minutes of ESC call ...
Michael Meeks
michael.meeks at collabora.com
Fri Sep 4 06:11:49 PDT 2015
* Present:
+ Caolan, Sophie, Stephan, Christian, Michael M, Kendy, Jacobo, Miklos, Eike,
Norbert, Robinson, Kendy, Muthu, Bjoern, Lionel (late)
* Completed Action Items:
+ give Mark Hung commit access & mail him (Norbert)
+ have a look at revert notification (author + committer + the person who reverted) (Norbert)
[ still not done. but, fyi. in 2015 so far there have been 150 reverts
79 of these were self-revert.
These involved 43 original author
and 21 'reverter'.
of the 71 that were not self-revert, 40 or so where revert involving core dev
on both side.. that should be talking to each other already.
That leave us with ~4 revert a month involving non core dev bein reverted.
Analyzing the data I also found a problem:
commit da0a44b415f630ce39617b8c8f0c366177a7705e
Author: Andras Timar <andras.timar at collabora.com>
Date: Sat May 23 14:26:27 2015 +0200
tdf#91429 do not convert ooxml preset shape modifier values
Revert "Resolves: #i124703# adapt coordinate range from ms relative..."
This reverts commit c764a3d978beb2e6197a8d3f7df53d81ebf72467.
The reverted commit did not fix anything in LibreOffice in the
first place. The document referenced in #i124703# is as good
without the patch as with it. But it caused a regression.
The original author of da0a44b415f630ce39617b8c8f0c366177a7705e
was Armin Le Grand, surely we do not want to send automated 'beware of revert' email in that case :-)
So, all in all, I now think that it is not worth the aggravation, the extra load and the risk of mishap.
And that we should stick to a social rule: notify the relevant parties when reverting.
=> please be polite when you revert & notify the person ]
+ 'make -k stagingcheck' prototype (Bjoern)
http://cgit.freedesktop.org/libreoffice/core/commit/?id=c1f47481eb78df3d73bed1da5020ed6ea565a999
[ just like subsequent-tests, but for unstable tests, a holding-ground
for not perfectly stable tests; first write tests & then fix it etc.
Known-probably failing tests etc.]
+ re-arrange the help XML for the menu changes (Jay)
[ committed the first help fix yesterday evening.
Review & encouragement appreciated: nice work Jay !
https://gerrit.libreoffice.org/#/c/18288/ ]
+ promote Development Mentoring Lead Job Offer through all channels: Universities, Companies, Clubs, ex-GSoCers etc. (all)
https://blog.documentfoundation.org/2015/07/15/tdf-freelance-job-opening-201507-01-development-mentoring-lead/
[ need more applicants, help appreciated pwrt. finding experienced candidates.
application period is theoretically over. ]
* Pending Action Items:
+ look at 'make check' failures on win32 again if interested (Michael S)
+ have a look at the CppUnit to implement 'expected failure' (Jan-Marek)
[ investigated the current CPPUNIT version and Qt test suite,
will be able to hack on that the next week @DebCamp.
was busy at the conference. ]
+ review Giuseppe's webdav work in gerrit (Kendy/?)
+ https://gerrit.libreoffice.org/#/c/17189/
[ Merged a couple of patches (Michael) ]
+ UserAgent - drop bundled-languages (Michael)
+ UserAgent - produce a patch for review (Michael)
+ add a conditional to disable google wearable bits for freedroid (Cloph)
+ ask students to write their own wrap-ups to aggregate (Cedric)
+ rename blocker -> very-critical (Robinson)
[ several pieces to update; waiting for patch ]
+ restrict access to very-critical-ity (Robinson)
[ testing patch - eager to do it all at once ].
+ create new queries for MAB-like bits for each release (Robinson)
+ failing that a description of the problem & sample queries.
[ have some queries they don't act the same way as before
+ https://bugs.documentfoundation.org/buglist.cgi?f1=version&o1=regexp&priority=highest&product=LibreOffice&resolution=---&v1=^5.0.*
+ that's not-resolved high-priority bugs in 5.0 -> the new MAB.
AI: + find & reward / promote whomever did the 1000'th (Robinson)
* Release Engineering update (Robinson)
+ 4.4.6 (last of the 4.4.x line) - status
+ RC1 - week 40 (week starting 28th September)
+ 5.0.2 - RC1 - freeze Thursday
+ plan to tag soon.
+ some translations cut off
+ Rene provided his logs, this complains about non-matching tags, but
the XML seems OK. He'll try with LO included tool chain to double-check (Christian)
+ ? Rene to use patch to use lo-shipped components to build l10ntools
http://anonscm.debian.org/cgit/pkg-openoffice/libreoffice.git/commit/?id=53d3a4e18d95579050d7e6302bda2ab416381d3a
+ closed.
+ Late OpenGL Status (Michael)
+ loads of bugs fixed this week:
losts contexts, threading issues, driver crashes,
white-listing / black-listing
=> Windows 7+ and v. new GL drivers.
