IRC logs for #baserock for Wednesday, 2015-03-25

*** zoli__ has quit IRC00:01
*** zoli__ has joined #baserock04:30
*** zoli__ has quit IRC04:43
*** zoli__ has joined #baserock05:06
* straycat clicks a change07:23
straycatoh, no, can't see that, my session's expired07:23
* straycat goes back to the list of changes07:23
* straycat clicks a change07:23
straycatcan't actually see the diff, just a load irrelevant information about the change07:24
* straycat clicks the file07:24
* straycat clicks back07:24
* straycat clicks the other file07:24
* straycat clicks back07:24
* straycat looks for the change this change depends on07:24
straycatoh no my session expired again07:25
straycatOkay screw this, I'm just going to fetch refs and read the commit log from now on, and if we want inline comments 2.11 will let us write a client for that, or maybe once we get https gertty will turn out to be okay07:30
* straycat uses gerrymander to see if anyone's commented on any of his changes07:32
paulsherwoodouch. straycat - maybe stick with the ML, until others convince you that the rough edges have been ironed out?07:40
* straycat nods07:43
* paulsherwood can see how fresh-faced young sofware engineers happy to try anything turn into bearded ones who only use commandline tools and bark a lot :)07:45
Kinnisonpaulsherwood: yep07:49
Kinnisonpaulsherwood: Did you get anywhere with that suggestion from me last night?07:50
*** mike has joined #baserock07:54
*** mike is now known as Guest7058207:55
*** a1exhughe5 has joined #baserock08:06
*** gfinney has joined #baserock08:15
*** gfinney has joined #baserock08:15
*** mdizzle has joined #baserock08:46
*** bashrc_ has joined #baserock09:02
*** gary_perkins has joined #baserock09:04
*** ssam2 has joined #baserock09:17
*** ChanServ sets mode: +v ssam209:17
jonathanmawSotK: I'm trying to get GENIVI Demo Platform working on a Jetson, and note that you've got some definitions for the GENIVI Demo Platform. How complete are those definitions?09:24
*** rdale has joined #baserock09:27
ssam2weird error from system-version-manager:
*** tiagogomes_ has joined #baserock09:29
ssam2this must be due to a change of behaviour in the 'mount' command09:30
ssam2we upgraded to util-linux 2.25.2 on 19th Feb09:31
richard_mawthat's not necessarily a `mount` failure, we saw something similar with a btrfs bug09:34
ssam2right... this is indeed btrfs09:36
ssam2seems there were quite a lot of changes in libmount, but nothing that says 'stop people mounting stuff twice'09:36
ssam2wow, actually there are loads of errors in 'dmesg'09:37
SotKjonathanmaw: I don't remember entirely... I think I'd got to the point where I could deploy a system with Weston/Wayland + Qt + the GENIVI browser POC and was trying to get fuel-stop-advisor to build nicely09:46
*** edcragg has joined #baserock09:48
ssam2seems that mounting the filesystem on a different machine and/or running btrfsck has fixed whatever problem I was having09:48
*** Krin has joined #baserock10:10
*** mauricemoss_ has quit IRC10:31
*** simonh_ has joined #baserock10:31
*** wschaller has joined #baserock10:33
*** locallycompact has quit IRC10:40
jjardonHi! Im trying to add some comment to a morphology but doesnt seems to work: this works fine though . Any idea what is the prefered way to do this?10:49
paulsherwoodKinnison: i got into a tangle10:51
Kinnisonjjardon: You're missing a backtick10:51
Kinnisonpaulsherwood: oops :)10:51
paulsherwoodKinnison: you did manage to completely shock me10:52
Kinnisonpaulsherwood: Oh?10:52
* Kinnison hides the taser behind his back10:52
paulsherwoodKinnison: i started by doing
Kinnisonyeah that won't work at all10:53
paulsherwoodit seems to me if stage1 and stage2 are only build depends, it *should* work?10:53
Kinnisonbecause build-essential-bootstrap contains no non-bootstrap chunks10:53
paulsherwoodinteresting 'logic' here :-)10:54
KinnisonAs I said last night, bootstrap chunks exist *ONLY* as explicit build-dependencies of non-bootstrap chunks in the same stratum10:54
* Kinnison perhaps didn't say that clearly enough10:54
