Bob,
This makes sense to me now from what I have been seeing. I when I use an  
XPointer it comes along with the ID. If I use the XInclude only, there is no  
need for an ID in my files so I did not put one in. That explains why it is 
 happening in my system. It really had nothing to do with the XPointer 
issue on  my system. That I see now. 
 
So now that I know what is happening, I am more than fine with this  
version. I can re-org my document to avoid most of the warnings if I get tired  
of 
seeing them (except the bookinfo one).
 
Thanks for helping me through this Bob. 
 
Regards,
Dean Nelson
 
 
In a message dated 10/30/2015 2:33:42 P.M. Pacific Daylight Time,  
b...@sagehill.net writes:

Hi  Dean,
I'm a little confused about what is going on here.  The internal  
processing steps are:

1.  The XML document is  parsed.

2.  Any <xi:include> elements are replaced with the  content.

3.  The XSL stylesheet processes the result into  fo.

4.  FOP processes the fo to pdf.

So at steps 3 and 4,  the xi:include and its attributes not longer exist, 
having been replaced  by the content pointed to.  So neither docbook-xsl 
nor FOP even see  the xpointer attribute.

The changes to fox:destination came from a bug  report that indicated 
that many elements were not generating  fox:destination, so links into 
the document were not working.  When I  looked, the stylesheet was 
outputing fox:destination for a select group of  elements, the hierarchy 
elements like book, chapter, section.  That  left out figure, table, and 
a lot of other potential targets for  cross-document links.

So I changed the template in mode="fop1.foxdest"  to recurse through the 
resolved document and generate a  <fox:destination> for every xml:id in 
the document.  That  certainly didn't leave any out, but now I realize 
there are some elements  whose template may not copy their id to the FO 
output, possibly because  the element is not actually output (such as an 
info element).  So you  get a fox:destination pointing to a non-existent 
id.  These are  harmless, but do generate the warning noise.

It would be hard for the  template with mode="fop1.foxdest" to 
distinguish between these two kinds  of elements, because in its context 
it would not know if a given source  element was generating output.

I've never been clear why FOP needs  fox:destination for every id. Other 
FO processors don't seem to need such  help identifying ids in the PDF.

Bob Stayton
Sagehill  Enterprises
b...@sagehill.net

On 10/30/2015 12:13 PM,  deannel...@aol.com wrote:
> Thanks for the input Bob.
> I do want  to emphasize that these are only FOP warnings and not errors.
> In  looking at the FO output, 1.79.0 puts out a list of
> "fox:destination"s  that have the IDs of each link (this is not in 1.78.1
> output). Some of  these are resolved but others issue a warning:
> [WARN] FOUserAgent -  Destination: Unresolved ID reference "doc.bookinfo"
> found.
>  [WARN] FOUserAgent - Destination: Unresolved ID reference  "met-company"
> found.
> [WARN] FOUserAgent - Destination:  Unresolved ID reference "met-author"
> found.
> [WARN] FOUserAgent  - Destination: Unresolved ID reference "met-address"
> found.
>  [WARN] FOUserAgent - Destination: Unresolved ID reference
>  "met-copyright" found.
> These are IDs of XIncluded files with  XPointers
> If I change these xincluded/xpointers files to simple  Xincludes then the
> IDs are resolved. The ones that are NOT resolved in  my test files are:
> bookinfo
> corpauthor
> author
>  address
> copyright
> I see that there are significant changes in  fop1.xsl and many of them
> deal with fox:destinations. I wonder if my  stylesheets assume way too
> much about what elements are able to be  included in XPointers?
> Do the stylesheets preserve the ID of XPointers  in its output?
> Regards,
> Dean
> In a message dated  10/30/2015 11:03:15 A.M. Pacific Daylight Time,
> b...@sagehill.net  writes:
>
>     Hi Dean,
>      I would suggest breaking down the process into steps and examining  
the
>     output of each step.
>
>   So I would run xmllint --xinclude on the file to generate a  document
>     with XIncludes resolved, in which you can  look for the id in 
question.
>     Then generating a  temporary .fo file using your stylesheet, and
>      checking
>     for the id.  And then passing that to  FOP to make the PDF.  That will
>     show which  stage is producing the error.
>
>     Bob  Stayton
>     Sagehill Enterprises
>   b...@sagehill.net
>
>     On  10/29/2015 9:43 AM, deannel...@aol.com wrote:
>      >  Bob
>      > Last night I ran the 1.79.0 load with my  system (xsltproc, DB
>     4.5, fop 1.1)
>   > For the most part, everything works as expected. However, I  am now
>      > getting a FOP warning when I use an  XPointer. It looks like it
>     adds it
>   > to the list of FOP destinations, but does not use it. Hence  :
>      >   <xi:include  xmlns:xi="http://www.w3.org/2001/XInclude";
>      >  href="../metriguard/loc-corpinfo.xml"  xpointer="met-company"/>
>      > and
>   >    <corpauthor  id="met-company">&CORP_TITLE;</corpauthor>
>     > produces this warning:
>      > [WARN]  FOUserAgent - Destination: Unresolved ID reference
>      "met-company"
>      > found.
>     > The FOP process produces a correct document, it just  complains.
>     Here is
>      >  the rub and why I am concerned: it works just fine with the  stock
>      > stylesheets, no warnings and also with  the 1.78.1 sytylesheets
>     WITH my
>   > customizations. So this has me a little stumped. I also  noticed
>     in the
>      >  release notes that there was some work in this area. Somehow, the 
 new
>      > stylesheet allow FOP destinations that  should not be there?
>      > Since it works with the  stock stylesheets, I am wondering how I 
can
>      >  troubleshoot this in my own stylesheets and am asking for some  
input.
>      > Any clues or ideas?
>   > Thanks
>      > Dean  Nelson
>      > In a message dated 10/21/2015 3:35:40  P.M. Pacific Daylight Time,
>      > b...@sagehill.net  writes:
>      >
>       >    It has been about a week since candidate DocBook XSL 1.79.0  
was
>      >     released, and it has  been downloaded about 150 times.  Has
>     anyone  who
>      >     has tried it run into  any problems yet?  If so, please let me
>     know  so
>      >    they can be fixed for the  1.79.1 release.
>      >
>       >    --
>      >     Bob  Stayton
>      >    Sagehill  Enterprises
>      > b...@sagehill.net
>   >
>      >  
---------------------------------------------------------------------
>   >    To unsubscribe, e-mail:
>   docbook-apps-unsubscr...@lists.oasis-open.org
>   >     For additional commands,  e-mail:
>      docbook-apps-h...@lists.oasis-open.org
>       >
>

Reply via email to