Re: XmlSerialzier

> i have merged back XmlSerializer to main package and modified XmlPullParserFactort
> to allow creation of XmlSerializer. please take a look on it - i will test it when i finish
> implementing serializer in XPP3. i will also write some simple conformance tests.
> 
> please take a look on modified Roundtrip sample that shows how serializer API can be used
> and send comments about XmlSerializer API and any unresolved issues about XmlSerializer
> as it would be good to get serializer API ready for new XMLPULL API release soon.

Great! The only problem I can see is default namespace handling 
in getPrefix; especially, default namespaces do not apply to 
attributes.... any suggestions??

Best,
Stefan


> 
> thanks,
> 
> alek
> 
> ps. here is current list of issues with resolution and one proposal for new issue:
> 
> --- GENERAL
> 
> // ISSUE: specify exactly when XmlSerializer must be supported
>     PROPOSAL: in J2ME both XmlSerializer and factory is option, on J2SE both are required
> 
> // ISSUE: in XmlPullParserFactory get/setFeature should it be used for XmlSerializer as well?
>  PROPOSAL: get/setFeature in factory is only applied to newPullParser()
> 
> //ISSUE: how to write epilog?
>  RESOLVED: write comment, pi or white spaces before calling endDocument()
> 
> //ISSUE: how to avoid possibility that user will call setPrefx and may override automatic prefix ...
>  RESOLVED: all calls must be before startTag()
> 
> //ISSUE: if namespace set in setPrefix was not used shoul dit still result in xmlns:ns='' declaration?
>  RESOLVED: YES -the prefix MUST be declared as may be used for attriburte values QNames etc.
> 
> //ISSUE: maybe replace flush() with closeStartTag() (if not used it is still called implicitly ...)
> //      this will allow to write unesscaped strings safely (like for example BASE64 or hexbin)
>  RESOLVED: left flush() with exact description what it does but also added text("") to close stag
> 
> 
> //ISSUE: how to write efficiently unescaped XML
>  RESOLVED: use flush() and write into stream ...
> 
> 
> //ISSUE: where setPrefix() should (or not) be called (before first startTag(), between attribute() ...
>  RESOLVED: setPrefix() MUST be called before startTag() is affecting only this startTag
> 
> //ISSUE: add close() that validates if XML document was writent correctly (finished depth == 0, etc.)
>  RESOLVED: instead of close() use endDocument()
> 
> //ISSUE: additional properties to specify indent character, indent size, attrib quotation character
>  RESOLVED: added get/setProperty to support it as options
> 
> 
> --- DOCUMENTATION
> 
> //ISSUE: document that element and attribute names and namespaces are not escpaed!
> //ISSUE: document that comment, CDSECT etc. are not escaped!!!
> //ISSUE: docuemnt that duplicate attribute declarations are not detected !
> //ISSUE: implicit closing of startTag() by text("") requires more description and exmaples!!!
> 
> ---- FUTURE
> 
> // PROPOSAL: maybe have unesapced(String), unesapced(char[], off, len) methods?
> 
> //ISSUE: have one method tagWithContent() that combines startTag, text and endTag  ...
>  RESOLVED: later ...
> 
> //ISSUE: should xml:space and xml:lang scope be maintained ?
>  RESOLVED: not in this version
> 
> 
> 
> 
> To unsubscribe from this group, send an email to:
> [email protected]
> 
>  
> 
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 
> 
> 


-- 
Stefan Haustein
Univ. Dortmund, FB 4, LS 8   tel: +49 231 755 2499
Baroper Str. 301             fax: +49 231 755 5105
D-44221 Dortmund (Germany)   www-ai.cs.uni-dortmund.de