paulsherwoodi hoped that ybd would be happier, but it just exposed other problems in the recursion10:54
* Kinnison has prepred a fresh baserock chroot and acquired ybd from your master10:55
* Kinnison may have a tweak10:55
paulsherwoodKinnison: i think you said it clearly enough... but it's still not clear enough to me. sorry, i can't think about this now.10:55
paulsherwoodas i said, this little exercise did also blow a hole in the ybd recursion implementation, so it's worse than i thought10:55
KinnisonOOI have your fixes to build-essential's b-deps got into maste ryet?10:56
paulsherwoodno, i don't believe so10:56
* Kinnison shall have to operate from your branch then10:57
* paulsherwood doesn't quite know why... it had 2 +1s :-)10:57
KinnisonIt has two +1s but Sam's comment indicates he'd like a fix to the comments in the morphology10:57
paulsherwoodah, missed that10:57
jjardonKinnison: sorry, I copy incorrectly; still doesnt work with the missing backtick addded:
Kinnisonjjardon: aah10:58
Kinnisonjjardon: that's because the way shell allows environment variables to be set requires that they be the first token10:58
Kinnisonjjardon: You probably need to shove an 'env' in there10:58
Kinnisonotherwise the ``d comment will be the first shell token10:59
Kinnisonhence your error10:59
* SotK wonders how sending a V2 patch to Gerrit works11:08
jjardonKinnison: thanks, that worked11:09
rdaleSotK: i got the impression that if you push it with the same change id as v1, it just works11:09
jjardonSotK: "force" push it to the same ref/to/master/<topic>11:09
jjardonSotK: I mean, simply push to that special ref, you dont have to use -f11:10
SotKjjardon: does it matter if the change-ids are different?11:11
KinnisonI think the changeids have t omatch for it to line up the patches11:13
* SotK wonders how his change-ids ended up changing in the first place, and fixes it11:13
jjardonSotK: no sure, until know I always rework the patches I already sent, so they have same change-id. Try it, you can alwyas discard the old patches if it doesnt work11:13
Kinnisonssam2: You +1d Paul's change for build-essential build-deps -- any hope that since it has two +1s it might get merged?11:17
jjardonKinnison: you gave me a +1 for the second commit of the series "no_name" ; that change is a consecuence of the first commit, but you didnt review it, maybe you want to take a look?11:20
KinnisonCan you give me a change URI?11:20
* Kinnison finds navigating the list of changes hard11:20
jjardon(seems the idea is not very popular already :))11:21
KinnisonYou appear to have two -1s already11:22
* Kinnison reads11:22
* Kinnison agrees with the -1s11:22
KinnisonDo you want me to review it formally?11:22
Kinnisonthe 'name:' field is meant to be a doublecheck that people haven't done something stupid :)11:22
jjardonKinnison: if you can, and change your review of the change in the README to be consistent11:23
ssam2Kinnison: I think that it needs a comment explaining why build-essential treats the build-depends field differently to all the other strata, so I don't want it merged as-is11:24
Kinnisonssam2: Aah, I saw it as a precursor to fixing it so build-depends always behaves like paul's change :)11:24
ssam2it could be... I'm not 100% sure on that now11:24
ssam2it certainly won't happen overnight anyway, so we need to document why things are inconsistent in the mean time if we merge this patch11:25
ssam2i think the next step towards making build-depends always behave like that would be to write a migration script11:26
jjardonKinnison: for me is a bit annoying because when  I create a bla.morph I have to put a field in the morphology like name: blah (and it has to be exactly the same)11:26
ssam2then we can see the actual implementation11:26
jjardonbut im fine with the -111:26
jjardonmaybe a better way to solve this is to allow morphologies to be inline in the stratum11:27
KinnisonI'd like that, but it's an even bigger change11:28
* Kinnison wants to see some smaller cleaner version updates happening first11:28
Kinnisonincl. migration scripts11:28
