commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Phil Steitz <phil.ste...@gmail.com>
Subject Re: [VOTE] Release Apache Commons Functor 1.0 based on RC1
Date Fri, 21 Oct 2011 16:13:46 GMT
On 10/21/11 8:49 AM, Emmanuel Bourg wrote:
> Le 21/10/2011 17:04, Matt Benson a écrit :
>
>> The most immediate thing is that [proxy] 2 needs a unary predicate.
>> It becomes ridiculous for every component we have to define such a
>> basic interface in a different way.  So [proxy] 2 can never be
>> released without a [functor] release, etc.
>
> Maybe the Predicate class from [collections] could be reused since
> [proxy] already depends on this component?
>
>
>> This is indeed sufficient for the immediate future, but again,
>> becomes
>> *insufficient* soon enough.  I don't personally see what purpose is
>> served by a 0.x release, but I have to say it's frustrating to
>> work on
>> a component and only when a release is proposed does everyone who
>> couldn't have cared less before come out of the woodwork to suddenly
>> find design issues (issues such as undocumented @SuppressWarnings
>> come
>> down to personal opinion and I would personally argue they shouldn't
>> be release blockers, particularly as more often than not the
>> justification is obvious).  For those of you who have suddenly
>> decided
>> you care about this component:  if we discuss and resolve the
>> fundamental design concerns (i.e. no promises beyond reaching
>> consensus), can you stay engaged long enough for us to get this
>> wrapped up?
>
> Sorry but that's what released candidates are for. You can't
> expect everyone to monitor every project continuously, especially
> the sandbox stuff. Releasing a RC is the right moment to ask for
> reviews since the code is arguably stabilized.
>
> You'll note that I didn't vote -1 on the release, I just spent 2
> hours to review the code and shared my observations thinking it
> would be useful, but I won't block the release if you think my
> points are irrelevant. Feel free to release Functor as is and I
> won't bother you again. But if you take my observations seriously
> I'll happily stay engaged with this project.
>
> As a side note, in general I wish there were more design/usability
> reviews on RC calls than checkstyle/compilation verifications.

+1 and thanks in advance for doing this on any of the components
that I work on.  I know there is a balance here between wanting to
get stuff out and differences in taste, etc.; but I have always
viewed it as a great strength of Commons that the sometimes very
small subcommunities around individual components can benefit from
the input and advice of the broader community.  More people looking
more deeply into more code in more components is all good for us. 

Per my earlier post today, a lot of this kind of review can and
should happen actually before RCs are cut and VOTEs are kicked off.

<shameless-plug>We are working on [math] 3.0 and [pool] & [dbcp] 2.0
now - code and API reviews are most welcome </shameless-plug>

Phil
>
> Emmanuel Bourg
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org


Mime
View raw message