incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jason van Zyl <jvan...@sonatype.com>
Subject Re: [VOTE] Approve the release of Shindig Incubator 1.0
Date Tue, 28 Apr 2009 20:27:27 GMT

On 28-Apr-09, at 11:50 AM, Henning Schmiedehausen wrote:

> Calm down, Jason. No one was attacking Maven.
>

I'm not upset, seriously. This is standard fare here, and only  
pointing out having any solution which involves Maven might better be  
on a path that actually involves Maven developers.

> The Apache Software Foundation requires a project (not just an  
> incubator
> podling. All projects) to release source in a form that can be used to
> recreate the binaries.
>
> For the current state of the Shindig release, this is not possible.  
> Noone
> was saying anything else.
>
> For Apache, we release source code that is immediately consumable to  
> users
> by downloading an artifact from our servers through www.apache.org/dist 
>  and
> potentially be able to rebuild the artifact.
>
> In general, this is a tarball / zip of the contents of a Subversion  
> tag.
>
> Everything beyond that is
>
> - optional
> - a convenience to our users
>
> This especially goes to
>
> - binary archives (whether these are .jar archives in Java or Binary  
> builds
> for a platform in non-Java land)
> - source/javadoc in another, better consumable form for IDEs
>
> Supporting these is nice to users, but the required part is the  
> tarball that
> I can go into and say <build-command> (be it ant, maven or make) and  
> get
> something that can be used.
>
> This is not the case for Shinig in its current state. Whether this is
> acceptable or not to the Incubator PMC members is another question.
>
>    Ciao
>        Henning
>
>
>
> On Mon, Apr 27, 2009 at 09:49, Jason van Zyl <jvanzyl@sonatype.com>  
> wrote:
>
>>
>> On 26-Apr-09, at 8:25 PM, Niclas Hedhman wrote:
>>
>> On Mon, Apr 27, 2009 at 6:46 AM, Vincent Siveton
>>> <vincent.siveton@gmail.com> wrote:
>>>
>>> You need to do a checkout of the tag to build it.
>>>> The source artifacts provide only java source, no build file.
>>>>
>>>
>>> -1.
>>>
>>> As others have pointed out, the ASF releases Open SOURCE, not Open
>>> Binaries and part of the policy is that the distributed artifact is
>>> first and foremost a buildable source tarball. Without it, it is  
>>> not a
>>> release.
>>> You may have system requirements ("thou shalt need Maven 2.0.6 or
>>> 2.0.9" ) and you should provide full build instructions to produce  
>>> the
>>> projects binary. And if you distribute a binary, it should be the  
>>> same
>>> thing that gets produced by following your instructions.
>>>
>>> And the above is not really up for debate either. At the end of the
>>> day, people will rely on tarballs, checksums (download integrity)  
>>> and
>>> signatures (manipulation integrity), and those are the primary
>>> artifacts that Apache Infrastructure will archive and get mirrored
>>> around the world.
>>> Maven artifacts are really nice to have for Maven users, but is
>>> exactly that; "Nice to have".
>>>
>>> Now, you are free to go banging on Maven's door that their built-in
>>> workflow doesn't support the Apache policy very well.
>>>
>>
>> Don't spread FUD like that. You don't have any idea how Maven  
>> releases work
>> so I'll take a moment and explain it to you.
>>
>> For a release like Maven we have N modules, where each module  
>> produces a
>> JAR, and each of those is released. Each JAR carries with it the  
>> POM inside
>> it, but is in a form which can be automatically utilized by IDE  
>> integration
>> to automatically be downloaded and integrated with debuggers. All  
>> the legal
>> bits are in the JARs and legally intact as it were. The blue print to
>> actually build that individual JAR is in the JAR by default in  
>> Maven. You
>> can't just unpack that source JAR and build it and that's for a  
>> couple
>> reasons: the first is that it's generally useless and the second is  
>> that we
>> create a source archive of the entire release with all the modules  
>> which is
>> what we recommend. As with Maven, the tagged sources for the build  
>> are
>> distributed along with the binaries. This is a matter of setting up  
>> a source
>> assembly, not hard to do.
>>
>> That said, show me any non-Maven project that makes individual JARs  
>> that
>> have the capability of rebuilding themselves. There aren't any here  
>> at
>> Apache. What gets produced is a overall source archive. And show me  
>> anything
>> as advanced and useful to developers where grabbing the source JARs  
>> for
>> debugging is transparent or materializing sources from binary  
>> artifact
>> coordinates is a button click. Again, nothing does that besides  
>> Maven and
>> the corresponding IDE integrations. So Maven adheres to any  
>> standard for the
>> overall release but goes way above and beyond to actually produce  
>> something
>> far more useful for actually doing work.
>>
>> So please don't try to explain to people what Maven does when you  
>> don't
>> actually understand it yourself. What gets released as the N  
>> modules is
>> complementary to aggregate release which is compliant with Apache.  
>> So if
>> Shindig doesn't have the overarching source archive that's not hard  
>> to add.
>> But there is not a single non-Maven build at Apache where a JAR  
>> emanating
>> from multi-JAR build where that JAR carries with it the information  
>> to build
>> itself. You are confusing an aggregate release with the releases of  
>> the
>> individual components which is what Maven users need to consume. We  
>> account
>> for both for the case where a user grabs the distribution to use,  
>> and the
>> case where someone is building a Maven plugin (most often in an IDE  
>> where
>> Maven is not installed) and transparently grabs the dependencies --  
>> the
>> individual JARs from a Maven release -- so they can build their  
>> project.
>>
>> We do the necessary and the nice to have. More working developers  
>> actually
>> care about the nice to have.
>>
>>
>> They won't be
>>> surprised, and IIRC there has been a lot of discussion over there to
>>> make it happen.
>>>
>>>
>>> Cheers
>>> --
>>> Niclas Hedhman, Software Developer
>>> http://www.qi4j.org - New Energy for Java
>>>
>>> I  live here; http://tinyurl.com/2qq9er
>>> I  work here; http://tinyurl.com/2ymelc
>>> I relax here; http://tinyurl.com/2cgsug
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
>>> For additional commands, e-mail: general-help@incubator.apache.org
>>>
>>>
>> Thanks,
>>
>> Jason
>>
>> ----------------------------------------------------------
>> Jason van Zyl
>> Founder,  Apache Maven
>> http://twitter.com/jvanzyl
>> http://twitter.com/SonatypeNexus
>> http://twitter.com/SonatypeM2E
>> ----------------------------------------------------------
>>
>> Simplex sigillum veri. (Simplicity is the seal of truth.)
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
>> For additional commands, e-mail: general-help@incubator.apache.org
>>
>>

Thanks,

Jason

----------------------------------------------------------
Jason van Zyl
Founder,  Apache Maven
http://twitter.com/jvanzyl
http://twitter.com/SonatypeNexus
http://twitter.com/SonatypeM2E
----------------------------------------------------------

We know what we are, but know not what we may be.

   -- Shakespeare


---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message