Trond Norbye <Trond.Norbye at Sun.COM> writes: >>> - Some minor tweaks/fixes, for example I've noticed that in 0.7 the >>> date shown next to a file in the webapp is the time it was indexed, >>> not the time it was last modified. Not sure if this is still the case >>> in HEAD. >>> >> > There is a RFE for this. The problem is that to fix this we need to > set the filemodification date on the file in the filesystem to the > date of the trunk revision. This isn't impossible, but no one > implemented it ;)
It should be easy enough to get that information from HistoryGuru. To avoid fetching the entire history for each file just to get the latest timestamp, we should probably add a new method to HistoryCache for this. getLatestModification(File,Repository) or something. >>> Regarding the Python analyzer, is anyone working on this already and >>> if so, anything I can do to help? Otherwise, I will read up on JFlex >>> and probably start off with an existing analyzer, maybe the Java or >>> Lisp one. What would be the best way to test my changes during >>> development without having to run the full stack -- is there a way to >>> run the analyzer on a single input file and check its output? >>> > > Create a junit test for it? JUnit isn't that well suited for visual inspection of the generated HTML during development, I would guess... The last time I wrote an analyzer, AnalyzerGuru had a main method that would analyze a file and print the generated HTML to System.out. That method seems to be gone now, though. I found it useful to have a small OpenGrok installation with just a few test input files for the analyzer. To test changes in the analyzer, I removed the old data root, performed a full indexing (which only took a couple of seconds on such a small data set), and reloaded the page in Firefox. Starting off with an existing analyzer sounds like a good idea. Copy and paste is the technique that's been used to write most of the analyzers. :) -- Knut Anders
