Hello.

Le lun. 17 janv. 2022 à 13:33, Thomas Vandahl <t...@apache.org> a écrit :
>
> Hi Gilles,
>
> > Am 14.01.2022 um 15:37 schrieb Gilles Sadowski <gillese...@gmail.com>:
> >
> >>> Maven artifacts are here:
> >>>   
> >>> https://repository.apache.org/content/repositories/orgapachecommons-1576/org/apache/commons/commons-jcs3/3.1/
> >
> > I only see one ".pom" file and one ".xml" file (with their respective
> > crypto sig).
> >
> … and modules below 
> https://repository.apache.org/content/repositories/orgapachecommons-1576/org/apache/commons/commons-jcs3

Actually they _above_ link, hence my confusion!
One has to click on "Parent  Directory" with target
  
https://repository.apache.org/content/repositories/orgapachecommons-1576/org/apache/commons/
in order to see the modules.

>
> > We should note that this command requires the environment variable
> >  JAVA_HOME
> > to be set (otherwise the build will fail).
> >
> Is this a special requirement of this build?

No; the remark is for making the reviewer's life easier, i.e just copy/paste
the command and see the build succeed.  Without the environment variable,
the "javadoc" step fails IIRC.

> If so, why?
>
> > Shall we stop cluttering what should be a summary of important changes,
> > easily readable by a human, with bot-generated messages and trivial
> > one-liners like "isEmpty()"?
> >
> I followed the directions for the release preparation at 
> https://commons.apache.org/releases/prepare.html and generated the release 
> notes with „mvn changes:announcement-generate -Prelease-notes“ and the vote 
> e-mail with „mvn commons-release:vote-txt“ as recommended. What is wrong with 
> this procedure?

It's not the procedure.  I simply stress FTR that the release notes
become less and less what they should be, that is a summary for
human consumption that attracts the user's attention on functional
changes. [The full set of changes is always accessible through the
commits log.]
To ensure clarity for users, changes that cannot affect them should
not be listed in "changes.xml".

>
> > Is the Log4j2 version update related to the latest security issue?
> > If so, that may have been important to note.
> Actually, no. log4j is an optional dependency.
>
> > Nit-pick: Rather than tagging them with "IMPORTANT CHANGE", it would
> > be clearer that important changes are mentioned in the description, at the
> > top of the release notes.
> This *is* part of the description. You cannot tag changes in changes.xml. I 
> remember discussions about making releasing easier. Every *manual* step such 
> as writing release notes does not follow this direction, IMO.

My opinion is that the release notes are essentially a "manual"
step (even though we get help from a tool that formats the contents
of "changes.xml") in the sense that the developers "talk" to the
users and tell them why they should upgrade.  So I think that
important changes should be mentioned, if just with a sentence
like "There are important changes, please read through the list
below" (and that's why the list should not be cluttered with trivial
changes).

>
> > There are a few changes marked incompatible.  Is it expected in a minor 
> > release?
> The japicmp report recommends a 0.1.0 release which it is.

I don't understand why, but OK then...

>
> >>> [ ] +1 Release these artifacts
> >>> [ ] +0 OK, but...
> >>> [ ] -0 OK, but really should fix...
> >>> [ ] -1 I oppose this release because…
>
> Would you please consider voting?

+1

[There are no technical issues, only communication issues that are
common problems with all the releases.]

Best regards,
Gilles

>
> Bye, Thomas

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org

Reply via email to