incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sebb <>
Subject Re: [VOTE] Release Kafka 0.7.1-incubating (Candidate 3)
Date Sun, 24 Jun 2012 14:32:15 GMT
On 24 June 2012 14:11, Emmanuel Lécharny <> wrote:
> Le 6/24/12 10:46 AM, Chris Douglas a écrit :
>> Kevan-
> Hi, jumping on this thread, because we have had similar discussions on
> another incubating project I'm mentoring...
>> This argument is logical, but that doesn't make it legal and thus
>> required. It is also a constraint foreign to most TLPs, making its
>> rigid enforcement in the incubator unfair and arbitrary. Not to be
>> glib, but you clearly do have time to pursue this broadly and across
>> the foundation if, in fact, you're right and many- if not most- ASF
>> projects are improperly managing this obligation.
> The fact that many projects at the ASF are not releasing correctly is not a
> good reason for not doing it correctly for Kafka. Moreover, I'm pretty sure
> that sooner or later, those projects will have to comply to what seems to be
> a reasonable requirement.


>> The references provided do not make the argument you're forwarding, at
>> least they do not distinguish it from the criteria we used for
>> evaluating Kafka 0.7.0. The claim is that Kafka "distributes" the
>> transitive closure of its dependencies NOT because they're mentioned
>> in the build source (where you carve out an exemption), but because
>> these deps are mentioned AND Kafka's build produces a server. This
>> distinction between "applications" and "libraries" is unmentioned in
>> the documentation. In fact, I cannot create an argument for
>> documenting the transitive closure of dependencies using those
>> references. The argument (as forwarded) requires this step to reach
>> its conclusion, but it appears to be novel and unsupported. The
>> passages you quote are already common ground; we both agree that the
>> artifact must contain an "appropriate LICENSE and NOTICE file", but we
>> disagree on the scope of "appropriate". I continue to disagree; I
>> don't think the case has been made.
>> The closest documentation I could find was here:
>> Which claims to be a draft of a document that also doesn't give
>> concrete guidance on this point.
> AFAICT, the important thing is that those who will download Kafka (or any
> other ASF project) can't be fooled when includig it into their own projects.
> The N&L files are here to facilitate our user's work, by listing all the
> external third party components we depend on, and that includes transitive
> dependencies.


The N&L files only apply to what is included in the release archive itself.

If there are other dependencies, these can be listed in a

There's an entirely separate issue here, i.e. which dependencies are
allowed for an ASF project, regardless of whether the dependencies are
bundled or not.

That is addressed here:

> It's also a guarantee that we have checked that no depency contains a
> transitive dependency that is *not* compatible with our licence. How good
> will it be to include a dependency that includes a GPL3 licensed 3rd party
> itself ?

See above.

> Having checked all the transitive dependencies ourself is the only way to
> provide a guarantee to your users, and by listing those transitive


> dependencies in the  N&L files is the only way to go.

No, because the N&L files only apply to what is actually being shipped.

For example, a source release of a project with no 3rd party source
will have N&L files that only mention the ASF.
It's important not to encumber the end-user with N&L information that
does not apply.

> At least this is my interpretation on the various discussions we have had in
> the ast two months while tryibg to cut a release on Syncope (incubator) and
> SSHD (a MINA subproject).
> Hope it helps...
> --
> Regards,
> Cordialement,
> Emmanuel Lécharny
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message