From: Andrea A. <aa...@op...> - 2008-05-06 19:47:59
|
Jody Garnett ha scritto: > Andrea Aime wrote: >> Jody Garnett ha scritto: >>> Hi Andrea; there are new "official" string functions in town - the >>> work Eclesia has been doing on SE1.1 has brought them to light. It >>> looks like this allows for some functions with a variable number of >>> parameters. So when you do boil this down to a single Expression we >>> may have a better function for you to use. >> >> I'll need to see some examples. Anyways, the current code is already >> parsing the mixed content >> >>> As for the meaning of the <Lable>....</Label> section I think we >>> would all like to the behave like a mini template; it output exactly >>> what is between the tags substituting <PropertyName> tags where >>> needed. Can you confirm for me that other expressions can be used >>> here - or is it just PropertyName? >> No no, whatever expression you want. > Good - thanks for the info. >> The only issue is that spaces are being eaten. Imho to make it a good >> mini template languages we'd have to preserve spaces instead. > Agreed; that sounds like a good move all around. Do you *exactly* have > to preserve whitespace or can we gobble up all \n\t characters as a > single ' '? ie. what is important here... One of my issues says people are trying to make up multiline labels, we'll need to preserve newlines in order to have those >> As an alternative, I can roll in a function that behaves like a real >> template language, freemarker. I have an experiment in a geoserver >> community module that looks more or less like this: >> >> <ogc:Function name="freemarker"> >> <ogc:Literal><![CDATA[<html><body> >> This is bridge <b>${FID.value}</b> whose name is <i>${NAME.value}</i> >> </body></html>]]></ogc:Literal> >> </ogc:Function> > That really does not look like a reusable style :-) Why not? > But you are onto > something, the use of a CDATA section to preserve a String, could that > be used to capture whitespace of the literals? I tried, it did not work, the parser does not make a difference between cdata sections and mixed content apparently. Cheers Andrea |