We've inherited an application (a fairly large one) that has a lot of unlocked reads and a few unlocked writes to shared scope variables (was written before all the problems of unlocked reads and writes came to light). We can catch all the unlocked writes, but there are a LOT of unlocked reads, and some of them are nasty (i.e., a CFIF Session... that wraps a LARGE amount of code). We're not sure of all the ramifications of bulk loading the session scope into the request scope yet - there may be writes to session that need to be read in the same template, etc. We'll figure these out as we rewrite and clean up. But the problem is, we need to have this application live right now, and it may be playing havoc with our servers. So, we're going to turn "automatic read locking" on for session, and "full checking" server and application. We've done this on our dev box, and caught all the unlocked writes. But, if we do this to our production box, what is the performance impact of this? Is it similar to just having the locks in there? I understand that the pcode 'compiler' puts the locks in where it needs when it compiles the template, and sometimes it doesn't always put it in the most optimized place, right? So this would be similar to us putting a CFLOCK around that huge CFIF Session.whatever block, right? Would this be better or worse then "Single-Threading" sessions? Thanks in advance! -- Edward Smith Internet Broadcasting System Director of Architecture http://www.ibsys.com ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Structure your ColdFusion code with Fusebox. Get the official book at http://www.fusionauthority.com/bkinfo.cfm Archives: http://www.mail-archive.com/[email protected]/ Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists
