Hi Pedro, NB: I am not the maintainer of the Kupfer package, so I will not decide the resolution of this bug.
I think your hope in reducing Kupfer's memory use is misplaced. You are correct that a minimal configuration of Kupfer can run without the keybinder module, and a minimal configuration of Kupfer can also possibly use only small amounts of memory. However, Kupfer's implementation details are such that our compromise is to use a bit more memory, and instead use a very dynamic implementation. The python implementation of Kupfer is not very memory efficient; neither is the architecture where all python modules are loaded permanently into the same process. Using less plugins, you will see less of the latter problem. The fact remains for me, that next to other memory usage patters in Kupfer, the keybinder module accounts for almost nothing. Much worse are all the shared libraries that we can't control (like gdata for some Google plugins, or other of the convenience modules that are used). If you are a coder, you can see that there remains interesting problems to hack on. A good start would be a reliable to *measure* memory usage, without even that starting point, it is hard to improve the situation. best regards, ulrik -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
