Re: Proposal for fixing intra-query memory leaks

Karel Zak <zakkr@zf.jcu.cz> writes:
>  The chunk header must be relevant and same for all allocation methods and
> must be independent on AllocSel (aset.c) code. Now, it is not. Am I right?

Right.  Actually it would work to have additional data before the
standard link-to-context-plus-chunk-size fields, if a particular method
really needed more per-chunk data.  But I doubt we'd ever need that.

And, as you saw, the links to the pfree and prealloc routines would need
to be present in standard places in the context header.
        regards, tom lane