Re: [docbook-apps] Re: [docbook] Re: unable to convert DocBook to PDF

[Jeffrey Walton <noloader@gmail.com>]

Hi Chris,

Thanks for your help.

> You could use an XInclude stylesheet with xmllint[1] (which is
> most likely installed in your Ubuntu distribution)
You were right. A quick grep of apt-cache turned up nothing, but it
was installed (presumably under a tools package name?).

>> You can change your toolchain, modify your toolchain to preprocess the
>> documents (or, equivalently, change your documents), or give up.  Oh, or
>> dive into modifying OpenJade — which is legendarily underdocumented.
This hack worked. Its kind of ugly, but it got me to where I needed to
be until I learn how to do things the DocBook5 way.

$ xmllint --xinclude book.xml > book-pp.xml ; docbook2pdf book-pp.xml
; mv book-pp.pdf book.pdf ; rm -rf book-pp.xml book-pp.tex

(posted for others who might stumble upon the thread).

Jeff

On Sat, Oct 22, 2011 at 10:39 PM, Jeffrey Walton <noloader@gmail.com> wrote:
> On Sat, Oct 22, 2011 at 10:25 PM, Christopher R. Maden <crism@maden.org> wrote:
>> On 10/22/2011 10:08 PM, Jeffrey Walton wrote:
>>> OK. You've already lost me (my apologies).
>>
>> Short version: you’re using a toolchain based on 1996 standards,
>> processing documents based on 2006 standards.
> :)
>
> I ordered "DocBook 5: The Definitive Guide" on Thursday (I try to
> support authors). Hopefully my dumb questions will end soon.
>
>>> OK. I think the prudent thing to do is file a bug report (and accept
>>> the flames) against OpenJade or docbook2pdf.
>>
>> OpenJade has not been updated since 2003.  Its underlying technology,
>> DSSSL, has one main mailing list, which has had a couple of dozen posts
>> in the last ten years.  I would not hold my breath.
>>
>>> As a dumb user, I just want things to work on a simple example file.
>>> My thinking is that word processors 'just work', and this is not much
>>> different (though I'm not total sure what 'this' is). I think its XML
>>> based publishing.
>>>
>>> Its OK if folks want to abstract and layer things. Unfortunately, the
>>> design and implementation cleverness does not work out of the box in
>>> this instance. I don't think I asking for too much here - C/C++
>>> beginners expect their "Hello World" programs to work out of the box
>>> too (if they have their tools installed correctly).
>>
>> I concur with your expectations.  However, throwing C99 code against a
>> strict ANSI C compiler might not work.
>>
>> Actually, there’s a very useful metaphor here: XInclude is like C
>> preprocessor instructions, which DocBook is like C itself.  If you had a
>> C compiler that didn’t understand the preprocessor directives, it would
>> choke and die on normal C code; OpenJade was last updated three years
>> before the preprocessor directives, and completely fails to understand
>> them.  Running an XInclude normalization (like using xmllint) is
>> equivalent to preprocessing the files.
> OK. So something like the following should work (or at least move me forward):
>
>    xmllint book.xml | docbook2pdf > book.pdf
>
> (I think I need to hunt down xmllint, so I can't test it at the momnet).
>
>> To be fair, the O’Reilly files should not be labeled as being DocBook
>> files; they are DocBook + XInclude files, combining two distinct
>> dialects of XML.  OpenJade and the DocBook DSSSL stylesheets only
>> understand the DocBook half, and are bewildered by the XInclude part.
> OK, thanks.
>
>> You can change your toolchain, modify your toolchain to preprocess the
>> documents (or, equivalently, change your documents), or give up.  Oh, or
>> dive into modifying OpenJade — which is legendarily underdocumented.
> I'd like another option, please.