Andre, I ran into the same problem when analysing 3500 stacks distributed over 3 computers, over 1000 stacks on each. Rev crashed. In the repeat loop version you could watch the memory go down in the activity monitor (on a Mac) and once it went into virtual memory soon after it crashed. Than I did a send in time version and it went smoothly and fast even though it was over a local area network. I made a bug report as a minor bug #6631 which has a demo stack that shows how repeatedly loading a stack and deleting it afterwards crashes revolution. And there is the send in time version that avoids crashing. Essentially it is like the code Richard has proposed.
Once you get the idea of send in time constructs they are a) easy to do and b) keep you out of trouble because Revolution can do it's housekeeping while you do your long and tedious computing. It may be a design flaw to put that amount of data into that many individual stacks but there were reasons for me to choose this approach. And with the send in time approach all went well. regards Bernd Andre.Bisseret wrote: > > Bonjour, > > I have an app including about five hundred stacks (main stacks with > substacks); they are physician's clients files. > In order to produce statistics I have a handler which search the > values of several measures in the main stack and one substack of each > clients. > > I am using a repeat loop to search the stacks (useful to say that, as > soon as the values are picked up, I "delete" the main stack). > > Despite this, my app systematically crashes before completing the all > set of stacks. > > In the archives of the list, I found the following : > --------------------------------------- > Re: Releasing Memory and Virtual Memory > > J. Landman Gay > Tue, 11 Aug 2009 15:27:28 -0700 > > BNig wrote: > > Derek, > do you import the 100 or so images in a repeat loop? > If so try whether adding a wait 2 milliseconds in the repeat loop > helps. > Or alternatively try to do it in a "send" structure. In my > experience Rev > has problems releasing memory in a repeat loop. When I changed to > a "send to > me in 2 milliseconds" no more problem with memory build-up. > > I agree, and I'm pretty sure this must be what's wrong. Rev doesn't do > any garbage collection until a handler ends. The most reliable way is > the "send in <time>" structure. > -- > Jacqueline Landman Gay | [email protected] > HyperActive Software | http://www.hyperactivesw.com > ----------------------------------------- > > But after a lot of unsuccessful trials I am completely lost and seem > unable to think correctly anymore! > > including "wait 2 milliseconds" in the repeat loop does not help. > > I don"t see how I could avoid the repeat loop ? (how to use a "send in > time" structure ? is it instead of the repeat loop ??) > I tried to proceed one hundred stacks at a time only, but here again I > don"t see how to avoid a repeat loop ? > > What am I missing ? > > Thanks a lot in advance for any advice > > Best regards from Grenoble > André > -- View this message in context: http://www.nabble.com/memory-saturating-with-a-repeat-loop-tp25082952p25086324.html Sent from the Revolution - User mailing list archive at Nabble.com. _______________________________________________ use-revolution mailing list [email protected] Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-revolution
