Tim, I still a question about Vladimir's patch - it just does not work in my configuration, and I cannot figure out why. This JIRA remained as the only questionable point in zip/jar staff, the rest of cleanup was merged in a cumulative patch at [1].
Generally, Vladimir's optimization looks correct for me. From the other side, while looking into the crash reason I have noticed that the code caches recently used entries, so I want to check a performance benefit reported by Vladimir myself to understand why this caching is not effective. It makes sense to check with zlib optimizations enabled. Honestly, I need to catch up with another portion of development [2] before returning to Vladimir's patch. Thanks. [1] http://issues.apache.org/jira/browse/HARMONY-4569, http://issues.apache.org/jira/browse/HARMONY-5522 [2] http://wiki.apache.org/harmony/Startup_Performance#CDS On Thu, Mar 13, 2008 at 3:48 PM, Tim Ellison <[EMAIL PROTECTED]> wrote: > Dmitry Egorov wrote: > > Hello Mark, Tim, > > Thank you for your idea! We have discussed possible native > > optimizations of jar and zip code with Alexei Fedotov and I'm > > currently looking into this. > > Great. Looking through the JIRAs I see Alexei raised a 'double read' > issue. You still have outstanding questions about this Alexei? > > [1] http://issues.apache.org/jira/browse/HARMONY-5526 > > Regards, > Tim > -- With best regards, Alexei
