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

Re: Classpath and versioned libraries for Debian-Java



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Max Kellermann <max.kellermann@epost.de> writes:
<snip>

> > I have a question to you all. Should we allow the same interface in different
> > packages? It seems to create strange conflicts but sometime it is necessary,
> > or? What do you all think?
> 
> That's also one thing I was thinking about. If we say we don't include DOM and
> SAX in Xerces, it's more clean IMHO, but much more work when packaging. We
> can't just call 'ant jar' to create the .JAR files, we have to create them
> ourselves. I think it's ok to have Xerces include DOM, although it's not the
> Best(tm) way.

I don't think it is just more clean... I think it is a requirement.

Specifically because I have seen this become a problem in the wild :)

A good example would be a few months back I wanted to run the latest version of
Xerces with Tomcat.

Tomcat used projectx.jar (which included an older version of DOM) and was loaded
first.  When I tried to load Xerces it wasn't loading its own internal version
of DOM since the verion with projectx.jar was already loaded.

Thus I was getting an Error stating that a new method couldn't be found.

I think that Java is starting to suffer from the quote (from Knuth???)

"Those who do not learn from UNIX are destined to repeat its mistakes"

> > Do you think the above description is covered in the proposed policy. It
> > should be but I'm not sure if it is clear enough.
> 
> Well, I think it's not fully covered, but it's possible with your proposed
> policy. It's not about replacing the policy, the policy is good, it's about
> adding to the policy.
> 
> > > - library F and library G may both provide the sub-library H, of the same
> > > version - they may both be installed, but not in the same classpath at the
> > > same time.  > > If they completely provide H, why can't they be in the same
> > > classpath? One > of them will override the other. Or am I wrong here?
> 
> Yes and no.
> 
> I most cases, e.g. the DOM interface classes, that's ok. They won't conflict
> because they're identical.

Not always... as above.  The point is that we could do it better. :)

> But look at the TRAX API, the TRAX classes included in Xalan call concrete
> Xalan classes in the TRAX factory classes (it has to, somehow). Another XSLT
> processor will call their own concrete classes in the TRAX. So TRAX of Xalan
> conflicts with TRAX of [fill in any other TRAX-implementing XSLT processor
> here]. The same with javax.xml in XML parsers. That's what I meant here.

OK.  Trax.... this is another one.
<snip>

- -- 
Kevin A. Burton ( burton@apache.org, burton@openprivacy.org, burtonator@acm.org )
             Location - San Francisco, CA, Cell - 415.595.9965
        Jabber - burtonator@jabber.org,  Web - http://relativity.yi.org/

How are you gentleman?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Get my public key at: http://relativity.yi.org/pgpkey.txt

iD8DBQE77ZG3AwM6xb2dfE0RArv/AKCTNklNF0wHInypCmO8mrfQWsVkRgCglBeO
PIG/UY5CQHS1Jh0SpjFdsPg=
=eNR4
-----END PGP SIGNATURE-----



Reply to: