Now code model consumes a lot of memory. And even if we reduce memory consumtion
twice - or thrice - it won't help. If user creates project that is large enough,
he'll finally get OutOfMemoryError.
The problem is in the code model implementation design - it's too monolitic now;
objects hard reference other objects, so the entire model should necessarily
reside in memory.
Code model implementation should be redesign so that the entire model won't
reside necessarily in memory.
Default setting for tasks is P3.