All times shown according to UTC.

Time Nick Message
00:50 DOS622 joined #buildbot
01:01 bstack joined #buildbot
01:39 bb-github joined #buildbot
01:39 bb-github [13buildbot-infra] 15seankelly comment on pull request #37 146cec902: People would not be logged into the service account very often, right? `sh` on FreeBSD 10.0 or 10.1 supports tab completion, solving I feel the most important issue. Using `csh` is a non-starter and I don't like using something from ports just in case. 02http://git.io/VLwZEQ
01:39 bb-github left #buildbot
01:41 bb-github joined #buildbot
01:41 bb-github [13buildbot-infra] 15seankelly comment on pull request #37 146cec902: It appears it is:... 02http://git.io/YJ14jw
01:41 bb-github left #buildbot
01:44 bb-github joined #buildbot
01:44 bb-github [13buildbot-infra] 15seankelly comment on pull request #37 146cec902: Extra word in this sentence. 02http://git.io/wMtFnQ
01:44 bb-github left #buildbot
01:55 sboudr___ joined #buildbot
02:10 ahammond I have configured a Try_Userpass scheduler and attempted to run buildbot try against it. I get the following logs which suggest to me that CWD where I run try is important, but I don't know how to set it up correctly. Help please? https://gist.github.com/ahammo[…]b57d7fe774112f017
02:27 ewong ahammond: firstly, does /Users/ahammond/src/af/hss-unified  exist?
02:28 ahammond: secondly, does it need a / at the end?
02:29 ahammond ewong that directory does exist on my dev laptop where I'm running buildbot try, it does not exist on the master or any of the slaves.
02:30 (as you might imagine, it's the base directory of the hss-unified git checkout, which is what bb is building)
02:30 ewong ahammond: sorry, where is that 'git' command being run?
02:30 ahammond git clone --branch develop /Users/ahammond/src/af/hss-unified .
02:30 ewong ahammond: as in which machine?
02:31 ahammond I'm running try on my laptop and it is connecting to the master. I assume this is being run on a slave as part of the build process
02:32 ewong ahammond: so that log you pasted is from your laptop?
02:32 ahammond the log is from the web interface to the master
02:33 ewong the log shows the process being run on your laptop, right?   what about what's in your twistd.log on your laptop's buildbot slave process?
02:33 (perhaps I'm not understanding how you've got your build process set up)
02:34 ahammond my laptop does not have any slaves or other stuff on it. the master and slaves are hosted on an openstack cluster.
02:34 I'll see if I can repo this right now. just a sec
02:35 ewong sorry.. I'm a bit confused right now.
02:35 ahammond yup. I change into my local checkout of the repo we're building from,
02:35 run buildbot try --vc=git --connect=pb --master=192.168.6.152:8031 --username=... --passwd=...
02:35 it says ob has been delivered / not waiting for builds to finish
02:36 then I go to the web interface on the master and see the waterfall display
02:36 it shows a build that has failed on the first step, which is the Git() step
02:37 unlike the builds which are scheduled by the GitPoller, this one has a patch.
02:37 ewong ahammond: well, in the first part, there is an error:  "stat: No such file or directory: /home/buildbot/srv-hss-unified-srpms-t​o-rpms-to-depot/deb9b1bebe64000eaad32f​5dd283e2afc31a9a55/.buildbot-patched"
02:37 ahammond ewong oooooh, that's interesting. I'm guessing my working directory changed because of the patch?
02:38 ewong ahammond: I dunno.  but something in that path doesn't exist.
02:38 and if that |git| command is being run on an open cluster, it isn't going to find /Users... on your laptop
02:38 KeiKun joined #buildbot
02:38 ahammond the .buildbot-patched part makes perfect sense for not existing: I didn't create it
02:38 right, I wouldn't expect it to.
02:39 I would expect it to check out the same repo it's already working against
02:39 again, a working example would be a huge help here. I feel like I'm doing something really obviously wrong, just no idea what it is.
02:40 ewong ahammond: I'm kinda dense right now.  but where is this repo?
02:40 ahammond: sorry, I mean..  git is supposed to clone from where?
02:41 ahammond the repo is hosted by our enterprise stash server (think github, but not as nice)
02:41 I have cloned it down from there to my laptop,
02:41 ewong ahammond: right, but the buildbot slaves have no way of cloning from your laptop..