jjardonso we can fix the problem of some morph files not being moved when you move chunks around11:28
KinnisonHowever I'm still way behind on things (ca. 1100 mails still left to deal with in baserock-dev :)11:29
* straycat would hope for more discussion of the branch morph rfc before any further version updates occur11:30
jjardonseems people is out of ideas about what "smaller" updates should be done next11:30
straycatesp since they'll inevitably conflict with the series and force me to rework it11:30
richard_mawstraycat: I'll take a look after I get back from lunch11:30
straycatrichard_maw, awesome thanks :)11:31
jjardonstraycat: I think you are mixing 2 things here: I think you want to discuss how we want to act when we break compatibility. That doesnt mean we can not keep doing (not breaking) changes.11:32
bashrc_looks like there are strange entries in self.pluginmgr.location11:33
bashrc_'/usr/lib/python2.7/site-packages/morphlib/plugins', '/', 'u', 's', 'r', '/', 'l', 'i', 'b', '/', 'p', 'y', 't', 'h', 'o', 'n', '2', '.', '7', '/', 's', 'i', 't', 'e', '-', 'p', 'a', 'c', 'k', 'a', 'g', 'e', 's', '/', 'm', 'o', 'r', 'p', 'h', 'l', 'i', 'b', '/', 'p', 'l', 'u', 'g', 'i', 'n', 's'11:33
Kinnisonlooks like someone extend()ed instead of append()ing11:34
straycatjjardon, I don't really follow, the discussion is really about whether we should try and get ourselves away from the problem of unversioned definitions by branching morph, such that one branch supports unversioned/v0 and v1 and the other branch supports v1,v2,...11:37
SotKKinnison: I sent a v2 (and a v3, I messed up a rebase)11:39
KinnisonI will try and get to that today11:39
jjardonstraycat: whats the problem with "unversioned definitions = version 0" ? My point is that when you want to make a change in morph and the result will not support version 0, then we should think about removing support for version 0, not before. It would be weird that we encourage people to use latest morph but they cannot build its no-so-old definitions11:43
straycatIt depends on whether we have users of unversioned/version 0 who need to be supported, if we do then branching seems better, so that we can continue to maintain version0/unversioned whilst still being able to reduce code complexity (breaking compatibility) in master12:00
*** wschaller_ has joined #baserock12:07
*** wschaller has quit IRC12:10
* bashrc_ spotted a bug12:11
pedroalvarezthat one must be the last thing that needs fixing to finish Morph :)12:19
pedroalvareztiagogomes_: I don't really know why, but your pysendfile can't be merged :/12:48
tiagogomes_pedroalvarez, does it show any error information12:49
pedroalvarezlooks like someone added another lorry after pymox, so it's a merge conflict12:50
tiagogomes_again ¬_¬12:50
* tiagogomes_ needs to be faster12:50
pedroalvarezat least franred spotted a Json error :)12:50
tiagogomes_Ok, I sent another patch12:53
pedroalvarezfranred:  can you change your -2?12:54
franredpedroalvarez, tiagogomes_, merged12:55
pedroalvarezI think we shouldnt use -2's when the patch can be merged with fixes12:55
pedroalvarezotherwise we need to wait for whoever has -2'ed  it to merge the patch12:55
franredpedroalvarez, I didn't know that the -2 stay after sending a reworked version I though it disappear as the -1s12:56
pedroalvarezedcragg:  I've reviewed your armv8 patches, they need some changes13:12
*** pdar has quit IRC13:17
*** pdar has joined #baserock13:26
jonathanmawhrm, gstreamer-0.10 is failing to build for me, because the version of the gstreamer-common submodule it uses has some kind of parsing error with bison's verson.13:38
jonathanmawthe correct solution is to update the 0.10 version of the gstreamer-common code to include this patch.13:39
jonathanmawhow do I edit a submodule's code so that it gets picked up by baserock. I gather my gstreamer-0.10 that I've checked out with `morph edit` will be used if available13:40
jonathanmawand that includes a ref to gstreamer-common13:40
jonathanmawif I add a new ref to gstreamer-common, can morph find it?13:41
ssam2submodules are a nightmare13:42
