It’s really hard to follow when people comment on emails _and_ on Confluence. Maybe just my client but there is no way to tell which is which in email.
Back to 1.1.1 but I wonder if our artifacts would run on 1.3 (except for the Shell). The sequence file write problem was supposed to be binary compatible but obviously not source compatible with previous builds. BTW can we forever put to bed the question of how serious a Spark version change is? Be afraid! On Mar 25, 2015, at 2:28 PM, Andrew Palumbo <[email protected]> wrote: I know that there is a bug in 1.2 that causes problems with CF. The patch for 1.2/1.3 turns out to be compatible with 1.1.1. I'll put a PR up for it, since we probably will want to commit it to master anyways. I was hoping that it would be possible to create a maven profile for the spark version and not have to worry about the 1.3 Branch... I'll take another look at the shell later and see if its something that I'm missing (they everything that we override for the shell package-private btwn 1.2 and 1.3)... On 03/25/2015 05:00 PM, Andrew Musselman wrote: > +1 am happier with 1.2 anyway, is a bonus to have 1.3 working halfway. > > On Wed, Mar 25, 2015 at 1:44 PM, Andrew Palumbo <[email protected]> wrote: > >> I'm good with that. The patch that I have on this JIRA gets everything >> ready for 1.2/1.3 and tests for everything on 1.1.1. So we can make that >> available. >> >> But the shell may need alot of work. >> >> >> >> On 03/25/2015 04:27 PM, Suneel Marthi wrote: >> >>> +1 Dmitriy >>> >>> Sent from my iPhone >>> >>> On Mar 25, 2015, at 4:25 PM, Dmitriy Lyubimov <[email protected]> wrote: >>>> I'd venture a herecy again. what if we put off 1.3 compatibility until >>>> better times and focus on 0.9...1.2.x compatibility we have now. >>>> >>>> Chances are by the time we are done with 0.10.x we'd need to consider 1.4 >>>> or 1.5 at the rate this project is bloating. Keep in mind that higher >>>> version != higher maturity, as it has been demonstrated by almost any >>>> project in hadoop ecosystem, and it is being demonstrated again with >>>> newly >>>> introduced 1.2.x bugs. >>>> >>>> On Wed, Mar 25, 2015 at 1:17 PM, Andrew Palumbo (JIRA) <[email protected]> >>>> wrote: >>>> >>>> [ >>>>> https://issues.apache.org/jira/browse/MAHOUT-1653?page= >>>>> com.atlassian.jira.plugin.system.issuetabpanels:comment- >>>>> tabpanel&focusedCommentId=14380674#comment-14380674 >>>>> ] >>>>> >>>>> Andrew Palumbo commented on MAHOUT-1653: >>>>> ---------------------------------------- >>>>> >>>>> So upgrading to Spark-1.3 does completely break our shell. >>>>> >>>>> Spark 1.3 >>>>>> --------- >>>>>> >>>>>> Key: MAHOUT-1653 >>>>>> URL: https://issues.apache.org/jira/browse/MAHOUT-1653 >>>>>> Project: Mahout >>>>>> Issue Type: Dependency upgrade >>>>>> Affects Versions: 0.10.0 >>>>>> Reporter: Andrew Musselman >>>>>> Assignee: Andrew Palumbo >>>>>> Priority: Critical >>>>>> >>>>>> Support Spark 1.3 >>>>>> >>>>> >>>>> -- >>>>> This message was sent by Atlassian JIRA >>>>> (v6.3.4#6332) >>>>> >>>>>
