From: Andrey Melnikov (melnikov_at_[hidden])
Date: 2005-08-16 08:17:05
Vladimir Prus wrote:
> On Wednesday 13 July 2005 15:38, greg_sylvain wrote:
>>Thanks for the reply. Sorry for the mis-direct posting.
>>LEIF's xsd2cpp tool generates a C++ class for each xsd:complexType
>>in the XSD schema. Accordingly, you can get a lot of classes from a
>>single schema file. All of the class files have a respective '.h'
>>header file for a respective implentation file ('.cpp').
>>In addition, the tool can take a configuration file that can re-map
>>any complexType element names to the respective C++ class name you
>>desire. For instance, all of our complexTypes are spelled out in
>>lower case and underscores, while our C++ class names use Hungarian
>>notation of mixed case class names. That it, the XSD complexType
>>ruleset_directive is remaped to become the C++ class RulesetDirective.
> This complicates matters quite a bit!
> At the moment:
> 1. Boost.Build should know the list of sources to compile before running the
> first command.
> 2. Boost.Build cannot extract anything from source files.
Do you mean that dependency scanners like boost-book.xinclude-scanner
won't help in this case?
> The best solution I have in mind is to build Boost.Jam with Python support,
> and write Python code that will scan the .xml file and extract the names from
> there. Would such a solution acceptable for you?
Is it possible to use BB native regular expressions in this case?
Regexps are not a conformant XML parser but they are TSTTCPW.
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk