Hi Dmitry,
> This is a flat list of problems (memory leaks):
>
> [...]
Thanks for your attention to the code correctness and quality.
In this project we typically don't bother about "cosmetic" memory
leaks - e.g. when several extra bytes are allocated during the process
startup, or many bytes are allocated in the short-living process (they
are going to be freed anyway when the process terminates).
If I didn't miss anything, your report contains only such cosmetic
leaks. If you believe I missed something, please don't hesitate
submitting a fix for a particular memory leak to pgsql-hackers@.