Lubos Kosco <Lubos.Kosco at Sun.COM> writes: > On 16.2.2010 18:38, Chandan wrote:
>> It is not exactly a bug, but the code that shows the >> matching line does no go deep into parsing out the >> comments or string literals. Doing so may slowdown >> the results. >> > > Then it seems like an RFE to me to improve the matcher ... I saw > lucene has some kind of highlighter for what we do with Matcher > classes ... maybe it'd make sense to do some research and use it > instead our own code; > because if we have the info about that this is a symbol search and we > mark it with green "local" word then Matcher could just print these > lines in case a symbol search is done, > hmm? Sounds like a good idea, but note that only the lines with definitions say "local". In this case, the search was for symbols, not for definitions, so some of the lines without "local" should still be printed. I don't know if we currently have enough information to do this from the matcher, but it would definitely be an improvement if we could. -- Knut Anders
