[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: libxp-java : XML 1.0 parser for Java



W liście z sob, 29-05-2004, godz. 14:22, Arnaud Vandyck pisze: 
> Nicolas Duboc <nicolas@duboc.net> writes:
> 
> > On Fri, May 28, 2004 at 05:00:23PM +0200, Arnaud Vandyck wrote:
> >> Nicolas Duboc <nicolas@duboc.net> writes:
> >> 
> >> Why sablevm and gij aren't alternatives for java1-runtime? (if there is
> >> no reason, can I add them as alternatives?
> >
> >    I have added kaffe on the Depend line because of the lintian warning
> > "virtual-package-depends-without-real-package-depends" [1].
> 
> [...]
> 
> >    I don't think adding other alternatives is useful. But if I'm wrong,
> > I will add it, of course.
> >
> > [1] http://lintian.debian.org/reports/Tvirtual-package-depends-without-real-package-depends.html
> 
> I agree. I tend to add other alternatives but it's not necessary.
> 
> I'll upload your package.
> 
> If someone think it's useful, file a wishlist bug against libxp-java.

Actually, especially in case of packages that work with free java
environments I think it IS useful.  

You wrote:
"I have chosen kaffe since I know the package perfectly works with kaffe
 VM and libraries. It also works with sablevm (it seems to not work on
 gij but I don't know yet why)."

This is IMO exactly the kind of informations you want to give to a user
by the Depends: alternatives.  In this case you would put something
like:

	Depends: kaffe | sablevm | java1-runtime

When I see an entry like that, I expect that the maintainer tested this
package with kaffe and sablevm jvms so if the package doesn't work w/
one of these anymore - it's a regression.  In such case, as a JVM
mantainer I'd want to hear about this regression.  And I guess it might
also be important for the maintainer of the package in question, as
strange things happen sometime, like latest gjdoc, which works *only*
with kaffe, while it used to work w/ others (but Arnaud handled it the
right way).

On the other hand, as a user, I treat putting names of alternative
packages into Depends field as an important information, which will help
me in case I had troubles running this software.  Maybe I should switch
to different JVM to run this app?

Summarizing: in both cases - for an enduser and for jvm maintainer
it is *worth* to have Depends: field describing what JVMs are expected
to work.

So please, include this information if you can,

				Grzegorz B. Prokopski

PS: And, obviously, if it hasn't been tested and confirmed that
a package works with some JVM, such JVM should not be explicitely listed
in the Depends: field.

-- 
Grzegorz B. Prokopski <gadek@debian.org>
Debian GNU/Linux      http://www.debian.org
SableVM - LGPLed JVM  http://www.sablevm.org
Why SableVM ?!?       http://devel.sablevm.org/wiki/WhySableVM



Reply to: