Hello

The only question left is whether we also put the apidocs into the
pear package. I've looked around a bit but judging on the ones I have
installed this seems not be usual. It would also be a problem as 
PEAR is done in the "default:package" and the apidocs in "site:site"
goal...

bye,

-christian-


Am Fri, 30 Oct 2009 06:54:37 +0100
schrieb Christian Grobmeier <[email protected]>:

> > 1. The source .tar.gz and .zip contain the site/coverage-report/
> >   directory. While including the generated phpdoc could be
> > considered useful for people who just want to use log4php and not
> > rebuild it, the coverage report seems pretty useless to me.
> 
> Yes, I think so.
> 
> > 2. Should we distribute the apidocs alongside the .tar.gz even if
> >   they could be generated by the user using maven?
> 
> I would include API docs because the risk that a user needs em is
> quite high.
> 
> > 3. The package-config.php and package.php are not included in
> >   the .tar.gz and .zip. IIRC because they are not needed by the
> >   endusers but it just came to my mind that this means that the
> >   user cannot generate the phpdoc using "mvn site" as this would
> > fail due to the missing files. Ok, maybe phpdoc is generated before
> > the pear package, I don't remember, but it looks a bit
> > unprofessional if we distribute the pom.xml but know that it will
> > fail with errors, or?
> 
> I agree again :-) IMHO a user should be able to run tests, coverage
> reports etc. all himself.
> This also goes for pear packages.
> 
> 
> > 4. The creation of the superfluous empty .jar will not annoy any of
> >   the reviewers, or? I have no idea how to turn it off and it would
> >   probably require an ant task to simply remove it.
> 
> Voting is upon the assemblies itself, not  on the creation process. We
> will not distribute this jar
> nor would we make it available for voting. Means, nobody will be
> annoyed of that.
> We can change the packaging mode to "dir" maybe and then only a
> directory will appear.
> If we want something else, we would have to create a custom lifecycle
> binding, but I don't think
> we need this at the moment.
> 
> Cheers,
> Christian
> 
> >
> > bye,
> >
> > -christian-
> >

Reply via email to