Here's the "complete" list of javax apis included with Glassfish, based on a 233mb source code CVS checkout...
javax.activation javax.servlet.jsp.jstl javax.resource (connector) javax.enterprise.deploy (deployment) javax.ejb javax.security.jacc javax.jms javax.mail javax.management.j2ee javax.persistence javax.servlet.jsp (jsr-152, JSP 2.0) javax.servlet.http (jsr-154, Servlet 2.4) javax.servlet.jsp (jsr-245, JSP 2.1) javax.transaction Obviously, verifying all the class files from the binary distribution vs what I compile out of CVS will be a bit of a chore. Don't suppose anyone has a method for comparing the contents of two file system trees? I can extract the class files from their distribution, build from source myself, and compare the file sizes etc assuming I can find a simple comparison process. Wayne On 2/27/06, Wayne Fay <[EMAIL PROTECTED]> wrote: > This is exactly why I said "we might not want to distribute as > javax.*". I am definitely concerned about ongoing maintenance etc. > Ideally we'd get the Glassfish project themselves to build the Jars > and submit to Maven repo. They are using Ant and Maven1 for their > build process, so they are familiar with the Maven repo concept. > > I will compile the sources, compare each to the binaries distributed > by Glassfish, and report back later today... > > Wayne > > On 2/27/06, Steve Loughran <[EMAIL PROTECTED]> wrote: > > Wayne Fay wrote: > > > > > > > > However, the CDDL source code license ensures we **can** download the > > > proper source, build/unit test, package, bundle with poms, and deploy > > > **those** executables from the repo. > > > > > > This is an important difference. That's why I originally said: > > >>> Assuming we all agree that we can do it legally, I'd be happy to build > > >>> the jars, write the poms, and add to Jira for uploading. > > > > > > Any more comments? :-) > > > > That would be progress. One thing to check is how much difference is > > there between a JAR made that way and a released JAR. In an ideal world. > > apart from manifest data, there would be no difference. > > > > But if there is a difference, there is a risk that something wont work, > > and then who is left fielding the problems? > > > > Maybe the artifacts should be published with a groupId that indicates it > > was rebuilt or something, so that glassfish-rebuilt-jta-1.0.3.jar is > > clearly different from jta-1.0.3.jar. > > > > -steve > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]