Hi Ed!

On Mon, Apr 3, 2017 at 1:09 PM, Ed Espino <[email protected]> wrote:
> I'm on the Apache HAWQ Incubator project. Our initial binary release will
> be validated against TLP Apache Bigtop. Coincidentally, our release paths
> are crossing as both projects are currently in the early stages of their
> respective voting processes. I have a few of very preliminary
> "Observations" to share as I am getting acquainted with Bigtop. I'm also
> fairly new to the Apache release processes and my comments might be
> slightly off point (so please bear with me).

First of all -- welcome to the Bigtop community and thanks for taking time
to share your feedback. See my comments inline below:

> LICENSE: I noticed the LICENSE file only contains the ASF v2.0
>          license.  Is there a need to mention the licenses of the
>          tools recommended for rapid provisioning (example; Vagrant,
>          Docker, Puppet) or other non-ASF projects?

Strictly speaking: no. Sine these tools are not being touched in any way
by the Bigtop codebase. This is similar to how we don't mention the
license of make when we ship makefiles.

> NOTICE: Copyright year needs to be updated (currently 2014)
>
>         Should it change to a range (2014-2017)?

Great catch! I filed https://issues.apache.org/jira/browse/BIGTOP-2731

>   OBSERVATION: The format of the hashes is not super convenient for an
>                automatic comparison. Being compatible with command
>                line tools would be helpful.

Good point, but our checksums are generated automatically by doing
a Maven release to the ASF Maven repo. That determines the format.
Once we complete our transition to Gradle in 1.3.0 -- I'll try to see if
we can make the format more friendly.

> ======================================================================
> GIT TAG
> ======================================================================
>
> Git tag matches src (except .git*)
> commit: 5a2a1a86b17f73260229137b0b008385fd32bfae
>
> OBSERVATION: I noticed the commit hash was not provided. I've seen
>              several podling votes with them. Should the hash also
>              been provided along with the corresponding tag?

Given that tags under rel/ are supposed to be immutable I don't see a strong
reason for why hash would be required. This wasn't the case a year or so
ago and that's probably why you see hashes still being given.

> ======================================================================
> Apache RAT
> ======================================================================
>
> The "gradlew rat" execution passed.
>
> OBSERVATIONS:
>
>  * Initially I ran using maven and have come to know I should use
>    gradle. For grins, I generated a RAT Report summary using maven
>    which intentionally did not have any exclusions identified.
>
>    Here is the "mvn apache-rat:check" summary which generated
>    a few followup observations (below):
>
>     *****************************************************
>     Summary
>     -------
>     Generated at: 2017-04-03T11:59:14-07:00
>
>     Notes: 97
>     Binaries: 9
>     Archives: 3
>     Standards: 1994
>
>     Apache Licensed: 1353
>     Generated Documents: 0
>
>     JavaDocs are generated, thus a license header is optional.
>     Generated files do not require license headers.
>
>     641 Unknown Licenses
>
>     *****************************************************
>
>  * Should the yaml files included in the release have ASF headers?  It
>    is possible the parsers have troubles with comment ASF headers. But
>    then again, maybe the current parsers can handle them.

Another great find. I believe originally we had problems with Puppet and Juju
YAML parser somehow choking on comments. I'm pretty sure that is no longer
the case -- but we need to make sure. I filed
https://issues.apache.org/jira/browse/BIGTOP-2732
to address this.

>  * There are three archives. Are these acceptable?
>
>    +
> bigtop-packages/src/charm/giraph/layer-giraph/resources/giraph-examples-1.1.0.jar
>    + bigtop-tests/test-artifacts/hadoop/src/main/resources/cachedir.jar
>    +
> bigtop-tests/test-artifacts/hadoop/src/main/resources/test_data/test.zip

I don't think these represent a problem, but you're right it would be
much nicer not
to have them on the exclusion list. We'll try to look into that as
part of BIGTOP-2732.
I think the ones in test are OK either way -- but the one in charms
can be fixed.

Thanks,
Roman.

Reply via email to