[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

Re: [xsl] Generating an internal subset?


Subject: Re: [xsl] Generating an internal subset?
From: dvint@xxxxxxxxx
Date: Wed, 30 Jun 2010 14:17:50 -0700

>> Just wondering if there is a better way to do this.
>
> That's best you can do if you really need an internal subset.
> (using pure xslt, saxon has an extension to do this I believe)
> Better perhaps is use
> <!DOCTYPE dmodule SYSTEM "local.dtd">
>
> which you can generate using <xsl:output doctype-system="local.dtd"/>
>
> then you just haave to write local.dtd to look like (in total)
>
> 	<!ENTITY % ISOEntities PUBLIC "ISO 8879-1986//ENTITIES ISO Character
> Entities 20030531//EN//XML"
>   "http://www.s1000d.org/S1000D_3-0/ent/xml/ISOEntities">
> 	%ISOEntities;
>
>
>
>
> Note though that if you do this (or what you suggested) any xml parser
> using the file that you generate will (by default) need a network
> connection to that server (or to the w3c server if you use the current
> versions)

Yeah I have that problem when I try to process the files from the system
I'm trying to mimic. this may be a holdover from the SGML version of these
schemas for s1000D. Everything is managed by a CMS/CSDB and I'm not sure
what assumptions it makes, so I didn't want to loose the definition. My
editing environment uses XML catalogs and such to handle these files
locally, but my XSLT processing is hanging up on them.


>
>
> If the file is being generated by XSLT, unless you go to a lot of
> trouble with disable-output-escaping, it will nothave any entity
> references in it, so you should not need to define any entities, so why
> do you need to refer to these entity definitions?
>

Mainly so I don't have to figure out any dependencies in the commercial
CMS. This is what their templates generate. Maybe I will go ahead and file
a bug report with them. the editor we use can actually generate UTF-8,
numerical character entities or these named entities, so this may be the
way the CMS tries to handle the different combinations.

this was also an interesting exeercise as I hadn't thought about creating
an internal subset like this and had previously done this as you suggested
with the doctype-system parameter on <xsl:output>.

After sending this question, I came across the need to extract references
in the XML file to entity references and generate the following:

<!ENTITY RQ4-B-VTC-ECS1-CANCEL SYSTEM
"q:\csdb\graphics\library\vtc\ecs1\rq4-b-vtc-ecs1-cancel.cgm" NDATA cgm>
<!ENTITY RQ4-B-VTC-ECS1-ECSFANSFWD SYSTEM
"q:\csdb\graphics\library\vtc\ecs1\rq4-b-vtc-ecs1-ecsfansfwd.cgm" NDATA
cgm>
<!ENTITY RQ4-B-VTC-ECS1-MANCNTRL SYSTEM
"q:\csdb\graphics\library\vtc\ecs1\rq4-b-vtc-ecs1-mancntrl.cgm" NDATA cgm>
<!ENTITY RQ4-B-VTC-MAINT-ECS1 SYSTEM
"q:\csdb\graphics\library\vtc\maint\rq4-b-vtc-maint-ecs1.cgm" NDATA cgm>

So this actually prevent me from creating an external DTD as you suggested
as this list of entites is unique to a particular file and I can't
anticipate what the complete list might be.

..dan


Current Thread
Keywords