Community
Participate
Working Groups
We now support camel case in Quick Outline and Quick Hierarchy but I think the filtering can still be improved significantly. Having to type everything from the beginning or type wildcards like * is both annoying and quite frankly needless. We should implement something like Subwords completion from Code Recommenders - http://blog.deepakazad.com/2012/05/subwords-completion.html We can implement the same filtering mechanism at other places as well like Open Type and Open Resource. But we can start with Quick Outline and Quick Hierarchy... Implementation details From bug 376674 comment 3 > > I thought the matching was only a regex of the the form 'e*x*p*r*e*s*s*i*o*n', > > is there more to it ? > > Yes, a bit more. In a nutshell, it counts how many pairs of 'ex' 'xp' 'pr' 're' > ... it finds in the completion string and ranks it according to the number of > pairs found. > > The regex is a "minimal requirements" filter that eliminates those completions > that do not contain at least all letters in the order specified by the token > (as you described above). Also from bug 377333 comment 2 > if some characters in a word are matched but not the first one we > should discard the proposal.
I think the Subwords completion engine should be migrated to JDT and replace the "standard" JDT completion engine, as it is much more powerful.