+ has been merged as a late feature
AI: + write up some blog / blurb on it (Michael)
+ huge number of stability & completeness fixes in the last days;
cf. tracker bug:
+ https://bugs.documentfoundation.org/showdependencytree.cgi?id=93529
+ couple more fixes coming in last-minute
+ lots of testing & fixing expected for RC2
+ please test (your default) settings with 5.0.2 RC1
<how to test added later>:
+ does it turn on GL rendering ? (check help-about for XXX-GL)
+ turn it on and have a go; but lower prio. for black-listed bits.
</added later etc. ;->
+ Android & iOS Remote
+ nothing new.
* UX Update (Kendy)
+ Upstreaming elementary theme -- tdf#92458, lp#1483914 (Bjoern)
+ Xubuntu are using a home-grown theme.
shipped as a separate package; has its own repo.
like to push that into our repo & fold them in.
+ some filter-branch magic to get it moved around etc.
+ asked for MPLv2; if not (but not ideal) add a custom license file.
+ does anyone have objections to commit access for this specific
bit to maintainers ?
=> just give the right few guys accounts (Bjoern)
+ no meetings for the holiday period
+ talked about templates:
+ didn't sort the wheat from chaff yet
+ encouraged Jay to push the best bits to master, and iterate there.
* Documentation (Olivier/Kendy)
+ move to flat ODT progress / filter issues ? (Michael)
+ Regina asked if it's possible to maintain help source as ODF
+ flat-odf - so filters can be used off-line
+ an easy win.
+ already have the filters; if anyone wants to do that: go for it.
* Crashtest update (Caolán)
+ 1 import failures, 1 export failures, 5 coverity warnings
+ updated test cases, 1 new import failure, fixed
+ one old test case odb export failure, not reprodicible
+ new coverity warnings all fixed
+ all looking good on fuzzing ?
+ went through hanging docs on the week-long fuzzing run.
+ escher issues peaked
+ docs. that are hanging in drawinglayer now
+ performance bottlenecks in drawinglayer.
+ endless sub-division (Michael)
+ had a run-in with polygon sub-divider: (Caolan)
+ dashed 19km straight-line
+ might want to re-visit the 1st attempt;
+ pushing the size of view-port all the way down.
+ writing an Excel loader next.
* GSoC (Cedric)
+ Failure rate of students this year.
+ Initially 12 students
+ 1x failed at mid-term
+ Another failed at final.
+ Almost another fail at final.
+ Kendy on vacation last week
+ but reasonably ok with the outcome.
+ concerning - the last 2x troublesome ones:
+ were returning students -> so we messed up selection ?
+ we expect students to be self motivated & responsive.
+ Hopefully we'll have more TDF mentoring B/W next year (Michael)
+ mentor summit: Cedric & Moggi confirmed going.
* Hackfests (Bjoern)
+ Aarhus LibreOffice conference bits ?
+ Hackfest topic: "C++11 in LibreOffice"
+ look into easy, easy C++11 EasyHacks for that
+ like Stephans "init list ctors for UNO-Sequences", or
using type-alias etc.
+ !!!please file easy hacks if you have ideas!!! (Bjoern)
+ Hamburg, Germany: Oct 23rd -> 25th
+ https://wiki.documentfoundation.org/Hackfest/Hamburg2015
+ would be great to see everyone there.
+ next venues / suggestions
+ Belgium another candidate ...
* l10n Update (Sophie)
+ nothing new.
* Conference (Sophie)
+ looking great; hope to have a programme by Monday.
* Developer Certification (Stephan/Bjoern/Kendy/Thorsten)
+ pending / shelved for now - wait 1 week.
* Jenkins update (Norbert
from:Thu Aug 27 14:04:44 2015
master linux rel jobs:223 ok:209 ko:13 fail ratio:5.82% break:7 broken duration:3.22%
master linux dbg jobs:128 ok:110 ko:10 fail ratio:7.81% break:6 broken duration:7.47%
master mac rel jobs:207 ok:196 ko:11 fail ratio:5.31% break:6 broken duration:4.14%
master mac dbg jobs:194 ok:181 ko:13 fail ratio:6.7% break:6 broken duration:12.52%
master win rel jobs:80 ok:39 ko:41 fail ratio:51.24% break:8 broken duration:60.66%
master win dbg jobs:150 ok:88 ko:62 fail ratio:41.33% break:13 broken duration:54.82%
master win64 dbg jobs:144 ok:56 ko:88 fail ratio:61.11% break:11 broken duration:62.2%
lo-4.4 mac jobs:20 ok:14 ko:6 fail ratio:29.99% break:4 broken duration:19.94%
gerrit win rel jobs:331 ok:137 ko:193 fail ratio:58.3%
gerrit lin rel jobs:336 ok:278 ko:58 fail ratio:17.26%
gerrit mac rel jobs:335 ok:260 ko:73 fail ratio:21.79%
gerrit all jobs:364 ok:125 ko:237 fail ratio:65.1%
+ Windows terrible last week. 50-60% broken.
