/srv/irclogs.linaro.org/2018/01/23/#linaro-meeting.txt

stooartGood morning - we have njian and zhoyao01_ ? Possibly ZhengXu09:30
stooart?09:30
ZhengXugood morning09:31
ZhengXunjian is coming09:31
njianHi09:31
stooartGood afternoon ZhengXu, njian !09:31
stooartLet's get going.09:31
stooart#startmeeting09:31
linarobotMeeting started Tue Jan 23 09:31:47 2018 UTC.  The chair is stooart. Information about MeetBot at http://wiki.debian.org/MeetBot.09:31
linarobotUseful Commands: #action #agreed #help #info #idea #link #topic.09:31
zhoyao01_hi09:31
stooart#topic OpenJDK (East)09:31
stooartOh good, linarobot is working.09:32
stooartAs is tradition, do you want to go first njian?09:32
njianok09:32
njianJDK-8195588 patch has been pushed by Felix.09:32
stooartAh, Felix.09:32
njianI asked him for help.09:32
njianI was following JDK-8195685. jdk10 and jdk/jdk should work, while jdk/hs stills needs a repo sync for JDK-8195588. So currently, jdk/hs build is broken.09:33
njianI also found a potential issue today. While building jcstress with latest jdk10, it crashed.09:34
njianThis also seems to be related to "8174962: Better interface invocations", but I am trying to reproduce and root cause.09:34
njianAt least, without 8174962, jcstress build is OK.09:34
njianI submitted several jdk10 tests before, and all looked fine at that time.09:35
stooartUrk. I don't have a recent jdk10 build to check.09:35
stooartI'll be speaking to Andrew about this.09:35
njianOK, you can have a check, if you have a chance. I submit the issue as LEG-687 internally for now.09:36
njianAnd I think I will email adinn or aph for this issue.09:36
stooartPlease, do.09:37
njianok09:37
njianRegarding to SVE, I haven't got much time on that. Just make the jtreg test worked with armie.09:38
njianAnd there's some cases failure, due to JVM/armie ErrorHandling09:38
stooartIt is understandable that a dump of security code may cause issues, but it would be nice if there was better coordination so as not to break OpenJDK.09:38
njianYes09:39
stooartI think it'll be a topic at FOSDEM. I can badger people in person at least.09:39
njianYes09:40
stooartIs armie getting in the way of the framework? I imagine there will be some complications with how java is launched.09:40
njianYes, there seems to be some issue09:42
njianBut I don't know for details, needs investigation.09:42
njianIt's much better than I expected, at least most cases work.09:43
stooartOk, hopefully there won't be anything too intractable. If there are, it would be acceptable to exclude awkward  failing tests that have nothing to do with SVE changes.09:44
njianYes, I agree.09:45
stooartAnything you'd like to add njian?09:47
njianno09:47
njianthat's all from me.09:47
stooartOk, thanks njian, once you get so far I think it would be worthwhile you sharing instructions.09:47
njianinstructions? about armie?09:48
stooartzhoyao01_, how are you doing?09:48
zhoyao01_I'm working on fixing the error when ld/st merging is enabled for negative offset.09:48
zhoyao01_And it turns out that it is caused by the implementation of MacroAssembler::zero_memory() has an assumption about a group of str instructions size.09:49
zhoyao01_This assumption fails when pair of str are replaced by stp. So the error is fixed by disable stp merging for this function.09:49
zhoyao01_And the patch is under internal reviewing now.09:49
zhoyao01_That's all from me.09:49
stooartOk, sounds good.09:49
stooartThanks zhoyao01_ .09:49
stooartLast week I booked FOSDEM 2018 - https://fosdem.org/2018/schedule/track/free_java/09:50
stooartaph and ednevill are going.09:50
stooartI've been working on replacement automation for the M-LAB. I'm doing it a bit differently from how it was before, using the dsl-job building and pipelines.09:51
stooartCurrently just on a single jenkins node, but it will be quicker than the apm machine I ordinarily use.09:52
stooartI was further behind on the monthly reports than I thought, so I submitted them on Monday, as the deadline was Thursday last week (and not Thursday this week..)09:53
njianI checked the agenda of FOSDEM, seems that many familiar names will be there.09:53
ZhengXuIs there a chance to meeting John Rose in FOSDEM?09:54
stooartYes - I might have some questions regarding the treatment of platforms Oracle doesn't care about.09:54
stooartI don't know if John will be there.09:54
stooartBut, yes, I understand your thinking.09:55
ZhengXu:)09:55
njian:-)09:55
stooartThere are some old friends I want to speak to about on the same topic.09:55
njianAnd don't forget to share some pics. :-)09:55
stooartSure, will do. Brussels is interesting. It is just myself that is going. My wife and daughter won't be attending this year.09:56
ZhengXuYou can make your daughter become another Java expert. :)09:57
stooartMy patch hasn't been merged yet - aph had been busy on the interface issue. I'll ask him today, althought arguably it's not terribly important.09:57
stooartSome day!09:58
stooartOne thing I need to do is help Daniel find some work to do. Do you have any suggestions for what you think might be useful?09:59
stooartOtherwise, that is me. Automation and reports.09:59
stooartZhengXu, have you anything to add?09:59
ZhengXuBy the way, will Linaro automation system be publicly accessible? The reason I am asking is, njian saw some failure with jdk10 branches. But it is not very convenient to send email to describe the issue.10:00
ZhengXuIf we can just share a link with the error log, it should be much clear to others.10:00
stooartSemi-public. The jenkins won't be, but I intend on publishing to openjdk.linaro.org as before.10:00
stooartThe actual build failures I want to change, rather than just failing silently.10:01
ZhengXuError log be good enough. I think people won't care about jenkins configuration.10:02
ZhengXunothing else for my side.10:03
stooartIndeed. I need to be careful about what is public.10:03
stooartOk, great.10:03
stooartThanks everyone, and have a good evening.10:03
stooartOk linarbot, do your business...10:03
stooart#endmeeting10:03
linarobotMeeting ended Tue Jan 23 10:03:35 2018 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)10:03
linarobotMinutes:        https://irclogs.linaro.org/meeting-logs/linaro-meeting/2018/linaro-meeting.2018-01-23-09.31.html10:03
linarobotMinutes (text): https://irclogs.linaro.org/meeting-logs/linaro-meeting/2018/linaro-meeting.2018-01-23-09.31.txt10:03
linarobotLog:            https://irclogs.linaro.org/meeting-logs/linaro-meeting/2018/linaro-meeting.2018-01-23-09.31.log.html10:03
stooartAh good, it worked.10:03

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!