Commit Graph

3 Commits

Author SHA1 Message Date
fpeters 7d7f350fdb properly install things in the right places
python setup.py install --prefix /usr/local/ \
	--install-lib /usr/local/lib/expression/modules \
	--install-data /usr/local/share/
2004-06-15 07:39:16 +00:00
eraviart 9a13c2d3fe Changed the way descriptions & schemas are found.
Instead of using an heuristic to find them in the current directory or in
the modules or in the application or..., it now uses absolute URIs and
catalogs.
2004-06-11 15:17:37 +00:00
eraviart 42a4a98f59 Added XHTML DTDs & entities as XML catalogs. These catalogs are handled the
Debian way <http://debian-xml-sgml.alioth.debian.org/>.


Q: Why do you use manage XML catalogs the standard way instead of libxml2 way?
A: Because the Debian way is the standard (Oasis) way and it is partially
incompatible with libxml2 way.

Q: Why do you use XML catalogs?
A: Because I needed to store XHTML entities definitions in the local
filesystem.

Q: Why do you need to store XHTML DTDs & entities definitions in the local
filesystem?
A: For speed reasons. Otherwise calls to libxml2 function xmlReadDoc with
option XML_PARSE_DTDLOAD require HTTP calls to get the DTD and its entities.

Q: Why do you need to use xmlReadDoc function with option XML_PARSE_DTDLOAD?
A: Otherwise, the parsing of XHTML files fails when the files contain
entities like &eacute;, etc.

Q: Why are entities definitions needed? You don't need entities substitution
when parsing XML & XHTML files.
A: Because after spending several hours testing and reading libxml2
documentation, I discovered that even when not substituting entities,
libxml2 needs to know their definitions (the reason is it stores both
entities declarations and substitutions inside XML tree. See
http://xmlsoft.org/entities.html).

Q: Why do you use xmlReadDoc instead of xmlParseDoc?
A: Because after spending several hours testing and reading libxml2
documentation, I discovered that xmlParse functions are deprecated.

Q: Why don't you use htmlParseDoc (or htmlReadDoc) anymore to read XHTML
files? They worked well with entities without requiring DTDs.
A: Because html(Parse|Read)Doc ignore namespace declarations.

Q: Why don't you ignore namespace declarations when parsing XHTML files?
A: Because XHTML files are XML files. And we need to differentiate XML tags
with and without namespaces.

Q: Why did you repair XHTML namespaces handling today.
A: Because Fred asked for it yesterday: "je crée un fichier xml, je ne
spécifie pas de namespace, pq l'attribution à l'holder et à l'élément
du namespace xhtml ?".

Q: Did you spent the whole day programming this small feature request?
A: No, I also wrote this FAQ.
2004-05-23 23:13:05 +00:00