On Tue, Jul 23, 2019 at 3:37 PM leerho <[email protected]> wrote:

>
> NOTE 2: All of the code has been properly refactored with
> "org.apache.datasketches...".
> All source files have the proper Apache license and have been checked with
> the Maven Rat Plugin.
> The code passes all tests with a coverage of > 98%.
>

This is really great to have done. Nice!


> Git Tag for this release:
>
> https://github.com/apache/incubator-datasketches-memory/tree/1.0.0-incubating-RC2
>

I notice that throughout the git history every commit has version
1.0.0-incubating. You may have considered this already, but maven supports
1.0.0-incubating-SNAPSHOT which will avoid inappropriate caching. At
publish time it will be replaced with a timestamp, if I recall the exact
system. So it often makes sense for every mainline commit on the master
branch to have this version in the pom.xml and you just make a single
tagged commit not on master where the version has been changed.

master --a--b--c--d--e--f-- ...
                \     \
                                    x     y

All of a, b, c, d, e, f can have a SNAPSHOT version while x and y, which
exist just to build an RC, can have a non-SNAPSHOT version.

5. Repository: Maven Central (repository.apache.org):
>
> Upon acceptance the jar artifacts will be generated from the source
> repository and deployed
> to the Apache Maven Central staging repository:
>
>
> https://repository.apache.org/content/groups/staging/org/apache/datasketches/memory/
>

It might be nice to do the staging prior to the vote. Technically they are
"convenience binaries" but it is helpful to test them as well, to uncover
any problems before approving the release. Avoids voting +1 on source that
ends up building a bad binary.


>   $ mvn clean test -P strict
>
> Note also that when running the test suite, you might get the following
> message:
>
>   SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
>   SLF4J: Defaulting to no-operation (NOP) logger implementation
>   SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for
> further details.
>
> This is normal.  It just indicates that in your environment, you do not
> have a logger in your class-path so the default logger of a no-op is used
> instead.
>

You should be able to add an slf4j backend to the <scope>test</scope> in
your pom.xml so it does not depend on the user/contributor to do so.

The documentation for the DataSketches Memory component is part of the
> website.
>
> Overview documentation:
> - https://datasketches.github.io/docs/Memory/MemoryPackage.html
> - https://datasketches.github.io/docs/Memory/MemoryPerformance.html
>
> Javadocs:
> https://datasketches.github.io/api/memory/snapshot/apidocs/index.html
>

Is there a versioned revision to the javadocs that could also be previewed
as part of the RC?

8. The vote will be performed in two stages:
>    - This letter will be published on dev@ open for at least 72 hours or
> until necessary number of votes are reached.
>    - This letter will be published on incubator@general for at least 72
> hours or until necessary number of votes are reached.
>

As mentioned on the [VOTE] thread, votes should stay open for 72 hours.



>   Anybody can vote, but only IPMC member's votes count.
>

You might say only IPMC members' votes are binding. This is true only for
the vote on general@incubator. For the podling, the PPMC votes should be
considered binding as to whether or not the artifact should go on to the
IPMC vote.


> The Apache DataSketches Team
>

I believe you should sign your emails as yourself.

Kenn

Reply via email to