02:41 ahammond and of course the current, working master.cfg clones it from the stash server
02:41 bb-github joined #buildbot
02:41 bb-github [13buildbot] 15djmitche comment on issue #1424: My comment about a superset was regarding pre-processing the inputs to croniter.  If (as in `master` right now) we just take the inputs and pass them unchanged to croniter, then we can simply say "we take the same inputs as croniter".  If we pre-process, then we must have a more complicated "similar inputs to crontier except .., .., .." which adds complexity both for Buildbot's users and maintainers
02:41 bb-github left #buildbot
02:42 ewong ahammond: ok. the working master.cfg clones it from the stash server to where?
02:42 ahammond ewong to one of the buildslave servers.
02:43 whichever of the buildslaves happens to be running the build
02:43 if you're wondering how my laptop is involved at this point, well... so am I. :)
02:44 seems to me that the try command would send a diff, a current branch / revspec or whatever and that would be it.
02:45 ewong ahammond: I think I'm beginning to understand.
02:46 ahammond: your laptop sends a diff to the master.  the master in turns sends a command to the slave to start building the repo + diff.
02:47 ahammond ewong yes, I think that's the order of operations.
02:47 ewong ahammond: now the question remains.. why is it referring to /Users... in the buildstep
02:47 when it should be referring to where that stash is..
02:47 ahammond ewong, that's what I'd expect.
02:47 ewong ahammond: does the slave clone directly from the stash for every build?
02:48 ahammond ewong yes. Or at least there's a buildbot.stesp.source.git.Git() step as the first step in the build process
02:49 ewong ahammond: right.  then that step needs to be pointing to the actual repo and not your local laptop one
02:49 ahammond ewong that step _is_ pointing to the actual repo and always has been.
02:50 so, I already have a GitPoller scheduler in there that kicks off working builds. The whole thing works end to end. I'm just trying to add a Try_Userpass scheduler. :)
02:50 ewong ok. so that step cloned the stash to some dir on the slave.
02:50 ahammond ewong it appears to behave differently depending on what Scheduler calls it.
02:51 dang. I've gotta run. kids to be picked up and all that. Thanks for the help! I'll be on this again tomorrow.
02:52 mattiasr joined #buildbot
02:53 bb-github joined #buildbot
02:53 bb-github [13buildbot-infra] 15djmitche comment on pull request #37 14d565bdb: I changed the other two to `master` to match this. 02http://git.io/PR1Esg
02:53 bb-github left #buildbot
03:05 bb-github joined #buildbot
03:05 bb-github [13buildbot-infra] 15djmitche pushed 10 new commits to 06master: 02http://git.io/mnla1A
03:05 13buildbot-infra/06master 14e6e06da 15Dustin J. Mitchell: Run ansible in a crontask on service hosts...
03:05 13buildbot-infra/06master 14cc3b12b 15Dustin J. Mitchell: use a git:// URL
03:05 13buildbot-infra/06master 14efef58e 15Dustin J. Mitchell: Focus on local.yml for all Ansible runs
03:05 bb-github left #buildbot
03:16 djmitche uhoh: Checking integrity...Assertion failed: (pkgdb_ensure_loaded(j->db, p2, PKG_LOAD_FILES|PKG_LOAD_DIRS) == EPKG_OK), function pkg_conflicts_need_conflict, file pkg_jobs_conflicts.c, line 211.
