Have you tried updating to the latest version of IE? IE6 had some major problems with leaks
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Dan Sent: Thursday, February 18, 2010 12:59 PM To: RBASE-L Mailing List Subject: [RBASE-L] - Re: Memory Usage out of hand I did mention that yesterday that IE exhibits the same for that URL. Ok, so we are back to... I need to find a way to open a whole instance of R:Base to a third monitor. I can have it stay open fcr a while and then dump it and open another. I could do this from a batch file, but how to get it to open on a particular monitor. I was hoping after you closed the form, and hence the IE on the form, it would release. Any thoughts? Ps. I can't change the way the web page eats memory, that is created by some off site third party program, and google Maps. So assuming that I can't get Google, or GPS Visualizer to re-do their code. I need to find a work around. At 01:28 PM 2/18/2010, you wrote: >Good catch Dennis. I didn't think of the page being something Dan >created. Could be Object creation that is not being disposed when >the page goes out of scope. > >----- Original Message ----- From: "Dennis McGrath" <[email protected]> >To: "RBASE-L Mailing List" <[email protected]> >Sent: Thursday, February 18, 2010 1:14 PM >Subject: [RBASE-L] - Re: Memory Usage out of hand > > >I tried your url in IE. It eats memory every time I refresh the page. > >Dennis > >________________________________ >From: [email protected] [mailto:[email protected]] On Behalf Of Dan >Sent: Thursday, February 18, 2010 11:55 AM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > >Hmmm, > Completely different results than me then. > I let it loop while I was at lunch. 47 times through the loop and > my memory usage for that instance of R:base was 1.4 gigs. > >That is using GOURL in the timer eep > >Also turned off the timer and had two buttons as you suggested, each >time the button was pressed memory usage increased. > >Button eep was > >PROPERTY URLControl GOURL ' >www.vredevoogd.com/dan--mapall.html<http://www.vredevoogd.com/dan--mapall.html>' >return > > >At 12:42 PM 2/18/2010, you wrote: > >I tried switching sites back and forth like crazy, no leak using the GOURL > >Dennis McGrath > > > >________________________________ >From: [email protected] [ mailto:[email protected]] On Behalf Of >Dennis McGrath >Sent: Thursday, February 18, 2010 10:03 AM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > >Put a couple of buttons on your test form. >Each button will do a different url. >Test until it works and use whatever syntax you found. >Waiting for a timer makes testing very slow. > >Dennis McGrath > >________________________________ >From: [email protected] [ mailto:[email protected]] On Behalf Of Dan >Sent: Thursday, February 18, 2010 9:56 AM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > >Did I understand correctly? >1. remove the recalc variables >2. set the url object to have a component ID (I set it as URLControl) >3. in the timer eep use: >PROPERTY URLControl GOURL ' >www.vredevoogd.com/dan--mapall.html<http://www.vredevoogd.com/dan--mapall.html>' > >If this is the case, no errors generated, no increase in memory, >however no update to the web page at the time of the eep firing. > > > > >At 09:35 AM 2/18/2010, you wrote: >Try PROPERTY <Component ID> GOURL 'value' > >Dennis McGrath > >________________________________ >From: [email protected] [ mailto:[email protected]] On Behalf Of Dan >Sent: Thursday, February 18, 2010 5:44 AM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > >Dennis, >At the moment the ONLY thing on the test form is the web browser., >and the only variable to recalc is the value of the web page (as far >as I know this is what causes the refresh on the web page) > The web page changes every few minutes by the code that I took out > of this application and run separately on another instance of > R:base. If there is a better way to refresh the web page than > recalc variables, I am open to trying it. > >At 05:38 PM 2/17/2010, you wrote: > >Dan, > >Sounds like you are onto something. >Can you avoid using recalc variables? > >There are other ways to get data to display without resorting to that. >You can do your calculations in code and use property commands to >update the form controls. > >But, before you do that, look carefully at your form >expressions. Are the data types what your expect? >How many variable controls do you have on the form? >You might try making a copy of your form and start hacking variables >and control out until you find the culprit. > >Dennis > >________________________________ >From: [email protected] [ mailto:[email protected]] On Behalf Of Dan >Sent: Wednesday, February 17, 2010 4:15 PM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > > > > > >Ok, I did some more testing. I turned off all processing and just >let the form run and refresh the web page over and over and >over... No increase in memory at all... so the memory loss has to >be in my coding. > >Then I did that again with a 'recalc variables' in the eep.... >Boom Memory issue > > >Eep calls the form... timer eep has... > >set v vcount = (.vcount + 1) >pause 3 using .vcount > >RECALC VARIABLES > >return > >every round adds 500k > >So No Rbase processing going on... no while loops etc... Timer, >web page, refresh. > >Open the form manually, it goes up > >Close the form just back to the R:Prompt Memory not released > >Re-open the form Memory goes up. > >close it back to prompt... stays up. > > >Next Idea? > > > > >Yes, Temp tables are created >files imported, >lots of processing done, >files renamed >files ftp'd to an off site