Displaying #maven-dev/2017-01-02.log:

Mon Jan 2 09:26:58 2017  mbenson_:Joined the channel
Mon Jan 2 14:44:45 2017  tibor_:Joined the channel
Mon Jan 2 18:01:47 2017  tibor_:Joined the channel
Mon Jan 2 18:08:29 2017  hboutemy:Joined the channel
Mon Jan 2 18:08:50 2017  tibor_:@stephenc: Hi Stephen
Mon Jan 2 18:09:14 2017  tibor_:@hboutemy: Hi Herve
Mon Jan 2 18:24:39 2017  hboutemy:tibor_: Hi Tibor
Mon Jan 2 18:25:01 2017  hboutemy:I'll answer tonight
Mon Jan 2 18:25:26 2017  hboutemy:I'm quite ok with you, sadly...
Mon Jan 2 18:27:15 2017  hboutemy:but let's start more positively: Happy New Year
Mon Jan 2 18:27:25 2017  hboutemy::/
Mon Jan 2 18:47:59 2017  rfscholte:Joined the channel
Mon Jan 2 19:31:54 2017  stephenc:@rfschole @hboutemy @mbension_ @tibor_ have we people here?
Mon Jan 2 19:33:20 2017  stephenc:@ASFBot help
Mon Jan 2 19:33:30 2017  stephenc:ASFBot: help
Mon Jan 2 19:34:11 2017  hboutemy:stephenc: Hi Stephen
Mon Jan 2 19:34:17 2017  stephenc:aha!
Mon Jan 2 19:34:25 2017  stephenc:online at the same time... hurrah!
Mon Jan 2 19:35:12 2017  hboutemy:yes, it has been hard to find :)
Mon Jan 2 19:35:38 2017  stephenc:what a mess
Mon Jan 2 19:36:12 2017  hboutemy:yes :/
Mon Jan 2 19:36:21 2017  hboutemy:Christian does not learn
Mon Jan 2 19:36:24 2017  stephenc:so my plan is to act as a release manager for 3.5.0 and 3.5.1 (at least)
Mon Jan 2 19:36:52 2017  stephenc:well he was not the only one treating master as a personal development branch ;-)
Mon Jan 2 19:37:00 2017  hboutemy:notice that I think you talk about 3.5.1 when you should be speaking of 3.6.0
Mon Jan 2 19:37:18 2017  stephenc:why do you think it should be 3.6.0
Mon Jan 2 19:37:43 2017  stephenc:bug fixes and regressions do not a minor version bump make
Mon Jan 2 19:37:50 2017  hboutemy:because the first (good) reason you invoke to not put in 3.5.0 is that it is a change in behaviour
Mon Jan 2 19:38:05 2017  hboutemy:then such a change in behaviour must not go in 3.5.1
Mon Jan 2 19:38:09 2017  hboutemy:but in 3.6.0
Mon Jan 2 19:38:15 2017  stephenc:well the aim was that 3.5.0 should be a drop in replacement for 3.3.9
Mon Jan 2 19:38:21 2017  hboutemy:yes
Mon Jan 2 19:38:26 2017  hboutemy:then changes are for 3.6.0
Mon Jan 2 19:38:31 2017  stephenc:IOW swapping aether for resolver should not make a difference at all
Mon Jan 2 19:38:39 2017  hboutemy:3.5.1 is bugfixes over 3.5.0
Mon Jan 2 19:38:53 2017  stephenc:yes and the plugin classloader regression is a bugfix IMHO
Mon Jan 2 19:39:05 2017  stephenc:likewize the handling of zip files
Mon Jan 2 19:39:22 2017  hboutemy:there are a few cases where it can be discussed
Mon Jan 2 19:39:32 2017  stephenc:likewize MNG-6029
Mon Jan 2 19:39:36 2017  rfscholte:here is well
Mon Jan 2 19:39:42 2017  stephenc:those, to me, are all bugs
Mon Jan 2 19:39:46 2017  hboutemy:zip files MNG-5567 : I don't see much harm in this change
Mon Jan 2 19:40:21 2017  stephenc:hervé: I have only said for that one that I will not second it... I do not object... if somebody else says let's go woth that I'm fine
Mon Jan 2 19:40:29 2017  hboutemy:MNG-6029 is a bug at the origin, but we must master the impact of the fix
Mon Jan 2 19:40:54 2017  stephenc:yes MNG-6029 is something where we need to understand the impact
Mon Jan 2 19:41:09 2017  stephenc:but we will lose trust of the users if we keep jumping minor versions too
Mon Jan 2 19:41:10 2017  hboutemy:just to avoid that a "logical" fix has more impact than expected
Mon Jan 2 19:41:33 2017  stephenc:if we have a cadence of releases, users will forgive us more
Mon Jan 2 19:41:40 2017  hboutemy:sure
Mon Jan 2 19:41:53 2017  stephenc:I am fed up waiting for JvZ to actually deliver his promise of a cadence
Mon Jan 2 19:42:17 2017  stephenc:and it has reached "stephen will step up" levels of annoyance on my part
Mon Jan 2 19:42:27 2017  hboutemy:thanks :)
Mon Jan 2 19:42:48 2017  stephenc:so, as (provisional) release manager, I have my plan for the releases
Mon Jan 2 19:42:52 2017  hboutemy:in fact, the hard part on being release manager for Maven core is not about the process: it's the same like anything
Mon Jan 2 19:43:10 2017  hboutemy:but it's about figuring out feedback
Mon Jan 2 19:43:20 2017  stephenc:my proposed plan is 3.5.0 should not change dependency or classpath building in any way from 3.3.9
Mon Jan 2 19:43:45 2017  hboutemy:for a very long time, I personnally don't understand the impact of a lot of subtle changes, or some feedbacks
Mon Jan 2 19:44:00 2017  stephenc:bugs / regressions in dependency / classpath building are fine IMHO for 3.5.1 (if we understand the risk / impact)
Mon Jan 2 19:44:08 2017  rfscholte:stephenc: with this subset of changes, do you want to push RC's?