03:17 that happened on my VM and I figured "meh" and blew it away, but this is on service1
03:18 booo joined #buildbot
03:29 bb-trac [trac] #3089/enhancement (new) updated by dustin (Deployed on service3. there are issues on service1 and service2.  I didn't try on ...) http://trac.buildbot.net/ticket/3089
04:03 bstack joined #buildbot
04:21 nobrowser joined #buildbot
04:27 DOS622 joined #buildbot
04:32 yokel joined #buildbot
04:34 jaredgrubb joined #buildbot
04:36 cdiehl______ joined #buildbot
04:38 lewq joined #buildbot
04:39 ggherdov joined #buildbot
04:53 mattiasr joined #buildbot
06:40 KeiKun_ joined #buildbot
07:12 mattiasr joined #buildbot
07:16 in3xes joined #buildbot
07:22 mattiasr joined #buildbot
07:59 bb-trac [trac] #3089/enhancement (new) updated by sa2ajj (vm1?) http://trac.buildbot.net/ticket/3089
08:07 KeiKun joined #buildbot
08:17 bb-github joined #buildbot
08:17 bb-github [13buildbot] 15benallard comment on issue #1432: Don't we want to move the callback at the same level as the codebase generator and at the same time enabling the functionality for **all** changesources ??? 02http://git.io/BiU1oQ
08:17 bb-github left #buildbot
08:30 jemand_pribyl joined #buildbot
08:33 bb-trac [trac] #3071/defect (new) updated by Ben (The Traceback changed due to modification of the code, but is still there:    {{{  ...) http://trac.buildbot.net/ticket/3071
08:58 Beirdo joined #buildbot
08:58 Beirdo joined #buildbot
09:03 bb-github joined #buildbot
09:03 bb-github [13buildbot] 15benallard opened pull request #1455: We want the first PrevSuccessfullBuild, not the last one ! (06master...06patch-1) 02http://git.io/7C9CMw
09:03 bb-github left #buildbot
09:06 bb-github joined #buildbot
09:06 bb-github [13buildbot] 15sa2ajj comment on pull request #1455 14d326aaf: I'm not sure about the logic here.... 02http://git.io/1mqNWw
09:06 bb-github left #buildbot
09:11 bb-trac [trac] #3071/defect (new) updated by Ben (GH:1455 is for the second case (comment 8)) http://trac.buildbot.net/ticket/3071
09:13 bb-github joined #buildbot
09:13 bb-github [13buildbot] 15delanne comment on issue #1455: :+1:  02http://git.io/K9Afhg
09:13 bb-github left #buildbot
09:16 bb-github joined #buildbot
09:16 bb-github [13buildbot] 15benallard comment on pull request #1455 14d326aaf: Not **that** loop, the inner one. That's exactly what we're fixing there. 02http://git.io/gu2BnQ
09:16 bb-github left #buildbot
09:30 lissyx joined #buildbot
09:34 bb-github joined #buildbot
09:34 bb-github [13buildbot] 15sa2ajj pushed 2 new commits to 06master: 02http://git.io/sNxypQ
09:34 13buildbot/06master 14d326aaf 15Benoît Allard: We want the first PrevSuccessfullBuild, not the last one !...
09:34 13buildbot/06master 14393a154 15Mikhail Sobolev: Merge pull request #1455 from benallard/patch-1...
09:34 bb-github left #buildbot
09:34 bb-github joined #buildbot
09:34 bb-github [13buildbot] 15delanne comment on issue #1432: if we want this functionality to all changesources, then we have to update all changesources (because it's not done in the BaseClass). Unfortunately, I don't have enought time to do that. 02http://git.io/yJgmuA
09:34 bb-github left #buildbot
09:35 mjw joined #buildbot
09:35 bb-github joined #buildbot
09:35 bb-github [13buildbot] 15delanne comment on issue #1432: Note: the code is not finished and not ready to be merged. 02http://git.io/Fk0ZeQ
09:35 bb-github left #buildbot
09:38 benallard delanne: (Further testing #3071) What should happen when the two builds found are triggered through a different codebase ? (comment 7 of that issue)
09:39 In those case, we created an (artificial) sourcestamp, but no changes belong to it.
09:39 (sourcestamps are always artificial anyway ...)
09:40 delanne no idea
09:42 benallard, I don't understand how you can have cb to None like in
09:42 {u'openvas-libraries': None, ....}
09:43 benallard to make a build, we need sourcestamps for each codebases.
09:43 delanne yes
09:43 benallard To trigger a build, a change from one of the codebases is anough
09:44 delanne so, we can have a sourcestamps which is not 'linked' to a change ?
09:44 benallard i.e. we are creating a sourcestamp for that one change we know, and we ned to 'invent' a sourcestamp for the change we didn't get.
09:44 that's it.
09:44 delanne damn !
09:44 benallard (that's where the codebase dictionnary of the schedulers is important)
09:45 delanne I still don't understand why I don't notice this with my configuration
09:45 I use 2 codebase too
09:45 benallard Look at a build where the triggering codebase changed.
09:46 (When you force a build, there are no linked changes either)
09:49 delanne sure
09:49 what I mean, is that I never noticed your traceback
09:51 benallard sa2ajj: Thanks for merging !