ssam2you need a new commit of the gstreamer-0.10 repo that updates the submodule to the commit that you need13:43
ssam2morph effectively does 'git submodule init; git submodule update' to fetch the submodules, i think13:43
pedroalvarezthis commit fixed the parsing error in gstreamer-common:
jonathanmawso it'll be fetched from my local or remote artifact cache13:44
jonathanmawso I'll need a branch pushed to for gstreamer-common13:44
ssam2no, you'll need a branch pushed for gstreamer-0.10, if I understand correctly13:45
ssam2the commit is already present in gstreamer-common.git13:45
ssam2but you need a commit in the parent repo that references the submodule commit13:45
jonathanmawssam2: the commit in gstreamer-common is for later versions of gstreamer, not the gstreamer-0.10 branch. I tried changing gstreamer-0.10 to use that ref and it didn't go well.13:46
ssam2ah, right.13:51
ssam2is there any way to avoid using gstreamer 0.10? it's totally unmaintained13:51
ssam2I think13:51
jonathanmawI'm not sure.13:52
jonathanmawit looks like our qt strata use gstreamer-0.1013:53
radiofreejonathanmaw: just don't add qt5-tools-qtmultimedia13:53
radiofreebut if anyone wants to use qt5-multimedia then sadly it's still stuck at gstreamer 0.1013:54
* jonathanmaw needs to figure out if the GENIVI Demo Platform needs qt5-multimedia.13:55
paulsherwoodi believe it does13:56
SotKI think that is what I found last year13:57
pedroalvarezthere has been some porting effort:
jonathanmawthe yocto GDP seems to use meta-qt5, which refers to qtmultimedia13:57
radiofreeit's a massive pain in the arse13:57
radiofreeqtwebkit uses gstreamer 1.0, that's it13:57
jonathanmawyep, this meta-qt5 includes patches to port qtmultimedia to gstreamer 1.013:58
radiofreeare they the ones by yoann something?13:58
* jjardon wondering if they sent those patches upstream13:59
radiofreejjardon: if they're by yoann, then they are qt patches13:59
jonathanmawradiofree: yep, by yoann13:59
jjardonradiofree: Martin Jansa13:59
Krini need to get pyserial into baserock as a dependency of tempest. there are no official git repository of it, only .tar.gz on the main site and an svn repo. what would be the best was of going about getting this into baserock?13:59
radiofreejjardon: that's the patch to add the patches :P14:00
SotKKrin: write a lorry for the SVN repo and test it on your devel machine to see if it works/isn't really slow I think14:00
SotKthen use SVN if it works, but the tar if it doesn't14:00 indicates there's a wip/gstreamer-1.0 branch14:00
jjardonoh, rigth. So why they are not upstream yet?14:01
radiofreeyep, it's also on if it ever loads for you14:01
Krinwe can lorry in SVN's? well thats something i didnt know. thanks SotK14:01
radiofreejjardon: it is being worked on by digia, it's probably just not ready yet14:01
radiofreeeven though it's been a wip branch for ages now...14:01
radiofreejonathanmaw: i would have thought yocto are patching it because they don't include gstreamer 0.1014:02
radiofreeif should work fine with 0.1014:02
radiofreeIT'S HAPPENING!14:02
jjardonoh, nice, seems to they restart to work on in last decemnber14:04
radiofreeit's in the 5.5 branch now
tlsaCan anyone review please?14:18
tlsait's to update nano tarball to v2.4.014:22
rdalewhen we update a tarball based lorry git rep, are the previous commits kept, or is the whole repo overwritten?14:23
richard_mawprevious commits are kept14:23
rdaleah ok14:23
pedroalvareztlsa: looks ok +114:28
ssam2tlsa: looks fine. I can merge it if you push it to Gerrit14:28
tlsaoh, right I forgot about gerrit14:29
tlsai'll need to get hooks, amend a change-id onto it and stuff14:30
pedroalvarezseriously, is not that much14:31
ssam2but merging it will be 1 click, instead of a lengthy process14:31
tlsapedroalvarez: yeah, didn't say it was14:31
tlsajust that I forgot about it, and was going to push direct to master since its just a one line change14:32
*** edcragg has quit IRC14:36
*** edcragg has joined #baserock14:38
tlsassam2: done:
radiofreehow do i change my e-mail address in gerrit14:41
ssam2merged, thanks14:41
