memory - why isn't it getting freed

Robert Brenstein rjb at rz.uni-potsdam.de
Thu Aug 12 04:51:13 EDT 2004


I did not mean to imply that Valentina's cache keeps growing. I had 
once a situation where there was a memory leak as a result of some 
weird interaction between MetaCard and Valentina. I had a field in a 
window that was displaying live log of on-going activities. If 
valentina was off, all was fine. If valentina was on, updating that 
field was leaking memory (and the field had a fixed number of lines 
displayed, for each added at the end, one was deleted at top). I 
ended up eliminating all live updates of log fields. Truly strange. 
The leak was definitely on the MetaCard's side but under no other 
circumstances I managed to reproduce it otherwise (and I really tried 
as I wanted Scott to fix it) and even in that program changing some 
unrelated scripts would trigger this behavior on and off. There was 
aparently some strange memory fragmentation occuring that was 
resulting in MetaCard acquiring more and more of it.

Robert


>Chris,
>
>Also probably worth noting- with Valentina you can control the size 
>of the cache, so if your data is small and you want to conserve 
>memory, you can always make the cache smaller. Otherwise, shutting 
>down the database is actually a pretty normal way to dispose of a 
>cache- most databases want to keep the cache big and persistent to 
>increase performance and so releasing the cache is actually one of 
>the primary purposes of a shutdown method (other than closing 
>connections to database files).
>
>FWIW.
>
>>Thanks, Robert.  I was able to make a called to Valentina("Shutdown"), and
>>that will at least clear my cache. Then I'm able to reinitialize the
>>Valentina kernel after that.
>>
>>Strange behavior going on here, but at least there's a workaround, for which
>>I'm very grateful.  I was having much stress over this one. :-)
>-------------------------------------
>Brian Yennie
>Chief Technology Officer
>QLD Learning, LLC
>(941)-928-7127
>briany at qldlearning.com
>--------------------------------------
>
>_______________________________________________
>use-revolution mailing list
>use-revolution at lists.runrev.com
>http://lists.runrev.com/mailman/listinfo/use-revolution



More information about the use-livecode mailing list