09:53 bb-github joined #buildbot
09:53 bb-github [13buildbot] 15benallard comment on issue #1432: I just mean: Why not implementing that in the base class ? 02http://git.io/N9oBOg
09:53 bb-github left #buildbot
10:14 benallard delanne: Maybe a solution would be to try to match the got_revision value ...
10:15 (instead of the parent_changeid)
10:18 That's not necessarily easier though. got_revision is a Property, Properties are not suposed to be stable ... It's mostly for display or for tweacking the build, but not meant to be used by buildbot internals ...
10:19 I just mean, I'd prefer not use it, unless we make it a row somewhere ...
10:21 Tomasu joined #buildbot
10:32 KeiKun_ joined #buildbot
10:37 bb-trac [trac] #3071/defect (new) updated by sa2ajj (Landed.) http://trac.buildbot.net/ticket/3071
11:13 lissyx joined #buildbot
11:15 Alina-malina joined #buildbot
11:18 MiniCow joined #buildbot
11:25 bb-github joined #buildbot
11:25 bb-github [13buildbot] 15gangefors comment on issue #1424: dayOfWeek is always processed before passing to croniter since buildbot uses mon=0 instead of sun=0. If buildbot had used mon=1 and sun=7 croniter would accept that wihtout processing any input from buildbot. Since buildbot have a different offset for dayOfWeek it needs processing, but this is done only if it's an integer (single or list of).... 02http://git.io/pBHmWQ
11:25 bb-github left #buildbot
12:08 KeiKun joined #buildbot
12:10 Varriount_ joined #buildbot
12:20 asherkin joined #buildbot
12:33 Nickeeh joined #buildbot
12:39 asherkin joined #buildbot
12:47 AlexanderS joined #buildbot
12:52 eXeC64 joined #buildbot
12:56 eXeC65 joined #buildbot
13:26 mjw joined #buildbot
14:06 lissyx joined #buildbot
14:08 steveeJ I've hit the bug http://trac.buildbot.net/ticket/3009 :/ and I'm not able to reproduce the workaround
14:09 during the build I create a directory called "output"
14:11 then I'm trying to upload all the files that are inside that directory
14:12 http://pastebin.com/Haig2Uya
14:13 infobob http://paste.pound-python.org/[…]S4oZsB8bMayRzfz0/ (repasted for steveeJ)
14:22 Amadiro left #buildbot
14:23 Jc2k_ joined #buildbot
14:24 bb-trac [trac] #3009/defect (new) updated by sa2ajj (Looks like we have another user with the same problem: ...) http://trac.buildbot.net/ticket/3009
14:25 [Arfreve1] joined #buildbot
14:25 yokel_ joined #buildbot
14:29 sa2ajj steveeJ: sorry, i made a note in the bug, but i can't help you :(
14:32 steveeJ sa2ajj: thank you though!
14:32 Alina-malina joined #buildbot
14:33 * sa2ajj is currently focused on bb infrastructure, so it's a bit difficult to dig in the code...
14:35 steveeJ since there is a posted workaround I must be doing that workaround wrong too
14:36 I understand that my build is happening in ${buildslave_basedir}/build/
14:36 so my output is really ${buildslave_basedir}/build/output
14:36 sa2ajj nope
14:37 ${buildslave_basir}/${builder-name}/build
14:37 jemand_pribyl sa2ajj: Error: "{buildslave_basir}/${builder-name}/build" is not a valid command.
14:37 sa2ajj in ${buildslave_basir}/${builder-name}/build
14:37 * sa2ajj is curious what that error was about :)
14:37 steveeJ which is really strange, since I have used `pwd` before to print out the full path and that didn't occur to me like that
14:38 sa2ajj that's interesting
14:38 steveeJ I'll give you output in a second
14:38 sa2ajj i'd be surprised if it's w/o ${builder_name} part
14:39 if it were, then running several builders on the same buildslave would cause a lot of mess...
14:39 steveeJ oh, you're right. the buildername is indeed in there
14:40 this is the property: upload_files: ['/var/lib/buildslave/rootfs_amd6​4/build/output/hello_world.txt', '/var/lib/buildslave/rootfs_​amd64/build/output/random1', '/var/lib/buildslave/rootfs_​amd64/build/output/random2']
14:41 the workaround suggests to "specify slavesrc paths relative to the builder's basedir"
14:42 * sa2ajj . o O (rootfs... so many old memories... hmm...)
14:42 sa2ajj then it should be something like 'rootfs_amd64/...'
