auto-verification of style.
           + third – re-base button auto-fixes style in gerrit reviewer
    + question is who is interested in doing the work.
       + Miklos happy to do git-hook hacking
       + if people want features they are more than welcome to impl.
    + propose merge step#1 without clang-format enforced (Michael)
       + don’t like the auto-re-format of others’ commits (Stephan)
          + if want to have enforced – make sure all commits is formatted right
            or we get these issues.
       + we have a similar issue with warning fixings (Kendy)
          + starting with limited files.
       + concern that people wouldn’t install clang-format (Stephan)
          + if we have this merged with other stuff.
    + to what degree do we provision binaries ? (Michael)
       + builds on windows / linux / Mac (Miklos)
          + could add this in the hook ?
       + what do we do wrt. gerrit edits ? (Stephan)
          + very minority use-case (Michael)
          + one single place we check stuff going into Master (Thorsten)
             + gerrit is optional (Miklos)
                + you don’t have to go through it.
       + not eager to make the hooks more complex (Thorsten)
          + appreciate a move into the right direction, but …
       + getting somewhere we didn’t want to go (Kendy)
          + Rome: start small – with some few files.
          + enable just for a few files Miklos is responsible for
              + only for new files.
          + understanding from Rome (Stephan)
              + these test files don’t get a cleanup commit – where
                clang-format doesn’t need a follow-up commit.
          + very, very unlikely that 0.1% of web-edits convolve with this (Michael)
              + with the 1% of files that we cover with this..
       + getting this in optionally, and config in (Thorsten)
          + is quite valuable.
          + cases – where people push without git hooks setup at all.
             + true (Miklos)
    + thought only-new-files, Win, Mac binaries etc. was non-controversial (Miklos)
       + wait to find someone interested in the server-side hacking shows up.
    + propose to get the clang-format with it in as a hook on (Thorsten)
    + can we re-format only the changed lines ? (Michael)
       + tricky, and could go wrong (Miklos)
       + establish as a test-bed for the future for all files (Stephan)
          + no point in the experiment if we can’t make it 100%
       + we can generate data on how sensible if it works ? (Michael)
    + should we disable web-edit instead ? (Stephan)
        + web-edit has CI running make check (Miklos)
    + concern wrt. checking clang-config bits in (Stephan)
        + fear creation of cleanup commits from people that break history.
        + possibly too pessimistic.
    + with subset of files / script operates only on these files (Kendy)
    + in theory random people do random things (Thorsten)
    => merge step#1 and re-visit if we have problems in a week
        + as it is on gerrit, but do nothing if clang-format is not installed
        + ok with that (Thorsten, Stephan)
    + gerrit concern – if have no-one able to change it (Thorsten)
        + will have problems going forward, concern it bites us later.
        + longer discussion, trade-off via gerrit vs. not (Miklos)

* Crash Hunting (Caolan)
    + 0 import failures, 1 export failures (ODT export)
    + 8 coverity
    + 34 fuzzers, but now enabled ubsan so...
    + 34 fuzzers * 2 sanitizers (asan and ubsan) * 1 engine (libfuzzer)
      => fuzz-apocalypse 44 bugs on Tue+Wed
      => 7 today so far, hopefully tailing off
      + doubled the number of fuzzers we have active
      - xls now added, will run tomorrow I imagine
      + MS word format work is good.

* Crash Reporter (Xisco)
    + http://crashreport.libreoffice.org/stats/?days=128#

    + http://crashreport.libreoffice.org/stats/version/5.3.6.1
        + 1527 (last 7 days)
    + http://crashreport.libreoffice.org/stats/version/5.4.1.2
        + 3168 (last 7 days)
    + http://crashreport.libreoffice.org/stats/version/5.4.2.2
    + 2141 (last 7 days )

* Hackfests & Events (Thorsten)
    + new plan from Bjoern:
       + https://skyfromme.wordpress.com/2017/08/27/hackfests-the-next-generation/
         + good suggestions from Regina
    + Hamburg – 1st new-style hackfest
         + 9th-10th December possible, if it fits.
         + just after Beta2
         + https://wiki.documentfoundation.org/Hackfests/HackfestTNG (Eike)
    + next after FOSDEM
    + Germany / France / Italy – considered for spring