Mon Jan 2 19:44:24 2017  stephenc:robert: I will do one RC at least
Mon Jan 2 19:44:32 2017  stephenc:soak the RC for one week
Mon Jan 2 19:44:42 2017  rfscholte:sounds good to me
Mon Jan 2 19:44:43 2017  stephenc:respin during the week if there are reported issues
Mon Jan 2 19:45:02 2017  stephenc:after a clear week without reported regressions on 3.3.9 I will then cut the release
Mon Jan 2 19:46:04 2017  stephenc:for 3.5.1, if people really feel strongly about getting bugs fixed, I am happy to start a 3.5.1 a week after 3.5.0
Mon Jan 2 19:46:25 2017  stephenc:if there is consensus that it should be called 3.6.0, I will follow the consensus
Mon Jan 2 19:46:47 2017  stephenc:but, to me, unless we are adding new APIs... it's just a patch version bump
Mon Jan 2 19:47:15 2017  stephenc:if we are changing existing APIs in a breaking way... it's a major bump
Mon Jan 2 19:47:50 2017  rfscholte:Is ANSI marked for 3.5.0? I've already showed it on several conferences with positive feedback. I haven't had any issue with it so far.
Mon Jan 2 19:48:35 2017  stephenc:I am not stepping up to second ANSI... I am not objecting to ANSI... if one of you two wants to second ANSI for 3.5.0 then - unless someone else objects - it's in
Mon Jan 2 19:49:10 2017  stephenc:My idea for the https://cwiki.apache.org/confluence/display/MAVEN/Roadmap+2017 is to have an agreed set of changes to be in scope for the release
Mon Jan 2 19:49:55 2017  stephenc:that way we can prevent some people - who's names rhyme with Shistrian - from doing a solo run again
Mon Jan 2 19:50:17 2017  stephenc:at least that is my hope
Mon Jan 2 19:50:23 2017  stephenc:we will see if it pans out
Mon Jan 2 19:50:36 2017  rfscholte:Is there no MNG for the Aether 2 Maven Artifact Resolver replacement?