14:42 if i understand the workaround suggestion right
14:43 steveeJ is there an environment variable specifiying the builders name?
14:43 s/fiying/fying
14:44 sa2ajj no, there's none
14:44 steveeJ or better a python variable that can be used in the step?
14:44 sa2ajj you can use a property though
14:44 http://docs.buildbot.net/curre[…].html#interpolate
14:46 steveeJ oh nice, there's a buildername property!
14:46 sa2ajj yes :)
14:48 steveeJ the property list does now contain: upload_files: ['rootfs_amd64/build/output/hello_world.txt', 'rootfs_amd64/build/output/random1', 'rootfs_amd64/build/output/random2']
14:49 still: http://paste.pound-python.org/[…]RiPxGGBRDJYa5sK1/
14:50 sa2ajj can you check the twistd.log at your buildslave?
14:50 you should see the exact path that causes this problem
14:51 steveeJ sh*t. I should have done that earlier :)
14:52 tatFile /var/lib/buildslave/rootfs_amd64/[ failed [Errno 2] No such file or directory: '/var/lib/buildslave/rootfs_amd64/['
14:52 it seems like the Interpolate which should render the list property to a list doesn't work as expected
14:55 interpolate is probably using the string and not a python list of strings
14:57 tromey joined #buildbot
14:59 steveeJ the remaining question is, how do I receive a python list of strings with Interpolate?
15:01 sa2ajj good question
15:02 immediate thing _to try_ is FlattenList
15:03 http://docs.buildbot.net/curre[…].html#flattenlist
15:03 though i'm not 100% sure that it'll produce what you expect
15:03 steveeJ I've seen that. but I do not have a list at all, but a string
15:04 sa2ajj well, your property is a list, no?
15:04 steveeJ I think as soon as this is figured out it should be added here: http://docs.buildbot.net/lates[…]opertyfromcommand
15:04 since this is where I grabbed the glob2list method
15:05 sa2ajj be careful w/ latest docs: it's for the master code
15:05 steveeJ the example shows how to produce a list property but doesn't how to use it ;)
15:05 sa2ajj yes, that's bad :(
15:07 i try to update the docs every time a small change makes a big difference, but usually it actually requires a bit bigger change :(
15:09 steveeJ sa2ajj: a native python way would involve regex to parse the string
15:10 re.findall("'.*'", s)
15:12 sa2ajj slightly different direction: is there any reason you cannot use http://docs.buildbot.net/lates[…]p-DirectoryUpload ?
15:13 you do not seem to filter any filenames
15:13 steveeJ good point. I need all the uploaded files in the json output of the build
15:14 sa2ajj right
15:17 mjw joined #buildbot
15:19 Alina-malina joined #buildbot
15:21 sa2ajj steveeJ: one question, how the property is shown in your build status page?
15:24 steveeJ upload_files: ['/var/lib/buildslave/rootfs_amd6​4/build/output/hello_world.txt', '/var/lib/buildslave/rootfs_​amd64/build/output/random1', '/var/lib/buildslave/rootfs_​amd64/build/output/random2']
15:25 this is copied from the "steps/setproperty/logs/property changes"
15:25 sa2ajj good
15:26 i think 'interpolate' was an overkill for your situation
15:26 steveeJ please enlighten me :D
15:26 sa2ajj can you try to just use slavesrcs=Property('upload_files') ?
15:27 KeiKun joined #buildbot
15:27 steveeJ of course
15:27 [Arfrever] joined #buildbot
15:27 sa2ajj (i'm not sure if it's an enlightenment of any kind though :))
15:29 steveeJ that's looking good!
15:30 sa2ajj good :)
15:30 steveeJ very good indeed! the absolute path is working too
15:30 thanks a lot
15:30 sa2ajj no worries :)
15:31 so i'll try to add an example there that uses a plain property :)
15:34 steveeJ that'd be nice
15:38 FWIW, using the absolute directory is working around the bug #3009 too
15:38 no need to specify an empty-string workdir
16:03 Abazigal joined #buildbot
16:08 Abazigal hi there; I'm considering buildbot for a replacement of the custom workflow I have, and there's smth I don't know how to do:
16:08 I have some GIT repos watched by the poller, and in a given folder in each repos, some .spec files
16:09 and I want to have one build job per spec file
16:10 the problem is that I can't know the list of specs (devs create it when they want) so I can't have a static list in my master.conf
16:12 sa2ajj sorry, i can't suggest anything right away
16:12 Abazigal so what I would like to have is a mechanism in the scheduler level which looks in the specs folder and dynamically creates the corresponding builds
16:12 sa2ajj a couple months back, one of the users managed to do some dynamic builders, however he did not publish results, so i can't point you to anything
16:14 maybe not dynamic builders as such, but some logic that'd produced a dynamic number of builds, iirc
16:14 Abazigal sa2ajj: ok, so, would you say that buidbot is not adapted to my need ? or do you think I will be able to achieve what I want (in a "quite nice" manner) with buildbot ?