* mentoring/easyhack update (Shinnok)
         committer... 1 week 1 month 3 months 12 months
                  open 62(3) 117(11) 118(12) 124(10)
               reviews 591(91) 1653(67) 5111(116) 17793(162)
                merged 284(55) 881(37) 2734(85) 10878(113)
             abandoned 25(4) 65(5) 190(2) 638(16)
           own commits 447(64) 1328(142) 3490(441) 13724(108)
        review commits 71(9) 239(19) 803(68) 2722(22)
        contributor... 1 week 1 month 3 months 12 months
                  open 16(-6) 40(-7) 42(-7) 46(-7)
               reviews 847(43) 2472(114) 6899(310) 22576(325)
                merged 32(16) 88(8) 345(24) 1375(18)
             abandoned 2(-2) 24(0) 97(1) 342(1)
           own commits 31(15) 88(7) 350(28) 922(17)
        review commits 0(0) 0(0) 0(0) 0(0)
    + easyHack statistics:
       needsDevEval 30(30) needsUXEval 0(0) cleanup_comments 200(200)
       total 255(255) assigned 11(11) open 211(211)
    + top 5 contributors:
          Gelmini, Andrea made 19 patches in 1 month, and 166 patches in 1 year
          Carl, Jens made 12 patches in 1 month, and 119 patches in 1 year
          Johnny_M made 7 patches in 1 month, and 43 patches in 1 year
          Udaree Kanewala made 6 patches in 1 month, and 6 patches in 1 year
          Jim Raykowski made 5 patches in 1 month, and 9 patches in 1 year
    + top 5 reviewers:
          Pootle bot made 260 review comments in 1 month, and 571 in 1 year
          Behrens, Thorsten made 145 review comments in 1 month, and 842 in 1 year
          Grandin, Noel made 141 review comments in 1 month, and 1336 in 1 year
          Stahl, Michael made 125 review comments in 1 month, and 1883 in 1 year
          Holešovský, Jan made 94 review comments in 1 month, and 1023 in 1 year
    + big CONGRATULATIONS to contributors who have at least 1 merged patch, since last report:
          Róbert Bán robert.ban102@gmail.com
          Gábor Koruhely koruhelygabor@gmail.com
          Furkan Tokac furkantokac34@gmail.com
          KARA Furkan Ahmet furkanahmetkara.fk@gmail.com
          Umann Kristóf dkszelethus@gmail.com
    + ssrlabs – thanks but out of mentoring B/W for them.
    + ping people directly ? (Heiko)
      + ask beginner developers – check this ticket …
      + a gentle encouragement to consider the next ticket is great (Michael)
         + problem compiling the list (Shinnok)
         + encouragement to build relationships & encourage contribution (Michael)
            + would be great if Shinnok can do same for non-UX (Heiko)

* Commit Access
    + propose Regina Henschel for commit access (Thorsten)
        => approved.
        => write a welcome mail (Thorsten)
    + propose Jens Carl (Markus)
        + doing lots of unit tests work (Eike)
        => approved.
        => write a welcome mail (Eike)

* Developer Certification (Stephan/Bjoern/Kendy/Thorsten)
    + next step is a meeting.

