Jericho HTML Parser

Hi Martin,

excellent, it works pretty fine, thank you so much.

> i.e. the tag <cb:message key='id007' text='Order it' /> would have getName()=="message".

Actually, it would have getName()=="cb:message" (just a note for another reader)....

> the isServerTag argument is set to true, which allows the tags to be found inside attributes of normal tags without deregistering the normal tag types.

A little bit confusing argument name, isn't it?

> This is quite messy as registering tags is a static function so they will remain registered for all Source instances. This might be a problem in a multithreaded environment.

Well, i think it is worse than a problem. Even in singlethreaded environment. ;-)

I need to parse several tens or hundreds of JSP files.

And I should register/deregister StartTagTypeJSP (and special StartTagTypeJspDirective, see another topic please) for every file. Or to remember the last prefix and register/deregister StartTagTypeJSP when the prefix was changed.

Are you going to remove the "static mess" ? For example to use the "Factory" design pattern?

I would preffer the code like this:

Map<String, SourceFactory> prefix2factory;

...
for every File:

SourceFactory fac = prefix2factory.get(prefixFound);
if (fac == null) {

  fac = SourceFactory.create();
  fac.registerTagType(new StartTagTypeJSP(prefixFound));
  prefix2factory.put(prefixFound, fac);
}

Source parser = fac.createSource(new BufferedReader(new FileReader(aFile)));

> Personally I think the use of xml-style JSP tags inside normal tag attributes is bad practice, even if it is quite common.

You are probably right.

But I think it is good idea to keep the business code and texts for an end user together, it is much easier for a developer. Those texts (in HTML tags) should be localised to several languages.

Do you have a better solution for this instead of JSP tags inside normal tag attributes? I would really appreciate your advice. :-)

Thank you again. Mike.