16:16 sa2ajj buildbot is a _framework_, so it should be possible to implement it.
16:16 what i meant is that i can't point you to an example that would get you started in the right direction
16:18 Abazigal ok ok
16:19 anyway, thx for the answers
16:19 bb-trac [trac] #3113/support-request (v:) created by sa2ajj (ML configuration should somehow be cleaned up) http://trac.buildbot.net/ticket/3113
16:26 brson joined #buildbot
16:30 Alina-malina joined #buildbot
16:30 Alina-malina joined #buildbot
16:30 steveeJ Abazigal: what do these specs specify?
16:30 Abazigal steveeJ: RPM builds
16:30 sa2ajj creation of rpm files
16:33 how i understand the situation: a single change may result in building several packages
16:33 however it's not known at the time of configuration what to check
16:33 (what directories to check)
16:33 Abazigal yep and yep
16:34 well, I DO know that in /this/folder/specs, there will be spec files
16:34 in each of my repos
16:35 bb-trac [trac] #3089/enhancement (new) updated by dustin (service1 and service2 are now handled.    vm1 isn't in local.yml, so I haven't ...) http://trac.buildbot.net/ticket/3089
16:36 sa2ajj Abazigal: i'd suggest you write a mail to buildbot-devel
16:36 there's a larger number of users there and somebody might have done something similar
16:36 (and not all of them follow this channel :))
16:38 bb-trac [trac] #3089/enhancement (new) updated by dustin (https://github.com/buildbot/bu[…]bot-infra/pull/41 for that) http://trac.buildbot.net/ticket/3089
16:38 bb-github joined #buildbot
16:38 bb-github [13buildbot-infra] 15djmitche opened pull request #41: set up vm1 as a servicehost, too (06master...06add-vm1) 02http://git.io/uwv_sg
16:38 bb-github left #buildbot
16:39 bb-github joined #buildbot
16:39 bb-github [13buildbot-infra] 15sa2ajj closed pull request #41: set up vm1 as a servicehost, too (06master...06add-vm1) 02http://git.io/uwv_sg
16:39 bb-github left #buildbot
16:39 Abazigal sa2ajj: I will consider that; thx again
16:47 sa2ajj no worries
17:16 lissyx joined #buildbot
17:25 steveeJ Abazigal: if the existence of that file is known and the repositories are known, what's the unknown again?
17:28 _VV joined #buildbot
17:45 bb-github joined #buildbot
17:45 bb-github [13buildbot-infra] 15sa2ajj opened pull request #42: bump buildbot.net zone serial number (06master...06dns-serial-bump) 02http://git.io/4JAPqQ
17:45 bb-github left #buildbot
17:49 mattiasr joined #buildbot
18:03 sa2ajj djmitche: ayt?
18:23 djmitche sa2ajj: yep
18:23 now :)
18:24 sa2ajj i work on ns1 now
18:25 and it seems the current role is not really up to the task :)
18:25 localhost is not available in the jail and i tested it in my vm where it is
18:25 so when i finish today bind will be running with the correct zone (see pr above), but i'll also have to submit a pr to take that into account
18:26 nds is stopped and disabled there
18:26 jamey joined #buildbot
18:27 jamey I'm trying to use buildbot latest so I can use github authentication
18:28 but when I click on "login with GitHub", it starts the url with "/undefinedauth" instead of "/auth"
18:28 does this sound familiar to anyone?
18:28 KeiKun_ joined #buildbot
18:29 jamey I have not been able to build the frontend, which makes debug a bit harder
18:31 djmitche sa2ajj: sounds good
18:31 nothing's pointing to it for production yet
18:31 so no harm in an intermediate state
18:32 sa2ajj yes, that's what i thought
18:32 (re production :))
18:46 djmitche sa2ajj: did you deploy the crontask config to vm1 after merging PR41?
