[Ironpython-users] How to free memory in ironpython

Djordje Spasic issworld2000 at yahoo.com
Mon Nov 7 16:25:20 EST 2016


Hello,

I have an issue. I am using ironpython 2.7 editor called through Rhino 5 application, and I load a single .dll file from an application called MapWinGIS. Like so:
dll_filePath = "C:\ProgramFiles\MapWinGIS\Interop.MapWinGIS.dll"clr.AddReferenceToFileAndPath(dll_filePath)import MapWinGIS
The problem is that I can only call a particular method once.
If try to do that the second time, the same method returns "False". When method works correctly (the first time) it return: "True".

The problem goes away if I restart the Rhino 5 application (and ironpython 2.7 editor along with it).


Why is this happening? I contacted the developers of upper mentioned MapWinGIS application, and they said that they can not replicate this behavior and that I should ask this question on ironpython mailing list. They suggested that I need to somehow "free the memory before you do the second run of the method".

So to make my problem more clear, this is what happens:

If run a script containing the this particular method twice, the first call of the method returns "True", the second one "False":

variable1a = "something1a"variable1b = "something1b"returnedValue1 =MapWinGIS.someClass.someMethod1(variable1a, variable1b)  # returns "True" variable2a = "something2a"variable2b = "something2b"returnedValue2 = MapWinGIS.someClass.someMethod1(variable2a,variable2b)  #returns "False"
If run the same script again, both upper calls of MapWinGIS.someClass.someMethod1 method return "False".
If I restart my Rhino 5 application, then again the same thing happens: "True" is returned on first MapWinGIS.someClass.someMethod1 call, while "False" is returned on the second call. If run the script again, then they both return "False". For every other run of the script, I get "False" on both calls. And this continues until I restart Rhino 5.

I perfectly understand that it is impossible for anyone from Ironpython mailing list to know the exact cause of this problem.

But can you at least help me with "free the memory before you do the second run of the method" part?

I tried deleting all the variables after the first call, calling the garbage collector and also removing the .dll file from the list of loaded assemblies. Like this:
import clrimport gc dll_filePath = "C:\ProgramFiles\MapWinGIS\Interop.MapWinGIS.dll"clr.AddReferenceToFileAndPath(dll_filePath)import MapWinGIS variable1a = "something1a"variable1b = "something1b"returnedValue1 =MapWinGIS.someClass.someMethod1(variable1a, variable1b)  # returns "True" # delete all variablesdel dll_filePath; del variable1a; del variable1b; del returnedValue1; del MapWinGIS # remove the .dll from the listof assembliesfor i,assembly in enumerate(clr.References):    if assembly.GetName().Name == "Interop.MapWinGIS":        clr.References.RemoveAt(i)        break # collect garbagegc.collect()   dll_filePath2 = "C:\ProgramFiles\MapWinGIS\Interop.MapWinGIS.dll"clr.AddReferenceToFileAndPath(dll_filePath2)import MapWinGIS variable2a = "something2a"variable2b = "something2b"returnedValue2 = MapWinGIS.someClass.someMethod1(variable2a,variable2b)  #returns "False"

But it is not working (I still get the "True" on first method call, and "False" on second).
I also tried to exchange the variable1a and variable1b with variable2a and variable2b. Or even use the same variables in both MapWinGIS.someClass.someMethod1 calls. Still, no improvement.
The MapWinGIS.someClass.someMethod1 that I am calling is actually: MapWinGIS.UtilsClass.OGR2OGR.
I named it MapWinGIS.someClass.someMethod1 to make it easier for others to follow the code. Not sure if this was good idea. If it wasn't I apologize.


This issue is driving me crazy. I would be very grateful for any kind of advice.
Thank you in advance!


Kind regards,
Djordje Spasic

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ironpython-users/attachments/20161107/13361f32/attachment.html>


More information about the Ironpython-users mailing list