sergio's observations in reviewing the release

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

sergio's observations in reviewing the release

Joe Witt
sergio commented in his RC2 review:

+++
Things could be polished for future releases (all non-blocking, probably
more question for discussion than issues actually):

* role of DEPENDENCIES file, how it compares with the information already
available at LICENSE and NOTICE
* I'd not recommend to use the Maven for publishing source tarballs
  * voting releases should go to
https://dist.apache.org/repos/dist/dev/incubator/nifi/
  * passed release should go to
https://dist.apache.org/repos/dist/release/incubator/nifi/
* I'd exclude of your build process the hash generation for signatures
(*.asc.md5 and *.asc.sha1). The maven-gpg-plugin can be tricky, this is how
we solved it in another project http://s.apache.org/60h
+++

First, what was used to make the shortened URL?  That and the fact that it
resolved to a specific highlighted line in the pom was very cool.

The DEPENDENCIES file: I agree this is an irritant in our case (might be
great for other cases).  I'd love to exclude it if we can sort out how to
without resorting to doing anything too wild in our POM.  The current POM
configuration we have for the release process is quite simple and nice.

Excluding hash generation for signatures: Looks really easy to do thanks to
your URL so happy to incorporate that.

Thanks
Joe
Reply | Threaded
Open this post in threaded view
|

Re: sergio's observations in reviewing the release

Joey Echeverria
It looks like anyone can create a short url by entering it in at
http://s.apache.org.

On Thu, Jan 22, 2015 at 11:28 AM, Joe Witt <[hidden email]> wrote:

> sergio commented in his RC2 review:
>
> +++
> Things could be polished for future releases (all non-blocking, probably
> more question for discussion than issues actually):
>
> * role of DEPENDENCIES file, how it compares with the information already
> available at LICENSE and NOTICE
> * I'd not recommend to use the Maven for publishing source tarballs
>   * voting releases should go to
> https://dist.apache.org/repos/dist/dev/incubator/nifi/
>   * passed release should go to
> https://dist.apache.org/repos/dist/release/incubator/nifi/
> * I'd exclude of your build process the hash generation for signatures
> (*.asc.md5 and *.asc.sha1). The maven-gpg-plugin can be tricky, this is how
> we solved it in another project http://s.apache.org/60h
> +++
>
> First, what was used to make the shortened URL?  That and the fact that it
> resolved to a specific highlighted line in the pom was very cool.
>
> The DEPENDENCIES file: I agree this is an irritant in our case (might be
> great for other cases).  I'd love to exclude it if we can sort out how to
> without resorting to doing anything too wild in our POM.  The current POM
> configuration we have for the release process is quite simple and nice.
>
> Excluding hash generation for signatures: Looks really easy to do thanks to
> your URL so happy to incorporate that.
>
> Thanks
> Joe



--
Joey Echeverria
Reply | Threaded
Open this post in threaded view
|

Re: sergio's observations in reviewing the release

Joe Witt
I really should have tried that base URL.... yowza.  Thanks Joey.

On Thu, Jan 22, 2015 at 2:33 PM, Joey Echeverria <[hidden email]> wrote:

> It looks like anyone can create a short url by entering it in at
> http://s.apache.org.
>
> On Thu, Jan 22, 2015 at 11:28 AM, Joe Witt <[hidden email]> wrote:
> > sergio commented in his RC2 review:
> >
> > +++
> > Things could be polished for future releases (all non-blocking, probably
> > more question for discussion than issues actually):
> >
> > * role of DEPENDENCIES file, how it compares with the information already
> > available at LICENSE and NOTICE
> > * I'd not recommend to use the Maven for publishing source tarballs
> >   * voting releases should go to
> > https://dist.apache.org/repos/dist/dev/incubator/nifi/
> >   * passed release should go to
> > https://dist.apache.org/repos/dist/release/incubator/nifi/
> > * I'd exclude of your build process the hash generation for signatures
> > (*.asc.md5 and *.asc.sha1). The maven-gpg-plugin can be tricky, this is
> how
> > we solved it in another project http://s.apache.org/60h
> > +++
> >
> > First, what was used to make the shortened URL?  That and the fact that
> it
> > resolved to a specific highlighted line in the pom was very cool.
> >
> > The DEPENDENCIES file: I agree this is an irritant in our case (might be
> > great for other cases).  I'd love to exclude it if we can sort out how to
> > without resorting to doing anything too wild in our POM.  The current POM
> > configuration we have for the release process is quite simple and nice.
> >
> > Excluding hash generation for signatures: Looks really easy to do thanks
> to
> > your URL so happy to incorporate that.
> >
> > Thanks
> > Joe
>
>
>
> --
> Joey Echeverria
>
Reply | Threaded
Open this post in threaded view
|

Re: sergio's observations in reviewing the release

Sergio Fernández
In reply to this post by Joe Witt
Hi,

On 22/01/15 20:28, Joe Witt wrote:
> First, what was used to make the shortened URL?  That and the fact that it
> resolved to a specific highlighted line in the pom was very cool.

At http;//s.apache.org you have the official ASF URI shortener that
every committer can use, with an optional key for building meaningful
URIs. I use it a lot, I though you know it; I guess many guides do not
cover such things.

The highlighted line is regular feature of GitHub. Normally I use the
official ASF git for official things, but for showing such snippets
GitHub is much more useful.

> The DEPENDENCIES file: I agree this is an irritant in our case (might be
> great for other cases).  I'd love to exclude it if we can sort out how to
> without resorting to doing anything too wild in our POM.  The current POM
> configuration we have for the release process is quite simple and nice.

I though was a custom file in your release profile... Does it come from
Apache Parent POM? First time I see it.

> Excluding hash generation for signatures: Looks really easy to do thanks to
> your URL so happy to incorporate that.

Yes, that patch should be trivial to apply to NiFi.

Thanks guys for addressing with such diligence my comments.

Cheers,

--
Sergio Fernández
Partner Technology Manager
Redlink GmbH
m: +43 660 2747 925
e: [hidden email]
w: http://redlink.co