18:47 * djmitche does
18:47 bb-trac [trac] #3089/enhancement (new) updated by dustin (..and deployed to VM1.) http://trac.buildbot.net/ticket/3089
18:48 [trac] #3089/enhancement (closed) updated by dustin (empty comment) http://trac.buildbot.net/ticket/3089
18:48 sa2ajj yay
19:01 bstack joined #buildbot
19:01 bb-github joined #buildbot
19:01 bb-github [13buildbot-infra] 15sa2ajj pushed 1 new commit to 06master: 02http://git.io/2crSmQ
19:01 13buildbot-infra/06master 143afb82c 15Mikhail Sobolev: fix the dns role...
19:01 bb-github left #buildbot
19:02 * sa2ajj . o O (how did i went off my branch? :( )
19:02 bb-github joined #buildbot
19:02 bb-github [13buildbot-infra] 15sa2ajj 04force-pushed 06master from 143afb82c to 14224f0fe: 02http://git.io/gR8UtA
19:02 bb-github left #buildbot
19:03 djmitche heh
19:03 sa2ajj 5 second rule :)
19:03 djmitche patch looks good fwiw
19:03 sa2ajj needs testing
19:04 and something went wrong :(
19:05 wrong option
19:13 djmitche: for some reason ttl w/ bind is different from what i get from current ns
19:14 djmitche how so?
19:14 sa2ajj 655360 :(
19:14 and i can't find anywhere it being mentioned
19:15 (or i interpret results of 'dig' wrongly)
19:17 nevermind
19:17 wrong was the key
19:17 djmitche ok :)
19:18 sa2ajj final review and merge? :)
19:18 djmitche sure, link?
19:19 sa2ajj https://github.com/buildbot/bu[…]bot-infra/pull/42
19:19 you probably can re-configure (or check) he.net to use this primary
19:20 djmitche I think that will require some downtime
19:20 I'll need to do it when I have a few hours to clean up if there's a mess
19:20 -> not today
19:20 can you reassign the bug to me?
19:20 for that part?
19:21 bb-trac [trac] #3002/task (assigned) updated by sa2ajj (I have not re-created the jail yet (I still need to see how things work and probably ...) http://trac.buildbot.net/ticket/3002
19:21 sa2ajj sure
19:22 bb-trac [trac] #3002/task (assigned) updated by sa2ajj (As agreed with Dustin, he'll need some time for recongiguring 'he.net' DNS.) http://trac.buildbot.net/ticket/3002
19:26 tromey joined #buildbot
19:33 tromey joined #buildbot
19:34 bb-github joined #buildbot
19:34 bb-github [13buildbot-infra] 15sa2ajj closed pull request #42: bump buildbot.net zone serial number (06master...06dns-serial-bump) 02http://git.io/4JAPqQ
19:34 bb-github left #buildbot
19:34 1JTABEQ90 joined #buildbot
19:34 1JTABEQ90 [13buildbot-infra] 15sa2ajj pushed 4 new commits to 06master: 02http://git.io/qJCDeg
19:34 13buildbot-infra/06master 14764ade2 15Mikhail Sobolev: bump buildbot.net zone serial number...
19:34 13buildbot-infra/06master 147935152 15Mikhail Sobolev: apply base role to ns1...
19:34 13buildbot-infra/06master 14443538a 15Mikhail Sobolev: fix the dns role...
19:34 1JTABEQ90 left #buildbot
19:41 sa2ajj djmitche: is this ''sqlalchemy-migrate ==0.7.0, ==0.7.1, ==0.7.2'' supposed to work?
19:41 djmitche it does in my experience :)
19:42 sa2ajj then i do not quite understand what happens :(
19:42 maybe something like setuptools need to be upgraded?
19:42 djmitche is that not working somewhere?
19:43 sa2ajj nine
19:43 (i thought i said 'nine is broken' but i did not)
19:44 djmitche ah
19:45 sa2ajj and i found a problem in my bb-master role :(
19:46 djmitche http://nine.buildbot.net/#/bui[…]/step/1/log/stdio specifically?
