Re: [xsl] Form feed character () in decoded xs:base64Binary

["Imsieke, Gerrit, le-tex gerrit.imsieke@xxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>]

On 13.07.2020 20:55, Martynas JuseviD
ius martynas@xxxxxxxxxxxxx wrote:
> Ah, sorry :) I get it now.
>
> After I make the change:
>
> <?xml version="1.1" encoding="UTF-8"?>
> <!DOCTYPE xsl:stylesheet [
>      <!ENTITY rdf    "http://www.w3.org/1999/02/22-rdf-syntax-ns#";>
>      <!ENTITY rdfs   "http://www.w3.org/2000/01/rdf-schema#";>
>      <!ENTITY xsd    "http://www.w3.org/2001/XMLSchema#";>
>      <!ENTITY dct    "http://purl.org/dc/terms/";>
>      <!ENTITY skos   "http://www.w3.org/2004/02/skos/core#";>
> ]>
> <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
> version="3.0"
> ...
>
> I start getting this error:
>
> Error on line 248 column 46 of messages2trix.xsl:
>    SXXP0003  Error reported by XML parser: The entity "xsd" was
> referenced, but not declared.
> org.xml.sax.SAXParseException; systemId: file:/.../messages2trix.xsl;
> lineNumber: 248; columnNumber: 46; The entity "xsd" was referenced,
> but not declared.
>
> Line 248 contains "&xsd;dateTime".

I donbt know about any limitations of XML 1.1 parsers wrt plain 
character entities declared in the internal subset. There shouldnbt be 
any, according to my reading of the spec.

Can you extract only the single template that is responsible for base64 
decoding and LF stripping into a separate XSLT document (XML 1.1) that 
need not refer to any entities? Then xsl:include this document from the 
XML 1.0 XSLT document that contains what used to be line 248.

Gerrit

> On Mon, Jul 13, 2020 at 8:44 PM Imsieke, Gerrit, le-tex
> gerrit.imsieke@xxxxxxxxx <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
> wrote:
> > I was suggesting that you prepend <?xml version="1.1"?> to your
> > stylesheet document, hoping that you are then able to apply translate(.,
> > '&#xc;', '') to the decoded string.
> >
> > On 13.07.2020 20:26, Martynas JuseviD
ius martynas@xxxxxxxxxxxxx wrote:
> > > With xsl:output version="1.1", the form feed is not a problem - Saxon
> > > writes the decoded xs:base64Binary string without any replacements.
> > >
> > > However I'm getting weird parsing errors downstream in my RDF toolkit
> > > (which works fine with XML 1.0). I'll try to see what the problem is.
> > >
> > > On Mon, Jul 13, 2020 at 8:00 PM Imsieke, Gerrit, le-tex
> > > gerrit.imsieke@xxxxxxxxx <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
> > > wrote:
> > > > What happens if you use version="1.1" in the XML declaration (of the
> > > > stylesheet)?
> > > >
> > > > On 13.07.2020 19:54, Martynas JuseviD
ius martynas@xxxxxxxxxxxxx wrote:
> > > > > Hi,
> > > > >
> > > > > I'm transforming large JSON files with some email data using XSLT 3.0.
> > > > > They contain xs:base64Binary literals which I'm decoding using
> > > > > bin:decode-string() and want to include the decoded values in the
> > > > > output XML.
> > > > >
> > > > > The problem is that some of the decoded string values have illegal XML
> > > > > 1.0 characters in them, such as Form feed (&#xc;).
> > > > >
> > > > > I want to remove them but cannot find a way.
> > > > > I can't use translate(., '&#xc;', '') because the stylesheet would not
> > > > > be well-formed anymore.
> > > > > I can't even use replace(., codepoints-to-string(12), '') because I
> > > > > get this error (with Saxon 10.1 EE):
> > > > >
> > > > >        codepoints-to-string(): invalid XML character [xc]. Found while
> > > > > atomizing the second argument of fn:replace()
> > > > >
> > > > > Are there any native XSLT options here?
> > > > >
> > > > > Thanks.