Egor, I share your concern. Ever changing nature of today's world implies that optimizing JIT have to be aware of JVMTI, though I cannot request implementing that from the patch author.
On Sat, Dec 19, 2009 at 1:11 PM, Egor Pasko <egor.pa...@gmail.com> wrote: > On the 0x689 day of Apache Harmony Alexei Fedotov wrote: >> [added d...@] >> >> Egor, >> Thanks for asking for review. I miss code exercises. The code looks >> ok. BTW, the review took me here [1] - thanks Google for the wonderful >> instrument. >> >> As for your last question, TI appeared in java earlier than >> java.lang.instrument. This means implementing the latter using the >> former may make some sense. >> >> [1] >> http://www.google.com/codesearch/p?hl=ru#TasP9sO-cIM/trunk/vm/vmcore/src/jvmti/jvmti.cpp&q=load_agentlib&exact_package=http://svn.apache.org/repos/asf/harmony/enhanced/drlvm/ > > Thanks, Alexey! > > I am totally not an expert in TI, please, correct me. > > My concern was that in TI mode DRLVM runs slower. One reason to this > is that Jitrino.OPT does not support TI. It sounds like a high price > to pay for some innocent instrumentations. What can we do about this? > > -- > Egor Pasko > > -- With best regards / с наилучшими пожеланиями, Alexei Fedotov / Алексей Федотов, http://www.telecom-express.ru/ http://harmony.apache.org/ http://www.expressaas.com/ http://openmeetings.googlecode.com/
