Cc: [email protected] Hi Wolfgang,
Yes, my current thinking is that having a new deployment process in place that doesn't rely on on-demand builds anymore, we can now generate two files with apt: 1) the class file (the same as we would get from just running javac except that it would take care of controls embedded with the jws if there are any) 2) the object model in some serialized format (since it is not intended to be human-readable not modifyable I will probably use java.serialization) At deployment time, we can load the class and the object model and will have to all information that was in the source file, including the parameter names (which are not available from class files). This is the only way for us to have a consistent build/deployment model that integrates well with controls and comply with JSR-181 at the same time. Cheers, -michael PS: APT can generate multiple output files and will compile all java source files (if desired) as part of the process (see http://java.sun.com/j2se/1.5.0/docs/guide/apt/GettingStarted.html). -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Monday, December 13, 2004 8:19 AM To: [EMAIL PROTECTED] Cc: Michael Merz Subject: Re: [WSM] About Beehive-120 in Jira. I just wanna make one thing clear. so If one compiles jws with APT (the apt processor), the one can get two artifacts which are the class file of jws and the serialized file of the object model (AnnotationModel) of the jws. Then, one can deploy the serialized file as web service using the reflection processor. Is it what you mean ?? Thanks in advance. Wolfgang
