Community
Participate
Working Groups
3.2 M5 The Javadoc location path shouldn't be set if there isn't any Javadoc.
Deferred
Can this be addressed for 3.3? Note that it is also a performance issue.
Yes, we aim to please...
We might be able to leverage the OSGi execution environment access rules to determine which packages have javadoc/source (i.e. list of API packages). The current API on execution environments does not expose what we would need to do this, but we could do something internally, since the EE implementation and the 'standard vm type' are both in the launching plug-in.
Not planned for 3.3 - I'm afraid this would impact launching performance. Each time we launch a JRE with default libraries, the default libraries are computed. This includes the computaion of javadoc locations. We would need to examine the contents of each library to see how it matches with the API of the system library, which would be slow. The information could be cached - but then we run the risk of becoming incosistent with the file system. (I know later is deprecated, but for now that is our style of marking bugs for future consideration).
As of now 'LATER' and 'REMIND' resolutions are no longer supported. Please reopen this bug if it is still valid for you.