RFC[02] - nZyme C++ code segment

George Makrydakis gmakmail at gmail.com
Mon Apr 3 16:45:26 PDT 2006


George Makrydakis wrote:
> On 3/25/06, Jeremy Huntwork  wrote:
> 
>> Looks good so far, George. Very thorough. 
<snip>

Hello again guys. I know you are all very busy with jhalfs-X release and 
I am not top - posting again about nzyme till I present the working code 
set as it is getting finalized.

> Once my home boxes are all up again I will provide you with a more up
> - to - date version including entity dereferencing and a primitive
> bash output "filter" (works here, did not post yet because I am
> thinking it over).

Systems back online, got some spare time to work on nzyme and enhance it :)

A small public update:

nzyme is _completely_ ported to an OOP structure; I also got rid of the 
necessity of loading the entire file in memory as with the POC code 
segments and other minor POC deficiencies. I managed to create a set of 
classes that can be reused for parsing any XML document with the 
concepts laid out in the list so far. With the current set the goal is 
to treat each XML document in such a way that it is made up itself of an 
XML doctype object and a series of XML element objects. It is possible 
to recursively load different chunks  who belong to different XML files, 
like with any multi - file XML document parser. This is a must feature 
for CLFS - like xml structures. The resulting classes do not parse XML 
neither in SAX nor in DOM compliant ways, but what you get out of them 
respects data hierarchy without the need for buffers, tree structures or 
anything else that is step - intensive.

Everything has been debugged and I am also embedding documentation - 
like comments in the final code. I will be posting with attachments so 
that we avoid cut / paste phenomena on critical code sections :)

So far, the ultimate goal is to reduce the main parsing class in the 
entire set into a single - transaction process, reducing time and loop 
steps. Once you have such a representation say in a single nzyme.h file 
on how to manipulate XML documents through a set of straight - forward 
classes, it becomes easy to create client files that remove XSL dependence.

Thank you for reading this message, it just to inform everyone that work 
is proceeding as planned. I hope to wrap the entire work in a single(?) 
C++ header file by the end of this week together with a sample "client" 
file demoing the use of the classes. I hope that my work helps you take 
care of the XML processing part pretty soon, if you still wish to 
proceed with a C++ based tool for building lfs - like projects without 
relying on _big_ third party parsing solutions.

George Makrydakis

PS: jhalfs-X is excellent work.







More information about the alfs-discuss mailing list