Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo >> mail # dev >> moving rat to a profile?


Copy link to this message
-
Re: moving rat to a profile?
Since this came up in Review (https://reviews.apache.org/r/23391/), I'm
revisiting this thread.

In short, it seems there was some confusion on whether or not we had
reached consensus on enabling rat by default, with rat.ignoreErrors=true
set to default, and requiring any entities that should be responsible for
checking legal status of contributions to turn it in (set
rat.ignoreErrors=false for Jenkins, committers, etc.)

I think one problem is that there was not actually consensus. This was a
proposal made previously by Bill Havanki to do the above, after which we
agreed to minimally making changes to make "git clean -df" sufficient. But
there did not appear to be any consensus reached on disabling error
checking by default, as proposed. Billie had expressed concerns about
encouraging committers to not take care in checking the legal status of
their commits, and not only do I share those concerns, I'd also extend my
concerns to any contributor. This is simply something that everybody
contributing to ASF should be aware of. Committers have the extra
responsibility, but even contributors and potential committers should
understand the contribution requirements, the development process, and have
put in some minimal amount of thought about whether they have the ability
to contribute a particular piece of code legally.

Additionally, I'm concerned about who we'd be easing things for. There's
clearly an increased risk of commits/contributions without strict checking
for licenses (we've seen past commits with these errors... just do `git log
| grep -i license` to see that). So, we're accepting that risk by making
the easy route one that serves who, exactly? The ASF makes it quite clear
that SCM is for developers, not end users, and this is an SCM-specific
problem (specifically branch switching in git).

So, I do not think there is consensus. There are certainly outstanding
concerns for disabling the rat check by default. However, I think there is
room for further discussion and compromise, and I'd love to consider more
solutions (or discuss the merits of the proposed solutions). For instance,
one area of compromise I've conceded (as stated in the review) is that I
think it'd be a good idea to put the rat check in a profile which can be
optionally disabled, so that it doesn't take the 3 extra seconds to run,
which could impact some iterative testing (like running one IT at a time,
or doing automated `git bisect`).
Christopher L Tubbs II
http://gravatar.com/ctubbsii
On Thu, Jun 19, 2014 at 11:15 AM, Christopher <[EMAIL PROTECTED]> wrote:
 
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB