On 14 October 2017 at 13:28, Scott Deboy <scott.de...@gmail.com> wrote:
> Awesome! Happy to help here where I can, although I haven't worked in > Java in forever. > Are you using any other JVM languages regularly? Or have you gone down the native code route or something? > Chainsaw is really five parts: > > - Receivers which grab the data from whatever source > This is the part I'm most interested in improving first. We already have a pretty standard JSON layout (and XML and YAML) which now have standard ways to parse them in log4j-core's API. As I mentioned as well, supporting some binary formats would be useful here, too. > The sorting is wonky and slow, especially if sorting times (!) > We might be able to address this by abusing some big data tech like Apache CarbonData or similar columnar formats. Or even just using an embedded SQLite or H2 database or similar. > Chainsaw' support for positive and negative filters, combined with > colorizing, search and event annotation combine to give you a pretty > powerful set of tools for log analysis. It'd take a long time to > recreate what's there in another language. > I suppose this could be an advantage for going with Kotlin since there are ways to mechanically convert Java to Kotlin which produces pretty decent code most of the time from what I hear. Sticking with a JVM language for the project does provide the ability to incrementally migrate to another language since they all support linking and compiling Java source files as well. -- Matt Sicker <boa...@gmail.com>
