Hi Nicolas, On 7/18/06, Nicolas Modrzyk <[EMAIL PROTECTED]> wrote: > > I am using a limited set of flows to start on the openwfe engine. > It seems that on each start, the flow is compile from its xml source > code, to the openwfe objects.
OpenWFE is more an interpreter, but the compilation cost you refer to might be the one used to turn the file into some [JDOM] XML. > My question is the following: > -> How would I go to get the flow being cached, and reused in some > kind of way. Or, in other words, is there anyway to tell the engine > that the flow to start is the same as some really recent one, and > thus to reduce the compilation time ? > > Does it make sense ? Would we see speed improvements when starting > the flow ? The latest 1.7.1 jars (not yet in Maven unfortunately, but the upload request has been made) could provide you some perf enhancements. > In my case, we use only one flow definition over and over again with > new and different dynamic parameters... An immediate way of avoiding file -> XML 'instantiation' would be to load the core of your process at the engine level. I don't know how Magnolia does it, but does it use one and only one OpenWFE engine ? If yes, that could make sense and spare you some CPU cycles. I have to look at caching process definitions, early versions of OpenWFE had that, but everybody was asking to turn it off (especially on development/validation platforms). HTH, -- John Mettraux -///- http://jmettraux.openwfe.org ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ OpenWFE - Open source WorkFlow Engine OpenWFE-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/openwfe-users
