Re: [docbook] xi:include problem

[Alex Kleider <akleider@sonic.net>]

On 2013-12-12 23:50, davep wrote:
> On 12/12/13 19:04, Alex Kleider wrote:
>
> > The problem (already aluded to in the thread from which I've spun off
> > this one) with
> > a) using xmllint to expand xIncludes
> > and then
> > b) validating with jing,
> > is that the validation must be done on the output of xmllint and when
> > jing reports an error it provides a location in the output file, but 
> > we
> > want to fix the problem in the original file.
> >
> > The way I'm solving the problem is making each xIncluded file a valid
> > DocBook file and validating each included file before using xmllint 
> > and
> > then doing a final validation.  This restricts me to making each
> > xIncluded component an <article> since I think this is the lowest 
> > level
> > element that can be a valid DocBook.
> >
> >
> > Comments?  Are there better ways?
>
>
> Sounds sensible, though if you just want to 'stop' and see the errors,
> simply stop processing when (if) xmllint reports an error?
>
> I do agree about the intermediate file being a nuisance though.
>
>
> regards

I have figured out how to solve my problem.  It's probably 'old news' to 
most but ...

I use the following script:
"""
if [ -n "$2" ];
then
    outfile="$2";
else
    outfile="/dev/stdout";
fi
infile=$1

rng=/usr/share/xml/docbook/schema/rng/5.0/docbook.rng

xmllint  --relaxng $rng --xinclude $infile > $outfile
"""