Mon Jan 2 19:50:47 2017  stephenc:I couldn't find one
Mon Jan 2 19:50:59 2017  stephenc:hard to even identify the changes in the commit diff
Mon Jan 2 19:51:15 2017  stephenc:plus what the F is in resolver... have we even got a baseline release
Mon Jan 2 19:51:40 2017  stephenc:I wonder would it make more sense to move resolver into maven-core's git repo
Mon Jan 2 19:52:09 2017  stephenc:(I don't want to delay 3.5.0 while we wait for that though)
Mon Jan 2 19:52:40 2017  rfscholte:IIRC Hervé was preparing resolver by updating poms, licenses, etc. Then suddenly bugfixes were added, so missed the release opportunity
Mon Jan 2 19:53:01 2017  stephenc:ok, so we probably need to reset resolver as well
Mon Jan 2 19:53:30 2017  stephenc:0dc0ccaaddce7d80934191e8ea09d70ec8eec9a3 is likely the commit we want to release from
Mon Jan 2 19:53:55 2017  stephenc:though that includes some other changes
Mon Jan 2 19:55:04 2017  stephenc:would it be better to cut a release from b7fc57c7e108fb0fa55710834abd59ff9d423e3d with hervé's changes rebased on top and then replay the rest of bentmann's changes
Mon Jan 2 19:55:56 2017  rfscholte:0dc0ccaaddce7d80934191e8ea09d70ec8eec9a3 contains bugfixes, that's not the right one
Mon Jan 2 19:56:06 2017  stephenc:ugh! what happened to the tag for 1.0.2.v20150114
Mon Jan 2 19:56:18 2017  hboutemy:JAnsi was proposed by Guillaume and seconded by me
Mon Jan 2 19:56:37 2017  hboutemy:on the dev list
Mon Jan 2 19:56:39 2017  stephenc:https://github.com/apache/maven-resolver/commits/1.0.x aha
Mon Jan 2 19:57:03 2017  stephenc:@hervé: sorry didn't see that as a seconding... seemed more like a clarification mail
Mon Jan 2 19:57:20 2017  hboutemy:though +1 was equivalent to seconding
Mon Jan 2 19:57:35 2017  stephenc:I may have mis-read... lot of balls in the air
Mon Jan 2 19:57:56 2017  hboutemy:and I'm not clear on what time we let to eventually have some -1 before considering ok
Mon Jan 2 19:58:00 2017  stephenc:hervé: what do we want to do about resolver
Mon Jan 2 19:58:10 2017  hboutemy:IMHO, what will be important will be a vote on the result
Mon Jan 2 19:58:30 2017  hboutemy:I think we should start from 1.0.x branch
Mon Jan 2 19:58:33 2017  stephenc:hervé I am leaving 3 days after a seconding before we say the item is "safe"
Mon Jan 2 19:58:58 2017  stephenc:hervé ok, so we need to replay your changes onto https://github.com/apache/maven-resolver/commits/1.0.x
Mon Jan 2 19:59:19 2017  hboutemy:ok, waiting for 2 more days (Robert, don't hesitate to second also in between :) )
Mon Jan 2 19:59:35 2017  hboutemy:yes
Mon Jan 2 19:59:39 2017  stephenc:(once an item is "safe" people can still object, but somebody 'may' have started work on it already)
Mon Jan 2 19:59:54 2017  hboutemy:I'll do the commits
Mon Jan 2 20:00:01 2017  stephenc:though no work should be started in any case until after the reset
Mon Jan 2 20:00:26 2017  stephenc:hervé seems fine to start the work on https://github.com/apache/maven-resolver/commits/1.0.x now if you are happy to though
Mon Jan 2 20:00:57 2017  stephenc:that would mean that the first release of resolver would be, what 1.0.3?
Mon Jan 2 20:01:05 2017  rfscholte:I will go through the complete list
Mon Jan 2 20:01:37 2017  stephenc:robert: my idea was to try and find as quick a way as possible to get agreement on scope
Mon Jan 2 20:01:44 2017  hboutemy:stephenc: the 3 days period is what you mean by "(until 2017-Jan-04)" in the Wiki page for the "Agreed changes" section?
Mon Jan 2 20:02:05 2017  stephenc:hervé yes, after 3 days I will change the (?) to a :+1:
Mon Jan 2 20:02:26 2017  hboutemy:ok, I didn't understand...
Mon Jan 2 20:02:29 2017  hboutemy:now I get it
Mon Jan 2 20:02:49 2017  rfscholte:Is MNG-6007 the Aether to MA Resolver?
Mon Jan 2 20:02:52 2017  stephenc:I'm trying to stage the changes to JIRA ahead of the reset vote
Mon Jan 2 20:03:19 2017  hboutemy:MNG-6110
Mon Jan 2 20:04:06 2017  rfscholte:right, let me add that to the wiki
Mon Jan 2 20:04:14 2017  hboutemy: MNG-6110
Mon Jan 2 20:04:14 2017  hboutemy:Upgrade Aether to Maven Resolver 1.2
Mon Jan 2 20:04:32 2017  hboutemy:currently in the "Unscrubbed" section
Mon Jan 2 20:05:35 2017  hboutemy:stephenc: if we want to be sure, we need to avoid changes that were in Eclipse Aether unreleased
Mon Jan 2 20:05:50 2017  hboutemy:then for clarity, maven-resolver 1.0.3 would be more explicit
Mon Jan 2 20:06:02 2017  stephenc:hervé, yes I think we can update the jira title with the correct version
Mon Jan 2 20:06:17 2017  stephenc:I've moved it out of unscrubbed... needing a seconder!
Mon Jan 2 20:06:50 2017  hboutemy:and to finish the mess, Jason released 1.1.0 froma personal branch not in Eclipse git, and with a controversial commit by Christian...
Mon Jan 2 20:08:16 2017  stephenc:OK, I've moved ANSI in scope for 3.5.0
Mon Jan 2 20:08:37 2017  hboutemy:and before this last controversial commit, the changes were done by Benjamin, but I don't know why he didn't make it in 1.0.x
Mon Jan 2 20:09:14 2017  hboutemy:for sure, I trust tnat when Benjamin put something in 1.1.x instead of 1.0.x, he knew what he did...
