Re: jade catalog problems for Cygnus DocBook 4.1 XML

[Eric Bischoff <e.bischoff@noos.fr>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 3153 bytes --]

"Alan W. Irwin" wrote:
> 
> Eric, thanks for that earlier acknowledgment that you have at least seen my
> pdfjadetex bug report.  Since you have so much to do I will be patient, and
> I will try to help out as much as possible by what I hope are clear bug
> reports and examples. While we wait for you to get a chance to sort out
> these problems I have found, a subset of our team can build our
> documentation on Debian DocBook 4.1 XML instead.  But we are looking forward
> to when this will all work properly on Cygnus/RedHat as well.

Thanks for your patience. For the moment, I'm putting the pressure on
supporting Asian languages in printed output, and my TODO list is
growing ;-).

> This current post is a bug report for entities that Cygnus just cannot seem
> to find although they are defined within the catalog chain.  I believe these
> entity problems are completely independent of the pdfjadetex problem I
> reported before. I am only copying that other bug report below so you will
> know all the details of my system.  Sorry that I could not actually find a
> solution of these entity problems, but here are the symptoms for a simple
> example with egrave.
> 
> jade -c /etc/sgml/catalog -d
> /usr/share/sgml/docbook/dsssl-stylesheets-1.57/print/docbook.dsl -t tex
> /usr/share/sgml/xml.dcl hello_world.xml
> jade:hello_world.xml:6:26:E: general entity "egrave" not defined and no
> default entity

The cygnus packages come with a package named "docbook-utils", which
include scripts like "docbook2tex" (formerly db2tex) who make all this
catalog thingy transparent. Give them a try, it will avoid you building
complex command lines like this one.

I advice you *NOT* to use /etc/sgml/catalog. This is the "super-catalog"
that references indirectly *every* declared SGML file on you systems via
four levels of indirection. So it's both slow and is not guaranteed to
be bulletproof against declaration conflicts. It is meant as a "last
chance" solution.

> Perhaps there is some minor typo in this chain of catalogs, but I cannot
> spot it so I have no idea why egrave is undefined as far as jade is
> concerned.  This is just a wild guess, but perhaps jade cannot handle a
> four-level deep catalog chain as above?

It can ;-), but try using /etc/sgml/xml-docboook-4.1 to avoid
declaration conflicts.
Please also check that you have read access on every file on the chain.
If this does not work try declaring explicitely '-c
/usr/share/sgml/sgml-iso-entities-8879.1986/catalog" on the command line
and mail me privately.

> To test that wild guess I tried specifying two catalog entries as a
> colon-separated list, i.e.,
> jade -c \
> /etc/sgml/catalog:/usr/share/sgml/sgml-iso-entities-8879.1986/ISOlat1.ent \

The colon syntax works only for the SGML_CATALOG_FILES environment
variable. Use several -c arguments.

You're the first one to describe such a problem.
-- 
+----------------------------------------------------------------------+
| Éric Bischoff                              mailto:e.bischoff@noos.fr |
in case of problems: ebisch@cybercable.tm.fr
+----------------------------------------------------------------------+