On Tue, Jul 14, 2015 at 09:44PM, Olaf Flebbe wrote:
> Cos,
>
> As far as I understand a release is done, when a tag like release-1.0 is
> set, not when a branch is created. Since there is no release-1.0 tag, a
> release is not done. And the relase has still has no SHA, since we have no
> tag associated.
I didn't bring up the release or their tags anywhere in my email, so I am not
sure what's your point here... Besides, with git we aren't relying on release
tags as they are mutable; instead we are relying on commit SHA1.
> The merging of hotfix branches (suggested in the paper you mentioning) will
> do more harm to the log than a cherry-picking: On the release branch you
> will get two commits: The fix itself and a merge commit. With cherry-picking
> you will get only the fix itself.
>
> The hotfix branch does make sense when you'll need more than one commit to
> fix a issue on a release. But since we use the one-JIRA-one-Commit model, it
> does not make sense to create a branch for every commit needed on a release
> branch. We would be polluting the repository with stale branches, since
> we'll have to push each commit as a branch to the official repository. We
> cannot remove them later, even when merged completly. So I would say -1 to
> hotfix branches in Bigtop.
Branches are cheap and easy to clean - once the merge is done the branch will
need to be deleted. 2nd merge you're referring to is a non fast-forward
commit, which is annoying in a sense, but also helps to track the history of
the changes and where there are coming from. And yes - cherry-picking is
absolutely worst because it changes a commit's SHA1 and makes it impossible
for automatic tooling to trace the flow of the changes.
Cos
> Cherry-picking is valid if you'll need to fix releases through small changes
> (but only then). The fixes on bigtop where small....
>
> Olaf
>
>
>
>
> > Am 14.07.2015 um 21:15 schrieb Konstantin Boudnik <[email protected]>:
> >
> > I have committed BIGTOP-1923 to the master and cherry-picked it to the
> > branch.
> >
> > The reason I did the cherry-picking is because there was a number of other
> > commits cherry-picked before me. I think cherry-picking is an ugly way of
> > doing SVN-like commit merges. And we should avoid it if possible.
> > Definitely,
> > in case of release branches it should be doable via git branching model
> > like...
> >
> > http://nvie.com/posts/a-successful-git-branching-model/
> >
> > which preserves the history and commit SHAs. Can we please avoid
> > cherry-picking on release branches in the future? Please....
> >
> > Thanks
> > Cos
> >
> > On Thu, Jul 09, 2015 at 03:59PM, Evans Ye wrote:
> >> I have no concern if we can get the ignite 1.2 patch in quickly . :)
> >>
> >> 2015-07-09 7:45 GMT+08:00 Konstantin Boudnik <[email protected]>:
> >>
> >>> BIGTOP-1907 seems to be in. One last question - any objection of bumping
> >>> up
> >>> Ignite to 1.2 which got released just a couple of weeks ago? It has good
> >>> fixes
> >>> (what I've heard ;)
> >>>
> >>> Thoughts?
> >>>
> >>> Cos
> >>>
> >>> On Thu, Jul 02, 2015 at 12:07PM, Evans Ye wrote:
> >>>> Sorry to chime in late.
> >>>> Bigtop 1907 should be in so that bigtop toolchain can successfully
> >>> applied
> >>>> on openSUSE. The things I'm still concerning about is that we do not know
> >>>> whether openSUSE is OK to build Bigtop components. IIRC it have some
> >>>> problems.
> >>>> However, if we treat the words "supporting OSs" as OSs that can install
> >>> and
> >>>> run Bigtop instead of build. Then we should be ok now. I personally
> >>> prefer
> >>>> the latter unless SUSE expert can help to fix and unlock the release. :)
> >>>> 2015/7/2 上午6:50 於 "Andrew Purtell" <[email protected]> 寫道：
> >>>>
> >>>>> +1
> >>>>>
> >>>>>
> >>>>> On Wed, Jul 1, 2015 at 11:54 AM, Konstantin Boudnik <[email protected]>
> >>>>> wrote:
> >>>>>
> >>>>>> Agree. It's super easy to push site updates any time after the
> >>> release is
> >>>>>> out.