radiofreei can't - 500 Internal server error14:41
ssam2radiofree: it's tied to your OpenID I think14:41
radiofreethere's a "register new e-mail" option in the settings, but that 500's14:42
ssam2ah, probably because it can't send emails14:42
ssam2i'll hopefully fix that this week14:42
radiofreeso should we always morph check * gerrit branches now?14:44
radiofreeif we intend to send patches?14:44
radiofreewhen i click "sign in" it's automatically signing me in with the baserock openid :\14:45
radiofreei suppose this is some cookie thing14:46
tiagogomes_would a patch for adding tcpdump to tools be appreciated?14:49
*** jonathanmaw has quit IRC14:52
radiofreeso i did
radiofreeand it failed14:54
tiagogomes_wdutch, just saw your patch for libpcap. Are you planning to add tcpdump as well?14:54
radiofreeare you supposed to see some "change-id" in your commit message?14:57
ssam2you're supposed to have one, yes14:57
radiofreehow do you add that from an existing branch from gbo?14:57
wdutchtiagogomes_: I have no plans to - for openwrt I just need that lib14:57
ssam2radiofree: commit --amend, after installing the commit hook14:57
radiofreei have a branch on gbo, want to submit the top two patches, i did "reword"14:57
ssam2if you have the commit hook then reword might work too, I don't know14:58
radiofreei already have that14:58
radiofree ! [remote rejected] HEAD -> refs/for/master/qt5-upgrade (change 49 closed)14:58
ssam2the error message '(change 49 closed)' suggests that you're using the Change-Id from Change 4914:58
ssam2seems a bit weird though14:58
radiofreeit seems that the mirrored branchs on gerrit already have a change id?14:59
radiofreeno they don't15:00
* wdutch has a go at building a tcpdump chunk15:00
* radiofree is stuck15:00
tiagogomes_it builds fine as long you add  libpcap as its dependency15:01
pedroalvarezhehe, many of us using tcpdump these days :)15:03
radiofreeoh wait i think i did it wrong15:03
pedroalvarezwdutch: should be easy, let us know if you need help :)15:03
radiofreeit's split up my two commits into two different things and
wdutchtiagogomes_: do you think libpcap and tcpdump should share a lorry?15:04
tiagogomes_pedroalvarez is a must for having a clue of the dark magic that Neutron/openvswitch do15:05
radiofreehow does it handle patch series? do they all have to have the same change-id?15:05
tiagogomes_wdutch no strong opinion15:06
ssam2radiofree: each patch has to be an individual 'change' with its own change ID15:07
ssam2you can link them with 'topics' (push to refs/for/master/topicname)15:07
pedroalvarezwdutch: can we persuade you to use gerrit? :P15:07
ssam2radiofree: if you're working on a big feature you could also have a staging branch, so push everything to refs/for/staging/myfeature/... which hopefully saves half your feature getting committed to master and half of it not15:08
wdutchpedroalvarez: I forgot that things have changed15:09
* wdutch goes to read the wiki15:09
radiofreei see15:11
* radiofree doesn't15:11
radiofreei did push them to a single topic name though15:12
radiofree * [new branch]      HEAD -> refs/for/master/qt5-upgrade15:12
radiofreeit's fine for them to appear as two different things then?15:12
pedroalvarezradiofree: yes15:14
radiofreeah ok15:15
radiofreeso if my icu patch got another +1, it would be auto-merged?15:21
radiofreeor do both my patches need +2 before that happens?15:21
pedroalvarezthey can be merged separately (I think)15:22
pedroalvarezunless you create a dependency when submitting the patch for review15:22
*** zoli__ has quit IRC15:22
bashrc_apparently baserock uses storyboard for bug reports
bjdooksI'm getting:15:26
bjdooks[ TIME ] Timed out waiting for device dev-ttyS0.device.15:26
radiofreebjdooks: one moment15:29
radiofreethe rootfs is on a usb stick right?15:30
pedroalvarezKrin: have you actually tested that the svn import of pyserial worked?15:30
pedroalvarezKrin: how long it took?15:30
radiofreecan you do grep sha1 baserock/build-system-armv7bhf-rootfs.meta15:30
radiofreeoh btw it's not actually hf....15:31