19:46 sa2ajj https://github.com/buildbot/bu[…]bot-infra/pull/43
19:46 bb-github joined #buildbot
19:46 bb-github [13buildbot-infra] 15sa2ajj opened pull request #43: use the item value not key (06master...06fix-bb-master) 02http://git.io/7beu2A
19:46 bb-github left #buildbot
19:47 sa2ajj djmitche: that's something else
19:47 djmitche yeah
19:47 sa2ajj it's py2.4 incompatibility i believe
19:47 what i have is:
19:47 djmitche indeed
19:47 maybe we should drop that for 0.9.0
19:47 sa2ajj pkg_resources.DistributionNotFound: sqlalchemy-migrate==0.7.0,==0.7.1,==0.7.2
19:47 skelly slave compat?
19:47 KeiKun joined #buildbot
19:48 djmitche skelly: new virtualenv doesn't support python-2.4
19:48 sa2ajj and the fix for bb-master: https://github.com/buildbot/bu[…]bot-infra/pull/43
19:48 skelly ah yes, I found that out the hard part
19:48 djmitche pull 43 looks good
19:48 skelly er, the hard way
19:49 bb-github joined #buildbot
19:49 bb-github [13buildbot-infra] 15sa2ajj pushed 2 new commits to 06master: 02http://git.io/RN685A
19:49 13buildbot-infra/06master 1464c97a9 15Mikhail Sobolev: use the item value not key
19:49 13buildbot-infra/06master 14897b124 15Mikhail Sobolev: Merge pull request #43 from sa2ajj/fix-bb-master...
19:49 bb-github left #buildbot
19:49 djmitche sa2ajj: huh, I wonder if that's treated as "==0.7.0", as 0.7.0 is now gone
19:50 sa2ajj i'd expect that even if it's gone, it does not result in an exception
19:51 djmitche what's the context where you're seeing this errro?
19:52 sa2ajj slogin to s2.bb
19:52 sudo jexec 6 bash
19:53 su - bbuser
19:53 . venv/bin/activate
19:53 buildbot
19:53 djmitche huh
19:54 pip install -U setuptools?
19:54 setuptools does a great job of hiding its version numbers :(
19:54 and has several version numbers depending on how you look
19:54 sa2ajj is already up to date
19:54 djmitche hrm
19:54 so maybe that was *once* valid syntax and is no longer?
19:55 sa2ajj hmm
19:55 but how we manage to pass travis then?
19:55 it'd expect it choke on this
19:55 djmitche I don't know
19:55 maybe travis has an older setuptools?
19:56 sa2ajj that's possible
19:56 djmitche if you change hat to just say ==0.7.2, does it work?
19:56 sa2ajj trying
19:58 yes
20:00 djmitche ok, I'm more than happy to merge that fix
20:00 sa2ajj ok
20:03 bb-github joined #buildbot
20:03 bb-github [13buildbot] 15sa2ajj opened pull request #1456: simplify sqlalchemy requirement (06master...06sqlalchemy-alchemy) 02http://git.io/7t3cOA
20:03 bb-github left #buildbot
20:03 sa2ajj djmitche: ^^
20:03 bb-github joined #buildbot
20:03 bb-github [13buildbot] 15djmitche comment on issue #1456: good! 02http://git.io/j-r_AQ
20:03 bb-github left #buildbot
20:04 bb-github joined #buildbot
20:04 bb-github [13buildbot] 15sa2ajj comment on issue #1456: Let's wait for Travis though. 02http://git.io/QBWWuw
20:04 bb-github left #buildbot
20:05 sa2ajj i'll merge it tomorrow if nobody does
20:05 * sa2ajj is off for today
20:05 djmitche kk
20:16 jamey joined #buildbot
20:40 Arfrever joined #buildbot
20:51 jaredgrubb joined #buildbot
21:54 dancingPanda joined #buildbot
22:05 dancingPanda I want to give users the ability to specify if they wish to be notified of the build's status when they force build on a certain builder.  Does anyone know if there is an easy way to accomplish this using MailNotifier or something similar?
22:46 KeiKun_ joined #buildbot
22:49 sergiodj joined #buildbot
22:50 sergiodj hi there, again :)
22:51 in what situations should retryFetch trigger another fetch on the Git step?
22:52 because I am setting retryFetch to True, my fetch is (strangely, I admit) failing with a connection reset, but I am not seeing a retry hapenning
23:41 KeiKun joined #buildbot

← Previous day | Index | Server Index | Channel Index | Today | Next day → | Atom Feed | Search | Google Search | Plain-Text | plain, newest first