source >While loops > >are all part of the code that was running every 15 minutes. > >If I call the form manually, the memory is not released until the >R:base session is closed. > >This code has been running unchanged 4 about 3 years or so. Not >sure why just in the last three months it has become an issue. > >I am now running a sample of my code without the form, and putting >it through a timed loop to see if it in fact is the culprit. If so, >it might be just as easy as run them separate. I could easily close >and open a second R:base on a timer and leave the web form open all >day with no memory issue. Will do some thorough testing. > >Thanks everyone for your help. I will post my results. > >Dan > > > > >1. If you call the form manually, does memory increase when it is >running and release when you exit the form? > >2. Are you using a while loop, or my goto code in your eep? > >Dennis > >________________________________ >From: [email protected] [ mailto:[email protected]] On Behalf Of Dan >Sent: Wednesday, February 17, 2010 12:41 PM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > >Dennis you are right with my first test. > >Originally I would just refresh the web content... Memory usage increased > >Then I tried a closewindow, start new form... Had that in a wrong >spot, and opened more and more forms. (my bad) Memory usage increased > >Now the start eep that calls the form originally is set up like you >are suggesting... completely closing the form on each iteration of >the loop. Memory usage increases. > >I am guessing the ONLY way to get the memory to reset is to close >R:Base. If that isn't the case, that would be best for me as I don't >know how to open an instance of R:base to a particular monitor. > >Without fixing this, This application uses up memory so fast that >twice a day things fail. Have to manually start and stop it. Not >good for an automated process. > > > >At 01:28 PM 2/17/2010, you wrote: > >Dan, > >I think your problem is nextelshuttle never completely goes away >because it is calling another copy of itself. >You should never have code that calls another form after closing the form. >The fact that the memory is released when you completely exit your >routine is a hint. > >Instead, set up a loop which calls nextelshuttle. >The nextelshuttle timer should just closewindow after 15 minutes. >No other code should be rum after closewindow. > >This way, nextelshuttle will completely close and release memory >before being called again. >You will need some way of setting vQuit to YESat the end of the day >or when a button is pressed on nextelshuttle > >Label Loop >set up some vars... >set var vQuit = NO >edit using nextelshuttle >if vQuit = YESthen > goto Endloop >else > goto Loop >endif >Label EndLoop > > >Dennis McGrath > >________________________________ >From: [email protected] [ mailto:[email protected]] On Behalf Of Dan >Sent: Wednesday, February 17, 2010 11:59 AM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Memory Usage out of hand > >Ok, > I made your suggestion work, and closewindow is working now, but I > think I need to actually close R:Base to free up the > memory? (Memory usage goes up by 50,000 k every loop) So, if that > is the case, >how do I have R:base open to a particular screen (I run a 4 screen >setup here and the app I am working on has to be on the upper right, >or Monitor 3) > >At 11:30 AM 2/17/2010, you wrote: > >Ok, > > >Yes, Rbase Dat calls up the form: > >set up some vars... >edit using nextelshuttle >exit > >Nextelshuttle form timer does the 15 minutes, and calls an internal eep > >That eep, >does : > >Closewindow >run nextelshuttle.dan >return > > > >The nextelshuttle.dan >does lots of database stuff, and then >edit using nextelshuttle > > > >The closewindow does not close the window, but a second form is >generated (and third and so on for each loop) > >If you close the latest loop, all forms close. > >No memory is released till all forms close. > >What am I missing? >Dan > > > >At 11:00 AM 2/17/2010, you wrote: > >Somewhere you have a program that calls up your form, maybe RBase.dat >or something else. Do the following (or something like it depending >on how your >system works): > >your program: > >LABEL formagain > >SET VAR vExit TEXT = NULL >EDIT USING nextelshuttle > >IF vExit = 'yes' THEN > EXIT > RETURN >ELSE > GOTO formagain >ENDIF >RETURN > >In your Nextelshuttle form: >You would have a button to exit the form totally; that form's eep >code would be: > SET VAR vExit = 'yes' > CLOSEWINDOW > RETURN > >And somewhere there's code that does the 15 minute timer. In that code when >you've reached the 15 minutes, the eep code would just have: > CLOSEWINDOW > RETURN >or you can use the pre-defined "exit" action. > >Important: you need the "return" even if you just have closewindow on there. > >Karen > > > >So, now that we have surmised that my memory issue is the use of >IE, I am trying to find a way to have the program close and reopen >the form each loop through. > >Closewindow doesn't seem to be my answer. Anyone have any good ideas? > >Problem: >I run an instance of R:base that opens just one form 'edit using >nextelshuttle' >The form has two objects, a clock and a web object that displays a >city map with all our technicians on it. >that form has a 6 minute timer to run an eep that refreshes data, >imports info to the database, after the eep, the web refreshes. (this >is where the memory loss is) > >I was hoping a closewindow would work, but I cannot make it close the >form and restart a new form. I suspect though that I would need to >close that whole instance of R:base to free up the memory. > >Thoughts? > >
