W3C home > Mailing lists > Public > xproc-dev@w3.org > May 2009

Re: xproc as an alternative to Apache ant

From: James Sulak <jsulak@gmail.com>
Date: Wed, 27 May 2009 08:36:26 -0500
Message-ID: <7cb78b3b0905270636o2a7a172bje89568967061e082@mail.gmail.com>
To: Dave Pawson <dave.pawson@gmail.com>
Cc: xproc-dev@w3.org
On Mon, May 25, 2009 at 11:31 AM, Dave Pawson <dave.pawson@gmail.com> wrote:
>
> What about using them?
> IMHO the xslt model is ideal?
> defined in xml, used as $variableName
>
> Then when I import (or bring into the pipeline) the variables,
> I can resolve one against the earlier ones?
>
> <var name='x' value='$x/a/b'/>

You can refer to variables and options in xpath expressions, so you
could do this:

<p:variable name="$foobar" select="concat($foo, $bar)" />

using as many variables or options as you want. But you have to
predefine your 'compound' variables in your pipeline; you can't do it
dynamically at runtime.  Is that what you're talking about?

On a related note:  It would be nice if there were the equivalent of
XSLT attribute value templates, so we could avoid the verbose
"p:with-option" language.  For example, instead of this:

<p:store>
  <p:with-option name="href" select="concat($base_dir, '/',
$output_dir, '/', $output_file_name)" />
</p:store>

this:

<p:store href="{concat($base_dir, '/', $output_dir, '/', $output_file_name)}" />

I assume that it was considered and discarded for some good reason,
but I'm curious to know what it is.  AVTs would make pipelines much
less verbose.

-James
Received on Wednesday, 27 May 2009 13:37:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 May 2009 13:37:03 GMT