Mon Jan 2 20:09:30 2017  stephenc:hervé for sure this will be a lot of fun to untangle
Mon Jan 2 20:10:06 2017  stephenc:do we agree that it seems we have consensus on doing a reset on core & integration-tests ? €hboutemy @rfscholte
Mon Jan 2 20:10:17 2017  hboutemy:yes
Mon Jan 2 20:10:19 2017  rfscholte:yes
Mon Jan 2 20:10:59 2017  stephenc:do you think we need to reset resolver as well, or are we good releasing on the 1.0.x branch?
Mon Jan 2 20:11:52 2017  rfscholte:I'd prefer resetting, just to ensure those changes don't get released by accident
Mon Jan 2 20:12:17 2017  stephenc:@rfscholte: should we reset master then to the 1.0.x branch head?
Mon Jan 2 20:12:37 2017  hboutemy:yes
Mon Jan 2 20:12:51 2017  rfscholte:I agree
Mon Jan 2 20:13:03 2017  stephenc:OK, I'll put that into the vote
Mon Jan 2 20:13:24 2017  stephenc:I'm going to post a draft of the vote into the discuss thread tomorrow
Mon Jan 2 20:13:39 2017  stephenc:and then unless there are objections I will call the vote on Wednesday
Mon Jan 2 20:13:59 2017  stephenc:by then we should have a good idea roughly what will be going into the 3.5.0 release
Mon Jan 2 20:14:31 2017  stephenc:hopefully I will have figured out how to get the Jenkinsfile building the integration tests on ubuntu & windows with both Java 7 & 8
Mon Jan 2 20:14:47 2017  stephenc:and then we should be able to get a clean run of the integration tests
Mon Jan 2 20:15:08 2017  hboutemy:idea
Mon Jan 2 20:15:21 2017  hboutemy:before resetting, why not work on a branch?
Mon Jan 2 20:15:45 2017  hboutemy:and reset only when we are ok with the branch
Mon Jan 2 20:16:03 2017  stephenc:no issues with that. in fact once I call the vote I intend to start doing just that so that the reset is to the first commit on the branch
Mon Jan 2 20:16:20 2017  stephenc:that will stop anyone accidentally pushing to master after the reset
Mon Jan 2 20:16:51 2017  hboutemy:we'll need to make clear who commits what
Mon Jan 2 20:17:51 2017  hboutemy:like: the one who is in charge of a Jira issue is the one who is expected to commit (and eventually squash)
Mon Jan 2 20:18:12 2017  stephenc:yep... do you want to add your suggestions on that regard to the discuss thread
Mon Jan 2 20:18:43 2017  hboutemy:the discuss thread is already hard to read with seconds
Mon Jan 2 20:19:04 2017  hboutemy:the idea on who commits on the new master is just natural
Mon Jan 2 20:19:21 2017  hboutemy:the guy who did the work in the previous master is the guy who masters the change
Mon Jan 2 20:19:46 2017  stephenc:yep
Mon Jan 2 20:19:50 2017  stephenc:+1 on that
Mon Jan 2 20:20:20 2017  stephenc:with the proviso that one of us three step up if we don't have progress in time for the release
Mon Jan 2 20:20:24 2017  hboutemy:just natural => just in the vote to be explicit, since was implicit in your proposal
Mon Jan 2 20:20:47 2017  stephenc:ok, I was confusing with you suggesting something else
Mon Jan 2 20:20:56 2017  hboutemy:yes, once the natural job is done, we'll work on the less natural one :)
Mon Jan 2 20:21:21 2017  stephenc:ok. good to chat. hopefully we can find another time as we make progress
Mon Jan 2 20:21:28 2017  hboutemy:yes
Mon Jan 2 20:21:36 2017  stephenc:I need to spend some family time now for what remains of the evening
Mon Jan 2 20:21:40 2017  stephenc:ttyl
Mon Jan 2 20:21:45 2017  hboutemy:I tried to have Christian come to IRC, but failed
Mon Jan 2 20:21:51 2017  hboutemy:ok bye
Mon Jan 2 20:22:21 2017  rfscholte:there's a darts final with a dutchman, gonna watch that
Mon Jan 2 20:26:27 2017  hboutemy:darts?
Mon Jan 2 20:26:59 2017  hboutemy:ok, dutchman: local culture ;)
Mon Jan 2 20:31:25 2017  rfscholte::P
Mon Jan 2 20:32:03 2017  stephenc:ASFBot: pony
Mon Jan 2 20:32:25 2017  stephenc:ASFBot: pointer
Mon Jan 2 22:46:29 2017  hboutemy_:Joined the channel

Comments