* Jenkins / CI update (Christian)
    master linux rel jobs: 220 ok: 219 ko: 1 fail ratio: 0.45 % break: 1 broken duration: 0.68%
    master linux dbg jobs: 149 ok: 147 ko: 0 fail ratio: 0.00 % break: 0 broken duration: 0.00%
    master mac rel jobs: 149 ok: 138 ko: 10 fail ratio: 6.71 % break: 10 broken duration: 4.47%
    master mac dbg jobs: 148 ok: 108 ko: 39 fail ratio: 26.35 % break: 6 broken duration:38.19%
    master win rel jobs: 92 ok: 83 ko: 9 fail ratio: 9.78 % break: 9 broken duration: 4.91%
    master win dbg jobs: 91 ok: 83 ko: 7 fail ratio: 7.69 % break: 6 broken duration: 5.62%
    master win64 dbg jobs: 96 ok: 88 ko: 8 fail ratio: 8.33 % break: 8 broken duration: 5.90%
    lo-5.3 mac jobs: 3 ok: 3 ko: 0 fail ratio: 0.00 % break: 0 broken duration: 0.00%
    lo-5.4 mac jobs: 0 ok: 0 ko: 0 fail ratio: 0.00 % break: 0 broken duration: 0.00%
    branch gerrit all jobs: 0 ok: 0 ko: 0 fail ratio: 0.00%
    master gerrit lin jobs: 511 ok: 434 ko: 75 fail ratio:14.68%
    master gerrit plg jobs: 509 ok: 361 ko: 137 fail ratio:26.92%
    master gerrit win jobs: 508 ok: 399 ko: 104 fail ratio:20.47%
    master gerrit mac jobs: 511 ok: 300 ko: 204 fail ratio:39.92%
    master gerrit all jobs: 509 ok: 203 ko: 284 fail ratio:55.80%
    + looks worse than they are
       + high failure rate of Mac compared to the rest
  + queue is quite large – a windows build ran out of disk
     + have brought it back.
     + expect the latency to improve

* l10n (Sophie)
    + work ongoing.
    + script / gettext migration (Cloph)
        + very picky wrt. Line-endings – build breaks hard
        + need to improve on the script to fix these too.
    + bit more work needed on getting export automated.
    + more work than needed (Sophie)
        + more for Cloph – checking the strings before build (Cloph)
        + translators concerned – lots of work that should not be there.
           + mistaken fuzzy strings issue (Cloph)
              + plan to auto-remove the fuzzy marker.
              + hope to have it improved a lot this week.

* QA update (Xisco)

    + Added list of bugs with 10 people or more in CC
        + https://wiki.documentfoundation.org/QA/Stats/MostCCBugs
    + Added list of most duplicated bugs
        + https://wiki.documentfoundation.org/QA/Stats/MostDuplicatedBugs

    + unconfirmed down to around 450 as of today.

    + UNCONFIRMED: 497 (-36)
        + enhancements: 64 (-4)
        + needsUXEval: 0 (+0)
        + haveBackTrace: 9 (+0)
        + needsDevAdvice: 32 (+0)
        + documentation: 3 (-1)
        + android: 20 (+1)

    + Most Pressing Bugs: http://tdf.io/mostressingbugs
        + New:
             [ none ]

        + Older:
             + CRASH: When closing the document before closing calc warning dialog
                   + https://bugs.documentfoundation.org/show_bug.cgi?id=113160
                   + bisected to Armin (?)

             + CRASH: bad dynamic_cast! after redo operation
                   + https://bugs.documentfoundation.org/show_bug.cgi?id=108124
                   + Patch in gerrit: https://gerrit.libreoffice.org/#/c/43862/
        + would be great to have a view from Michael S.

        + Fixed:
             + Crash in: ScPreview::MouseMove(MouseEvent const &)
                   + https://bugs.documentfoundation.org/show_bug.cgi?id=113422
                   + thanks to Noel Grandin.

             + Crash when editing code “ActiveCell.String =” in BASIC IDE
                   + https://bugs.documentfoundation.org/show_bug.cgi?id=113519
                   + thanks to Julien Nabet

* QA stats
    + https://bugs.documentfoundation.org/page.cgi?id=weekly-bug-summary.html
      +148 +64 (-188) overall)
      many thanks to the top bug squashers:
       Xisco Faulí 86
       Maxim Monastirsky 7
       Heiko Tietze 7
       Buovjaga 7
       Adolfo Jayme 5
       Telesto 4
       Aron Budea 4
       Yousuf Philips (jay) 4
       Tamás Zolnai 4
       Caolán McNamara 4

    + top 10 bugs reporters:
       Yousuf Philips (jay) 20
       zyklon87 6
       Regina Henschel 5
       Telesto 5
       Christian Lehmann 4
       Ole Tange 4
       Xisco Faulí 2
       Aron Budea 2
       Jon R Kibler 2
       Dr. Matthias Weisser 2

