Jaroslav Luke-Bš wrote at 2005-6-22 14:51 +0200:-A >I have idea about bytecode cache in Zope, it is possible to implement? > >What I mean: python scripts (in ZODB), DTML documents/methods, ZPT will be >bytecode compiled in similar way like ordinary .py scripts into .pyc and then >run much faster than "noncompiled" ZODB objects. I think that it should have >significant speed improvement.
All of these objects are (in some way) already now "byte compiled". But only some of them store the compilation result inside the ZODB (ZODB based DTML objects and PythonScripts do). Thus, on initial access, they need to be compiled. Further access already finds the compiled code and can skip compilation. However, there is potential to reduce startup (!) time by storing the compilation result. That is the purpose of my "CompiledExecutables" <http://www.dieter.handshake.de/pyprojects/zope> When you look there, you find a note "currently broken". That is because the current version only works with our own slightly modified (fixed :-) ) Zope. It causes an infinite loop when used with a stock Zope. Someone reported the problem and I think I have a fix. I asked the reporter to try it out and report back. But never got an answer... maybe, because my eager spam filter did not like his mail service provider (all providers that relay spam and do not react on my complaints are banned to send me any more emails). If you are interested to take over and check my fix, let me know (it usually is better not to try to contact me privately (--> eager spam filter) but to send Zope related messages and questions to the mailing list). -- Dieter _______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )