[xml/sgml-pkgs] Bug#132989: how to add encodings

Ardo van Rangelrooij ardo at ardolabs.com
Sat Aug 12 04:48:19 UTC 2006

Joey Hess (joeyh at debian.org) wrote:
> Adding encodings seems pretty straightforward:
> 1. Go to ftp://ftp.unicode.org/Public/MAPPINGS/ and download
>    the files.
> 2. Run make_encmap from libxml-encoding-perl to generate xml file.
> 3. Edit xml file, adding expat="yes" attribute to the encmap tag.
> 4. Run compile_encoding from libxml-encoding-perl to generate enc file.
> I've done this for 8859-1, 8859-6, 8859-10, 8859-11, 8859-13, 8859-14,
> 8859-15, and CP1251[1]. 8859-13 is not available at that unicode.org
> site. This procedure appears to be correct; I get a file that is
> identical to the one I sent to #380426 for 8859-15. I'll attach a
> tarball of the new .xml and .enc files to this message.
> I don't really understand why libxml-parser-perl contains the
> precompiled encoding files instead of build-depending on
> libxml-encoding-perl and using its tools to build them. However, keeping
> things the way they are now while adding encodings is tricky since it
> would mean patching binary .enc files into the libxml-parser-perl source
> package.
> So I can see four alternatives:
> 1. Modify libxml-encoding-perl to not ship compile_encoding, make_encmap
>    and the xml files in /usr/share/doc/libxml-encoding-perl/examples,
>    but instead put the programs in /usr/bin and the xml files somewhere
>    in /usr/share where other packages can reasonably access them.
>    Then make libxml-parser-perl build depend on it and build the enc
>    files as part of its build process.
> 2. Introduce a new package that contains encodings not present in
>    libxml-parser-perl.
> 3. Get the missing encodings added to XML::Parser upstream.
> 4. Include the compiled encoding files in libxml-parser-perl using
>    uuencode or the like, and include the updated xml files in
>    libxml-encoding-perl for complteness.
> If I were maintaining these packages, I'd go for option 1. If I were
> NMUing[2], I'd probably go for option 4. I've also attached patches that
> implement option 4.


Thanks for looking into and for the suggestions.  This is really helpful. :-)

I'll be working on this over the weekend.  Thanks for the NMU offer.

Ardo van Rangelrooij                                     Debian XML/SGML Group
<ardo at debian.org>              <debian-xml-sgml-devel at lists.alioth.debian.org>
http://people.debian.org/~ardo/      http://debian-xml-sgml.alioth.debian.org/

More information about the debian-xml-sgml-pkgs mailing list