    + top 10 bugs fixers:
       Philips, Yousuf 4
       Zolnai, Tamás 4
       McNamara, Caolán 3
       Vajna, Miklos 3
       Justin L 3
       Szymon Kłos 3
       Nabet, Julien 2
       Adolfo Jayme Barrientos 2
       Tietze, Heiko 2
       Faulí Tarazona, Xisco 2

    + top 10 bugs confirmers:
       Faulí Tarazona, Xisco 49
       Philips, Yousuf 20
       Buovjaga 18
       Dieter Praas 9
       Foote, V Stuart 6
       Budea, Áron 5
       Raal 5
       Tietze, Heiko 5
       *UNKNOWN* 3
       McNamara, Caolán 2

* Highest-Priority bugs (aka "MABs"):
   + http://bit.ly/2lJAAJI
     5.3 : 1/46 - 2 % (+0)
     5.2 : 2/39 - 5 % (+0)
     5.1 : 1/35 - 2 % (+0)
     5.0 : 2/62 - 3 % (+0)
     4.4 : 3/76 - 3 % (+0)
     4.3 : 5/74 - 6 % (+0)
     4.2 : 6/134 - 4 % (+0)
     4.1 : 3/83 - 3 % (+0)
     4.0 : 4/83 - 4 % (+0)
     old : 26/258 - 10 % (+0)

* Bisected bugs open: keyword 'bisected'
   + more accurate - down to a single commit.
   + http://bit.ly/2dyIfDy
   + 417/1716 410/1698 410/1669 400/1654 385/1631 380/1614 371/1593

     done by:
        Faulí Tarazona, Xisco 11
        Raal 3
        Budea, Áron 2
        Kaganski, Mike 2
        Weghorn, Michael 1
        Telesto 1

* Bibisected bugs open: keyword 'bibisected'
   + http://bit.ly/2cSCXlS
   + 502/2320 494/2299 490/2265 476/2247 462/2225 461/2212 451/2191

     done by:
        Faulí Tarazona, Xisco 11
        Budea, Áron 3
        Raal 3
        Telesto 2
        Kaganski, Mike 2
        S. Christian Collins 1
        Weghorn, Michael 1
        Alex Thurgood 1

* all bugs tagged with 'regression'
   + 885(-3) bugs open of 6501(+16) total 7(+0) high prio.

     done by:
        Faulí Tarazona, Xisco 8
        Kaganski, Mike 2
        Raal 1
        Kosiorek, Bartosz 1
        Weghorn, Michael 1
        Telesto 1
        Stanić, Mihovil 1
        Foote, V Stuart 1
        Alex Thurgood 1

* ~Component count net * high severity regressions
   + http://bit.ly/1HWHb3E
     Impress - 2(-1)
     BASIC - 1(+1)
     graphics stack - 1(+0)
     Writer - 1(+0)
     LibreOffice - 1(+0)
     Calc - 1(+0)

   by OS:
     Windows - 0(+0)
     Mac OS X - 3(+0)
     All - 4(+0)
     Linux - 0(+0)

* ~Component count net * all regressions
   + http://bit.ly/1BUdI8i
     Writer: other - 157(+0)
     Calc - 138(+3)
     Impress - 93(-4)
     Writer: docx filter - 72(+0)
     LibreOffice - 71(+1)
     UI - 50(+0)
     Writer: doc filter - 36(+0)
     graphics stack - 36(+1)
     Draw - 33(+0)
     Base - 31(+1)
     Borders - 29(+1)
     Writer: perf - 27(+0)
     Writer: other filter - 26(+0)
     filters and storage - 25(+0)
     Chart - 19(+0)
     Crashes - 19(-1)
     BASIC - 18(+0)
     Printing and PDF export - 18(+1)
     framework - 5(+0)
     Linguistic - 2(+0)
     Extensions - 2(+0)
     Installation - 1(+0)
     sdk - 1(+0)