bjdooksi can run init=/bin/sh15:31
bjdooksecho hello > /dev/ttyS015:31
bjdooksand I get hello back15:31
radiofreegrep sha1 /baserock/build-system-armv7bhf-rootfs.meta15:31
bjdooks    "sha1": "eee1143e45133d74fd78fdda450f6d2e57a52028",15:32
radiofreeis that from before you get anything from systemd?15:34
radiofree(the [DEPEND])15:34
radiofreeWORKSFORME (on a jetson)15:34
radiofreebjdooks: do you have CONFIG_FHANDLE enabled?15:37
Krinpedroalvarez, it tool about as long as it took me to read the output when i lorried it on it's own15:40
rdaleit would be nice to lorry qtwebengine and include that with the qt 5.4 changes15:41
bjdooksradiofree: ah, no15:41
Krini did it as a seperate lorry to test because it was taking forever to lorry in all of python-packages.lorry15:41
pedroalvarezKrin: cool, that makes sense15:41
Krinbut yes it was very quick15:41
pedroalvarezKrin: I made a comment in your patch anyway15:42
*** edcragg has quit IRC15:43
*** ssam2 has quit IRC15:45
*** edcragg has joined #baserock15:45
wdutchcould somebody please give me an example of what refs/for/master/my-topic-branch could be in git push gerrit HEAD:refs/for/master/my-topic-branch? Everything I try is getting rejected15:46
SotKWhat error is it rejected with?15:47
pedroalvareztry without '/'s?15:48
SotKwdutch: you need the refs/for/master bit15:48
Zarawdutch: it's literally HEAD:refs/for/master/ followed by the name of your branch15:49
SotKso refs/for/master/williamholland/libpcap15:49
wdutchaha, thanks!15:49
Zara(I was unsure of that as well, and thought 'refs/for/master' might stand for something; we should make the wiki clearer)15:49
bjdooksThank you for trying Baserock.15:50
bjdooksbaserock login: root15:50
bashrc_initially I also thought it stood for something, but it doesn't15:50
bjdooksNo mail.15:51
bjdooksLinux baserock 4.0.0-rc4+ #23 Wed Mar 25 15:49:27 GMT 2015 armv7b GNU/Linux15:51
* radiofree needs to update his big endian jetson15:51
radiofreeLinux baserock 4.0.0-rc1-00031-g4e4e5ff-dirty #9 SMP PREEMPT Mon Mar 23 17:36:16 GMT 2015 armv7b GNU/Linux15:51
*** dabukalam_ is now known as dabukalam15:52
*** dabukalam has joined #baserock15:52
radiofreethink the jetson wins as a build platform though15:53
radiofree~ # cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq15:53
Krinpedroalvarez, fixed :)15:53
radiofreebjdooks: CONFIG_FHANDLE fixed it then?15:53
pedroalvarezKrin, how easy was to send another version of the same patch?15:54
pedroalvarezI wonder if it's too complex, given that bashrc_ has sent a different patch to gerrit, and abandoned the first version of it15:56
* SotK found it fairly painless earlier15:58
Krinnot difficult, just have to copy the change ID of the original patch and put it on the new patch, this fools gerrit into thinking it is the same patch15:59
*** ssam2 has joined #baserock16:03
*** ChanServ sets mode: +v ssam216:03
franredKrin, pedroalvarez, should we use rather than a tarball for testresources ?16:21
Krini don't know any more franred. i'll make it be however it needs to be to get merged.16:23
paulsherwoodwe should not lorry tarballs, normally unless there's a clear need, for (say) bootstrapping16:24
pedroalvarezor incompetent upstream16:25
Krinok paulsherwood i'll make those changes now16:25
*** zoli__ has joined #baserock16:25
ssam2wdutch: somehow you managed to submit your libpcap lorry patch to gerrit twice, and both have been merged16:27
ssam2seems harmless though16:27
straycatit's not exactly that they're incompetent, it's that they're not expecting people to build from their source16:27
wdutchI submitted it twice because I couldn't find it the first time and thought I'd done something wrong16:27
pedroalvarezstraycat: yes, maybe I was too harsh, sorry16:28
Krinwhat would the "type": feild be for ?16:29
*** zoli__ has quit IRC16:43
straycatshall we have a gerrit bot then?16:44
* straycat has some time to kill16:45
*** a1exhughe5 has quit IRC16:51
straycati can also just listen to drum and bass though16:51
pedroalvareza gerrit bot for changes merged?16:57
straycatand proposals16:59
straycat's what the openstack folks seem to do16:59
