[xml] Parsing XML in embedded environment

[David Kubicek <foceni gmail com>]

Hello,

I've been struggling with this for two whole days. I have read the whole 
of libxml docs and the library source, but I'm not any further.

We're writing an app for a small resource-limited device. It parses XML 
files. We don't need validation or any extras. Just get at the contents. 
The problem is that by default, libxml knows only the basic 5 XML 
entities. We cannot use the library's default file system DTD "catalog", 
because basically, there's no filesystem like that.

So I'm looking at how to implement this dynamically. Everything has to 
be dynamic, no file or even network access. We have to keep it as simple 
& fast as possible, so we're not using callbacks like SAX interface has, 
etc. We simply call "*xmlReadMemory*" and then recursively loop through 
the returned "*xmlDoc*" and get what we need. The problem is that libxml 
doesn't know our custom entities, so the parsed text is not correct. 
I've been hitting the wall ever since. I just can't find (or even think 
how to add it by modifying the library) any way of supplying libxml with 
the extra entities. Be it programmatically, via a function defining new 
entities, or by letting libxml pre-load our DTD that contains all the 
definitions (that would be the preferred way).

I've tried "xmlAddDocEntity(xmlDocPtr,...)", but that's obviously 
available only *after* the document is created/parsed and that's too 
late. We get "undefined entity" errors *during* parsing. Somehow, I need 
to setup libxml with a DTD, then call xmlReadDoc/Memory, and finally 
walk the returned tree and extract the tags we need. I'm sure I must be 
overlooking something, because this is a reasonable requirement - many 
apps can't access the file system and/or the network to allow libxml 
on-the-fly DTD access based on PUBLIC ID's and external catalogs.

Can somebody here help us solve this?

Thank you very much,

--
Dave Lister