> >>>>>> We are generating it our of the master anyway, so I see no reason to
> >>> hold
> >>>>>> 1.0
> >>>>>> because of the website changes.
> >>>>>>
> >>>>>> So, shall I spin-out RC1 then?
> >>>>>> Cos
> >>>>>>
> >>>>>> On Wed, Jul 01, 2015 at 11:17AM, Andrew Purtell wrote:
> >>>>>>> We can fix the site after the release. I agree it will be
> >>> important to
> >>>>>> get
> >>>>>>> the content up to date as close as possible to when the release
> >>> goes
> >>>>> out.
> >>>>>>> Are we doing svnpubsub? If so it should be straightforward to
> >>> update.
> >>>>> One
> >>>>>>> of us (maybe me) can regen while RC voting is underway and push
> >>> upon
> >>>>>>> release.
> >>>>>>>
> >>>>>>>
> >>>>>>> On Tue, Jun 30, 2015 at 10:53 PM, Olaf Flebbe <[email protected]>
> >>> wrote:
> >>>>>>>
> >>>>>>>> hi cos,
> >>>>>>>>
> >>>>>>>> thanks for asking.
> >>>>>>>>
> >>>>>>>> two of the new commits were critical since last minute changes
> >>> broke
> >>>>>>>> compilation.
> >>>>>>>>
> >>>>>>>> I see no further technical Jiras waiting for inclusion into a 1.0
> >>>>> tag.
> >>>>>>>>
> >>>>>>>> But I am missing one „huge“ commit:
> >>>>>>>>
> >>>>>>>> Our web site is essentially a heap of obsolete stuff. Since the
> >>>>>> website is
> >>>>>>>> somehow generated from git, we should consolidate our site as
> >>> well
> >>>>>>>> reflecting the 1.0 release (for instance removing 0.8 information
> >>>>> from
> >>>>>> the
> >>>>>>>> frontpage). @all: Is there already any progress in this
> >>> direction?
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>> olaf
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>> Am 01.07.2015 um 02:16 schrieb Konstantin Boudnik <
> >>> [email protected]
> >>>>>> :
> >>>>>>>>>
> >>>>>>>>> Guys,
> >>>>>>>>>
> >>>>>>>>> I have noticed that branch-1.0 is now having these three
> >>> committs:
> >>>>>>>>>
> >>>>>>>>> 11dc343 BIGTOP-1893: Compilation of hadoop-yarn-client failed
> >>>>>>>>> edc881d BIGTOP-1902: typo in
> >>>>>>>> bigtop-deploy/vm/vagrant-puppet-vm/vagrantconfig.yaml
> >>>>>>>>> 885cd8f BIGTOP-1896. bigtop_toolchain broken bei ant update
> >>>>>>>>>
> >>>>>>>>> Now I would have to update CHANGES.txt and RELEASE notes for
> >>> both
> >>>>>> master
> >>>>>>>> and
> >>>>>>>>> the branch. To me, those weren't that critical to include into
> >>> the
> >>>>>>>> release,
> >>>>>>>>> but I guess it makes no harm. Any other changes we have to
> >>> hold the
> >>>>>>>> release
> >>>>>>>>> for?
> >>>>>>>>>
> >>>>>>>>> Thanks,
> >>>>>>>>> Cos
> >>>>>>>>>
> >>>>>>>>>> On Sat, May 30, 2015 at 10:14AM, Konstantin Boudnik wrote:
> >>>>>>>>>> Ok, master is unlocked for 1.1.0-SNAPSHOT development
> >>>>>>>>>> I have also pushed branch-1.0 that has all the bits for 1.0
> >>> RC,
> >>>>> but
> >>>>>> I
> >>>>>>>> don't
> >>>>>>>>>> have time to finish RC publishing up right now - will try to
> >>> do it
> >>>>>> from
> >>>>>>>> the
> >>>>>>>>>> train, but who knows if I will have any connection there.