ssam2i'd prefer emails to an IRC bot, personally. If you have time to spare maybe you could look at finishing off infrastructure.git branch sam/gerrit-exim416:59
straycati'd prefer emails too! heh17:00
ssam2the 2 commits in that branch could be moved to definitions.git either, i just worked in infrastructure.git because it was close to hand17:00
straycati'd definitely like a bot, all the other projects have one17:01
pedroalvarezwe already have one! :) It's a supybot17:05
pedroalvarezbrlogger: hey!17:05
brloggerpedroalvarez: Error: "hey!" is not a valid command.17:05
pedroalvarezhe is still there :)17:06
straycatif there's a gerrit module for supybot then cool17:06
bashrc_ok, is there any special magic which needs to happen in order to enable a plugin? The evidence is mounting that the firehose enable is never called17:12
perrylbashrc_: how do you mean, enable? it could be that needs to go into morphlib/plugins if it isn't already17:14 enable function17:14
straycatpedroalvarez, ?17:17
perrylahh, i'm not sure in that case i'm afraid :(17:18
*** Guest70582 has quit IRC17:21
bashrc_and the morph plugin manager only runs after parsing of --config. Probably there's something I'm missing about how plugins are supposed to work.17:23
ssam2the plugin mechanism is provided by 'cliapp'. I doubt its innards are documented, but it's fairly simple Python code17:26
ssam2git:// branch baserock/morph17:26
bashrc_is there any parallelism in the way morph starts up?17:27
ssam2it's all 1 thread17:30
bashrc_so, it looks like it is reading the plugin paths but just doesn't find firehose17:30
*** mdizzle has quit IRC17:46
radiofreehmm.. you end up with a merge commit for everything now
straycatradiofree, yes but we can have an irc bot!17:47
ssam2radiofree: we had merge commits for everything before Gerrit, because the policy was to merge with 'git merge --no-ff'. Currently the Gerrit merge policy for all repos is set to 'merge-if-necessary', so we don't get a merge commit unless one is needed17:54
*** tiagogomes_ has quit IRC17:55
ssam2we could set the policy to rebase-before-merge to reduce the number of merge commits17:55
ssam2in fact, I meant to do that already on jjardon's suggestion but then didn't17:55
ssam2i'll do it tomorrow17:55
radiofreessam2: i mean *everything* in the patch series17:55
ssam2oh, because each commit is separate17:55
radiofreethere's "Merge "Update to qt5 to v5.4.0" and Merge "Fix build of ICU on Arm"17:55
ssam2yes, that's a bit ugly17:56
radiofreethat's going to look horrendous for a large patch series17:56
straycati think you're missing the bigger picture17:56
straycatwe can finally have an irc bot17:56
bashrc_if gerrit is the future then the existing niggles in it could be patched17:56
ssam2radiofree: like I said, I'll change the merge policy so that Gerrit rebases each commit before merging17:56
radiofreestraycat: :-o17:56
* radiofree drops all objectsion then17:56
ssam2there are only merge commits for those commits because they were not based off latest master at the time they were merged17:57
ssam2but gerrit has about 10 different merge policies, I think rebase before merge will solve that17:57
ssam2but, tomorrow, i'm off now17:57
*** ssam2 has quit IRC17:57
*** bashrc_ has quit IRC18:00
*** cyndis has quit IRC18:06
*** edcragg has quit IRC18:07
*** rdale has quit IRC18:12
*** cyndis has joined #baserock18:20
*** mdunford has quit IRC18:25
straycatpersia, right so it seems no one wants a 0.x branch18:29
* straycat shrugs18:29
*** zoli__ has joined #baserock18:30
*** wschaller_ has quit IRC18:50
straycathmm there might be a hack for gerrit that allows you to push a series as one change19:03
straycatapparently you can perform the merge locally then push the merge commit19:03
straycati also didn't realise but the reviews seem to be per change, so you can't actually give a review on a series as a whole, just the individual commits19:04
*** zoli__ has quit IRC21:20
*** gary_perkins has quit IRC21:43
*** gfinney has quit IRC22:18

Generated by 2.14.0 by Marius Gedminas - find it at!