RPM support update

[jon.vanalten at redhat.com]

----- "Dalibor Topic" <Dalibor.Topic at Sun.COM> wrote:

> jon.vanalten at redhat.com wrote:
> > Dalibor, we can adapt to parsing the output of jpkg in whatever
> format you end up implementing, but can you post with some idea of how
> you envision the output?
> 
> Hi,
> 
> I am thinking of the output being a simple list of property-like
> assignments, in the form of jigsaw.x.y.z=something which should
> be directly usable in scripts, translatable to XML via the
> Properties API, etc.

++

> 
> So, for example, dumping the metadata of a simple hello world
> module would result in the output
> 
> jigsaw.moduleinfo.id.name = org.openjdk.jigsaw.test.hello
> jigsaw.moduleinfo.id.version = 1.0
> jigsaw.moduleinfo.requires.1 = jdk.base @ 7-ea
> jigsaw.moduleinfo.requires.2 = something.else @ 0.4
> jigsaw.moduleinfo.mainClass =
> org.openjdk.jigsaw.test.hello.HelloWorld
> 
> with the hierarchical naming according to the data types in
> http://cr.openjdk.java.net/~mr/jigsaw/api/java/lang/module/ModuleInfo.html
> 
> where a set of items like the requires would be represented as
> a set of assignments with a different number at the end (j.m.r.1,
> j.m.r.2 above).
> 
> In terms of granularity, I think this is a preferable form to
> splitting
> the dependencies apart further - as each dependence itself consists of
> a
> set of dependence modifiers, a module name, and a version query, and
> I
> am not sure that
> 
> jigsaw.moduleinfo.requires.1.name = jdk.base
> jigsaw.modileinfo.requires.1.versionQuery = 7-ea
> 
> is a better representation for your needs then
> 
> jigsaw.moduleinfo.requires.1 = jdk.base @ 7-ea
> 
> what do you think?
> 

I agree with the one-liner requires.  I wonder if the same theory can be applied to the name/version lines too? ie:

jigsaw.moduleinfo.id.name = org.openjdk.jigsaw.test.hello @ 1.0

A couple of other questions, then.  I wonder what the "#" in j.m.r.# is needed for.  I can see if the version info was split off to a separate line, the number would be needed to match the requires.name with requires.version.  But with it all on one line, is this necessary?

Finally, is there the possibility of something like this:

j.m.r >= jdk.base @ 7-ea

(ie requiring some version greater than or equal to the version specified).


cheers,

jon

> cheers,
> dalibor topic
> -- 
> *******************************************************************
> Dalibor Topic 			Tel: (+49 40) 23 646 738
> Java F/OSS Ambassador 		AIM: robiladonaim
> Sun Microsystems GmbH 		Mobile: (+49 177) 2664 192
> Nagelsweg 55 			http://openjdk.java.net
> D-20097 Hamburg 		mailto:Dalibor.Topic at sun.com
> Sitz der Gesellschaft: Sonnenallee 1, D-85551 Kirchheim-Heimstetten
> Amtsgericht München: HRB 161028
> Geschäftsführer: Thomas Schröder, Wolfgang Engels, Wolf Frenkel
> Vorsitzender des Aufsichtsrates: Martin Häring