+ translates to gerrit
+ some of them infra problems; lots of time while
+ painful to trigger a new build when infra fails (Michael)
+ use the 'rebase' button to do that (Norbert) - just push it.
+ new H/W is in-place, and more coming.
+ if people push something that breaks -> it breaks master.
AI: + advertise that verification is now much faster (Norbert)
+ was awesome overnight: (Michael)
+ the new build times should look like (Norbert)
+ 30-40mins for Windows
+ mac / linux 5-7 mins gerrit
+ some Mac builders up to 1:30 with bad cache
+ 2x other Linux machines coming soon.
+ see Norbert's forthcoming talk in Arhus for more ...
* Perf-testing update ? (Norbert)
+ http://perf.libreoffice.org/
+ git://gerrit.libreoffice.org/perfdbmgr
+ now collecting make perfcheck result ... no graph yet
+ still some work needed to improve the web front-end
+ talk to Markus - wrt. how to do it.
+ do a unit-test & trigger a valgrind dump around a specific section.
+ can give nice narrow measurement - without drowning in cost of configmgr.
* new landing page website
+ http://devcentral.libreoffice.org
+ intented to present in one page all the tdf online-resource related
to libreoffice development.
+ patch welcome. devcentral.git under gerrit.
* Easy Hacks:
+ new easy-hacks thanks to:
+ Michael Meeks x2
+ http://bit.ly/1KR4HWX
+ create your own by setting whiteboard to eg.
"easyHack difficultyBeginner skillCpp topicCleanup"
* QA update (Robinson)
+ UNCONFIRMED at 524
+ growing backlog ATM
+ needAdvice at 15
+ Most Pressing Bugs:
http://ur1.ca/nkp0t
+ can we remove 'UNCONFIRMED' too ? (Michael)
+ ( we can, but UNCONFIRMED/highest begs for triage attention.. )
+ tdf#92478 - Scanning crashes on 64bit Windows 10 build
+ gentle nudge
+ tdf#93600 - Condition Formatting: Entering "=" in comparison value crashes program
+ fixed it & merged to -5-0 thanks to Bubli
+ tdf#91882 - UI: Dialogue for Advanced filter is transparent.
[ horrific idle loop -> fixed & in -5-0 (Michael) ]
* Bibisected regressions (Bjoern)
+ tagged the rendercontext specific ones to the tracker
+ 8 currently open to fix.
* QA stats
+ https://bugs.documentfoundation.org/page.cgi?id=weekly-bug-summary.html
+178 -94 (+84 overall)
many thanks to the top bug squashers:
Maxim Monastirsky 22
Michael Meeks 7
tommy27 6
raal 6
Timur 5
Vasily Melenchuk 3
Julien Nabet 3
Xisco Faulí 2
Björn Michaelsen 2
Beluga 2
Samuel Mehrbrodt 2
m.a.riosv 2
Yousuf (Jay) Philips 2
Alex Thurgood 2
Miklos Vajna 1
* Open 5.0 MAB
+ 1/12 1/12 1/12 1/12 1/12 1/12 1/12 1/12 1/12 2/12 2/12 3/12 5/11
+ https://bugs.documentfoundation.org/showdependencytree.cgi?id=86696&hide_resolved=1
* Open 4.4 MAB
+ 10/60 11/60 11/60 11/60 11/60 10/60 10/60 10/60 10/60 12/60 14/61 16/61
+ https://bugs.documentfoundation.org/showdependencytree.cgi?id=79641&hide_resolved=1
* Open 4.3 MAB
+ 59/168 59/168 59/168 59/168 61/168 61/168 61/168 62/168 62/168 64/169 65/169
+ https://bugs.documentfoundation.org/showdependencytree.cgi?id=75025&hide_resolved=1
* Bibisected bugs open: whiteboard 'bibisected'
+ 344/1028 345/1020 338/1006 328/991 314/964 309/956 311/951 305/936
+ http://tdf.io/bibibugs
* all bugs tagged with 'regression'
+ 624(+8) bugs open of 4158(+21) total 38(-2) high prio.
* ~Component count net * high severity regressions
Calc - 17 (+1)
Writer - 9 (+0)
LibreOffice - 5 (-2)
Impress - 4 (-1)
UI - 1 (+0)
filters and storage - 1 (+0)
Base - 1 (+0)
+ http://bit.ly/1HWHb3E
* ~Component count net * all regressions
Writer - 210 (+2)
Calc - 126 (+2)
Impress - 46 (-1)
Crashes - 43 (+0)
LibreOffice - 37 (-1)
UI - 38 (+1)
Base - 26 (+0)
graphics stack - 25 (+1)
Borders - 22 (-1)
Draw - 21 (+0)
filters and storage - 12 (-1)
Chart - 11 (+0)
Printing and PDF export- 10 (+0)
BASIC - 7 (+0)
Formula Editor - 5 (+1)
framework - 5 (+0)
Linguistic - 4 (+0)
Extensions - 2 (+0)
--
michael.meeks at collabora.com <><, Pseudo Engineer, itinerant idiot
More information about the LibreOffice
mailing list