cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Portier <...@outerthought.org>
Subject Re: [NEW] webapp config with xconf tool
Date Fri, 23 May 2003 11:49:20 GMT
Geoff Howard wrote:
> At 12:53 AM 5/23/2003, David wrote:
> 
>> >    This would be ignored in non
>> >    xerces parsers, but would only cause a UnknownHostException when
>> >    the parser attempts to retrieve the DTD from a file being patched
>> >    (from web.xml, for example).  If needed, other parser 
>> implementations
>> >    (keep in mind this is only during build) may have similar 
>> mechanisms.
>>
>> So would any other parsers ever be used during the build?
>> If not, then we can remove the xmlcatalog stuff from the
>> XConfToolTask.java ... less code, easier to maintain.
>> Anyway, it seems that that code is not actually doing local
>> DTD resolution.
> 
> 
> It's a good bet that xerces is the only parser that would
> ever get used during the build, but as soon as we make that
> assumption Murphy's Law will kick in.  I'm not sure it matters
> though since we can't redistribute that dtd and the whole point
> of this was to add convenience to the build for the uninitiated
> user.
> 
> The reason I left the xmlcatalog code in was that I could
> imagine it being useful hypothetically but wasn't quite sure.
> The "magic" xerces parameter only applies to dtd resolving.
> Of course, I'm having trouble thinking of concrete examples of
> entities that would need to be resolved from the outside world
> in the context of the patch tool (except the web.xml dtd).
> 
> So, if no one comes up with an argument in favor of xpatch using
> xmlcatalog, I'll take that stuff out.
> 

makes sense

> I think I also should turn entity expansion again, as I suppose that
> could be useful and there is really no reason to turn it off AFAICS.
> 

mmm, kind of depends on how you look at it:

just as with the default attribute values the end result is in 
logical XML terms equivalent... (so why bother?)

only (since yesterday) I look at this patching task as being more 
at the lexical level, and I would greately appreciate (for 
debugging) the output to have the most as possible visible 
resemblance to the different originals (so also keep external 
references outside?)

(but I can't be a hardliner on this statement cause e.g. CDATA's 
will probably get replaced by escaped entity-chars anyway, and 
probably other stuff not coming to mind now)

AFAICS now it is a matter of 'taste', hoping someone can build a 
stronger case either way

> Geoff

Geoff, other question: can't seem to find examples/docos on how 
to encode the new supported actions with xpatch, care to guide?
(I started on wikifying my other stuff and want to add a ref, or 
even slide it in)

-marc=
-- 
Marc Portier                            http://outerthought.org/
Outerthought - Open Source, Java & XML Competence Support Center
Read my weblog at              http://radio.weblogs.com/0116284/
mpo@outerthought.org                              mpo@apache.org


Mime
View raw message