> >>>>>>>>>>
> >>>>>>>>>> If anyone can pick up where I left-off - it'd be great: I am
> >>>>> really
> >>>>>>>> trying to
> >>>>>>>>>> get on my damn vacation ;) If not - I will try to find a bit
> >>> of
> >>>>> time
> >>>>>>>> next week
> >>>>>>>>>> for this.
> >>>>>>>>>>
> >>>>>>>>>> Thanks all for your help,
> >>>>>>>>>> Cos
> >>>>>>>>>>
> >>>>>>>>>>> On Sat, May 30, 2015 at 09:56AM, Konstantin Boudnik wrote:
> >>>>>>>>>>> Indeed.
> >>>>>>>>>>>
> >>>>>>>>>>> The one hurdle I am facing though is that with BIGTOP-1833
> >>>>> getting
> >>>>>>>> pushed to
> >>>>>>>>>>> the master (which I normally would appreciate ;) I have to
> >>> fiddle
> >>>>>> with
> >>>>>>>> my
> >>>>>>>>>>> local branches as they already had a commit for this fix.
> >>> Now,
> >>>>>>>> rebasing it
> >>>>>>>>>>> locally presents an issue where I need to fix some unpleasant
> >>>>>>>> conflicts in...
> >>>>>>>>>>>
> >>>>>>>>>>> Oh well, it will be done in a bit.
> >>>>>>>>>>> Cos
> >>>>>>>>>>>
> >>>>>>>>>>>> On Fri, May 29, 2015 at 09:37PM, Evans Ye wrote:
> >>>>>>>>>>>> I've committed the BIGTOP-1833 patch.
> >>>>>>>>>>>> It's so exciting that we're at the moment ready to release
> >>>>> bigtop
> >>>>>> 1.0.
> >>>>>>>>>>>> Please ping me if any help needed for release.
> >>>>>>>>>>>>
> >>>>>>>>>>>> Evans
> >>>>>>>>>>>>
> >>>>>>>>>>>>
> >>>>>>>>>>>> 2015-05-29 17:52 GMT+08:00 Konstantin Boudnik <
> >>> [email protected]>:
> >>>>>>>>>>>>
> >>>>>>>>>>>>> Guys,
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> I have made all preparations for 1.0 RC and changed
> >>> everything
> >>>>> we
> >>>>>>>> need to
> >>>>>>>>>>>>> move
> >>>>>>>>>>>>> to 1.1.0-SNAPSHOT in the master.
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> However, because BIGTOP-1833 hasn't been reviewed yet, I
> >>>>>> evidently
> >>>>>>>> haven't
> >>>>>>>>>>>>> pushed this change and other commits to the master nor
> >>>>>> branch-1.0. I
> >>>>>>>> would
> >>>>>>>>>>>>> unblock me immediately if someone can review BIGTOP-1833
> >>>>> changes.
> >>>>>>>> Then
> >>>>>>>>>>>>> patch
> >>>>>>>>>>>>> is rather trivial and I have tested it to make sure that
> >>> the
> >>>>>>>> packages are
> >>>>>>>>>>>>> produced.
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> I'd love to switch off to the vacation mode by the end of
> >>> the
> >>>>> day
> >>>>>>>> today but
> >>>>>>>>>>>>> have to wrap up the RC process (at least the branch
> >>> preparation
> >>>>>> part)
> >>>>>>>>>>>>> before
> >>>>>>>>>>>>> that. So if someone in Europe (during the day here) or
> >>> later in
> >>>>>> NA
> >>>>>>>> can do
> >>>>>>>>>>>>> this
> >>>>>>>>>>>>> at your early convenience - it'd be just great!
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> Thanks in advance!
