Hey Glenn, My suggestion, while you're waiting to hear from support, would be to run your app with a tool like Valgrind. It could be you've found a leak in the engine, but you might just have a buggy dependency in your Centos distro. Valgrind output may give you some hints as to where.
-- cb On Fri, Apr 13, 2012 at 6:53 AM, Glen Bojsza <gboj...@gmail.com> wrote: > My latest update / feedback on the issue... > > I tried monitoring the memory on the Centos 6.2 system with top -p $(pgrep > Xorg) and watching the %MEM. It was really interesting and informative... > when I launched the application I was down at 1.5% MEM and as I did > database queries and plots I saw it increase between 0.3 and 0.5 % each > time (This had several queries with large numbers of data). > > To address Tim's concern I stopped doing any graphing and solely did a > series of import snapshots. > > Background here: I am trying to produce a "poor man's" powerpoint where > each card is a slide and I have a text field that acts as a background. > > A import snaphot of the image on each card --> reduce the new image and > then place the image in the text field by replacing the first letter in a > line in the text field. > > Each image is associated to the card's name which is then associated with > the line of text in the field. > > Anytime the user wants to re-arrange the sequence of cards/slides they can > click and drag the associated image in the text field to the position they > want and this then triggers a "refresh" where all the card images are then > again having their image being done as an import snapshot and placed back > in the text field. > > So essentially, going through this cycle several times showed the memory to > start increasing approximately .7% each cycle. (this is where there are 20 > cards so 20 images) > > I increased the number of cards / images to 40 and the memory started to > increase 1.2% each cycle! > > Never has the memory decreased. > > I removed ALL images and cards except for the intro card and the memory > never decreased?! > > All I had to do is move the images in the text field as a user would when > wanting to re-arrange things and when I hit approx 80% MEM the application > stops and shuts down. > > MAC OS X experience: using the activity monitor and watching the > application I saw the memory increase a little and then remain fairly > constant while I cycle through the re-arranging of images in the text > field. > > The difference is that it did NOT just continually grow in size similar to > what I saw on CentOS. > > Also, when I deleted all the images except for the intro card the memory on > the OS X dropped to almost where I started. > > My conclusion is that the memory and application on OS X are acting as what > it probably should be. The CentOS is just a straight climb up ... which > means, if true, then any Livecode Linux apps will crash and it only depends > on when. Obviously some applications may only increase the memory in > decimal % so it would take a while and may never happen if the user closes > the app between uses. > > I think this is serious to the point that someone has to fix quickly > because Livecode sells this as a supported platform and if it fails this > easily then the overall credibility of Livecode and it's user community > suffers. > > With our app due shortly we can't take a chance that the user doesn't > experience this so we are not certain what to do. > > Glen > > On Thu, Apr 12, 2012 at 9:56 PM, Tim Jones <tolis...@me.com> wrote: > > > Glen, > > > > I sincerely doubt that this is related to the version of Linux you are > > using. This sounds like a potential memory leak in your graphing > > functionality. Are you using native LiveCode or someone's plugin / > stack? > > > > Tim > > > > On Apr 12, 2012, at 6:23 PM, Glen Bojsza wrote: > > > > > Hello, > > > > > > I have discovered a potential problem with my Livecode app running on > > > CentOS 6.2 > > > > > > Basic summary is that the application queries databases and generates > > > graphs. The application does snapshots of these graphs and allows the > > user > > > to position them or re-order them. > > > > > > After stress testing the app (which is due to be released this week) we > > > discover that after generating a volume of snapshots of the graphs (low > > > 100s) that the application crashes hard ... ie just shuts down. > > > > > > Even if, before generating more snapshots, you delete all existing > > > snapshots the application shuts down... deleting doesn't seem to do > > > anything in helping the application to continue to run?? > > > > > > Are there any suggestions that can be tried to solve this? > > > > > > From above it is easy for anyone to reproduce if they have a Linux > system > > > (we are locked to CentOS). > > > > > > regards, > > > > > > Glen > > > > > > _______________________________________________ > > use-livecode mailing list > > use-livecode@lists.runrev.com > > Please visit this url to subscribe, unsubscribe and manage your > > subscription preferences: > > http://lists.runrev.com/mailman/listinfo/use-livecode > > > _______________________________________________ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your > subscription preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode > _______________________________________________ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode