Re: Reading the specs raw from Github

Thanks Norm!

Serving it as text/plain does indeed seem to be the problem. Ideas for
workarounds here are welcome.

As for your other observation - that the document is supposed to be
completely flat, this is what my actual XProc has --

<p:for-each name="additional-steps">
      <p:with-input select="/steps/string(.)">
         <p:inline>
            <steps>ixml</steps>
            <steps>mail</steps>
            <steps>file</steps>
            <steps>os</steps>
            <steps>paged-media</steps>
            <!--<steps>rdf</steps>-->
            <steps>run</steps>
            <steps>text</steps>
            <steps>validation</steps>
         </p:inline>

Those strings are picked up and links are constructed. Each one gets its
own p:load. Everything works when there is no XLink, and breaks when there
is.

Thanks! Wendell


On Wed, Jul 2, 2025 at 1:43 PM Norm Tovey-Walsh <ndw@nwalsh.com> wrote:

> A few quick thoughts.
>
> Wendell Piez <wapiez@wendellpiez.com> writes:
> > Probably I am not the only one who can tell from "side effects" when the
> committee has been busy at work updating and improving things.
>
> Cool.
>
> > For example, recently a pipeline of mine broke ... apparently because
> relevant pieces of the specs in the 3.0-steps repository now use XInclude.
>
> It surprises me that you’ve only just encountered that. That’s been true
> for ages.
>
> >     <p:load
> > href="
> https://raw.githubusercontent.com/xproc/3.0-steps/master/step-text/src/main/xml/specification.xml
> "
> >
> >       content-type="application/xml"/>
>
>
> That document is supposed to be completely flat, so if there’s an errant
> XInclude in there, I think that’s a build problem.
>
> > This returns an error in both XML Calabash and Morgana IIIse.
>
> I shall have to look at the error to have better insight.
>
> > It returns a different error related to the same cause if you leave
> > 'content-type="application/xml"' from the p:load. (I blame Github.)
> >
> > What is up with this? Is there any way to read the raw XML off Github,
> as XML?
>
> Probably not, the seem to serve it as text/plain.
>
> > If not, are there alternative pathways -- short of cloning the
> repository and reading via file: not
> > http: ? i.e. alternative all-XProc approaches to reading the specs
> dynamically in place.
>
> What your doing should be possible. I’ll see where the error leads me when
> I have a chance to check.
>
>                                         Be seeing you,
>                                           norm
>
> --
> Norm Tovey-Walsh <ndw@nwalsh.com>
> https://norm.tovey-walsh.com/
>
> > Sarchasm: The gulf between the author of sarcastic wit and the person
> > who doesn't get it.
>


-- 
...Wendell Piez... ...wendell -at- nist -dot- gov...
...wendellpiez.com... ...pellucidliterature.org... ...pausepress.org...
...github.com/wendellpiez... ...gitlab.coko.foundation/wendell...