Hey Wolfgang, I think the idea was to just create the object model during the APT compilation step and then serialize it out using regular Java Serialization which should work once we get around to making AnnotationMetadata extend java.io.Serializable since most of the metadata is made up of Strings.
Then when the reflection processor is creating the Object model it can take a shortcut and load any available Serialized OM that might be present or do reflection as it currently does. Since the APT Annotation processor runs against the source it gets the param names that the reflection processor misses after compilation had removed them. No parsing involved just using an ObjectInputStream or something similar. cheers, Jonathan -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Monday, December 13, 2004 01:55 PM To: 'Beehive Developers' Subject: [WSM] About Beehive-120 in Jira. Hi there, The description of beehive-120 (http://nagoya.apache.org/jira/browse/BEEHIVE-120) in jira says "this can be fixed by storing the required info in a artifact that is generated when the jws-file is compiled". In what way do you think to retrieve the required info (e.g. parameters' name of a method ) from the artifact ? I think the only way to get the info is to parse the source code to get them... We have a couple of choices to accomplish it. 1. use a parser generated by JavaCC and customize it. ( the code generated by JavaCC is NOT restricted by JavaCC's license ) ( https://javacc.dev.java.net/servlets/ReadMsg?list=users&msgId=67734 ) 2. make our own parser from a scratch. What do you think ? I wanna know if there're any other ways that we can retrieve parameters's name of a method at compile time or runtime. Thank you. Wolfgang.
