lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steven A Rowe <>
Subject RE: Improvements to the maven build
Date Thu, 05 May 2011 14:31:18 GMT
Hi Ryan,

On 5/4/2011 at 7:14 PM, Ryan McKinley wrote:
> As a rule, everything should go through JIRA on its way to svn -- this
> is important so that we have somewhere to point for why we did things.
>  Even small things.

Your phrase "As a rule" provides wiggle room that we all use.  "Even small things."  Um, I
don't think so.  E.g. no-one is going to go through JIRA for a small typo fix.  This judgment
about what's big enough to warrant a JIRA issue is one each committer has to make.  As a result,
this argument (David's patch should have gone through JIRA because everything should go through
JIRA) doesn't work for me.

> With patches from contributors it is especially important they are
> added to JIRA because they need to grant the license to ASF.  Also
> attachments are often stripped from mailing list archives, so down the
> road its really hard to know what happened.

These are both excellent points.  Non-trivial non-committer patches should definitely go through
JIRA for these reasons.

> I understand the desire to keep maven support low key -- but we should
> do that with a good README in dev-tools.

I agree that the Maven build should be documented - I plan on putting something together soon,
as suggested by David.  This seems completely orthogonal to me, though, to the question of
using JIRA issues for Maven build changes.

> Even as officially non-official tools, it still gets into svn so
> we need a trail of where it came from and hopefully a log of why
> we thought it was important.

I agree in principle, but again, I'll continue to use my own judgment about whether to use
JIRA for small changes, especially to stuff under dev-tools/.


View raw message