> >>>>>>>>>>>>> Cos
> >>>>>>>>>>>>>
> >>>>>>>>>>>>>> On Tue, Apr 28, 2015 at 04:03PM, Konstantin Boudnik wrote:
> >>>>>>>>>>>>>>> On Tue, Apr 28, 2015 at 03:11PM, Andrew Purtell wrote:
> >>>>>>>>>>>>>>> Also, working on BIGTOP-1851 right now. Please pardon the
> >>>>>> delay, I
> >>>>>>>> was
> >>>>>>>>>>>>> sick
> >>>>>>>>>>>>>>> over the weekend.
> >>>>>>>>>>>>>>
> >>>>>>>>>>>>>> Thanks Andrew - hopefully are doing better now!
> >>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> On Tue, Apr 28, 2015 at 3:09 PM, Andrew Purtell <
> >>>>>>>> [email protected]>
> >>>>>>>>>>>>> wrote:
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> Cool, I went over to BIGTOP-1615 and it's already
> >>> resolved.
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> As for BIGTOP-1827, if it's just a package naming
> >>> problem we
> >>>>>> can
> >>>>>>>> just
> >>>>>>>>>>>>>>>> rename it, right? See my comment on the issue.
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> On Tue, Apr 28, 2015 at 2:40 PM, Konstantin Boudnik <
> >>>>>>>> [email protected]>
> >>>>>>>>>>>>>>>> wrote:
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Guys
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> I think we are in the good shape to cut-off 1.0 branch
> >>>>>> tonight. I
> >>>>>>>>>>>>> have ran
> >>>>>>>>>>>>>>>>> full stack build on Ubuntu and everything is building
> >>> fine
> >>>>>> now.
> >>>>>>>> Of
> >>>>>>>>>>>>> course
> >>>>>>>>>>>>>>>>> some
> >>>>>>>>>>>>>>>>> testing needs to be done ;)
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> One last unresolved blocker is BIGTOP-1827. If I don't
> >>> hear
> >>>>>> from
> >>>>>>>> the
> >>>>>>>>>>>>>>>>> Tachyon component's maintainers I will have to remove
> >>> this
> >>>>>> from
> >>>>>>>> the
> >>>>>>>>>>>>>>>>> release
> >>>>>>>>>>>>>>>>> 1.0 because the packages are broken and we can release
> >>> them
> >>>>>> like
> >>>>>>>>>>>>> this.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> BIGTOP-1615 (another blocker) is in PA state, so if
> >>> anyone
> >>>>>> can
> >>>>>>>> take
> >>>>>>>>>>>>> a
> >>>>>>>>>>>>>>>>> look -
> >>>>>>>>>>>>>>>>> it'd be great!
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Thanks!
> >>>>>>>>>>>>>>>>> Cos
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> --
> >>>>>>>>>>>>>>>> Best regards,
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> - Andy
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> Problems worthy of attack prove their worth by hitting
> >>>>> back. -
> >>>>>>>> Piet
> >>>>>>>>>>>>> Hein
> >>>>>>>>>>>>>>>> (via Tom White)
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> --
> >>>>>>>>>>>>>>> Best regards,
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> - Andy
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> Problems worthy of attack prove their worth by hitting
> >>> back.
> >>>>> -
> >>>>>> Piet
> >>>>>>>>>>>>> Hein
> >>>>>>>>>>>>>>> (via Tom White)
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> --
> >>>>>>> Best regards,
> >>>>>>>
> >>>>>>> - Andy
> >>>>>>>
> >>>>>>> Problems worthy of attack prove their worth by hitting back. - Piet
> >>>>> Hein
> >>>>>>> (via Tom White)
> >>>>>>
> >>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Best regards,
> >>>>>
> >>>>> - Andy
> >>>>>
> >>>>> Problems worthy of attack prove their worth by hitting back. - Piet
> >>> Hein
> >>>>> (via Tom White)
> >>>>>
> >>>
>
