Re: Report a potential memory leak in setup_config()

Andres Freund <andres(at)anarazel(dot)de> writes:
> When I'd looked at this last I apparently (looking at git stash, I ended up
> discarding this) decided that the best way would be to change replace_token's
> API to one that just processes one line at a time, with an outer loop that
> processes all tokens in a line. I'm not really sure why anymore.

Hmm, I did it the other way, as attached. This gets it down to

==3254266== LEAK SUMMARY:
==3254266== definitely lost: 342 bytes in 17 blocks
==3254266== indirectly lost: 152 bytes in 2 blocks
==3254266== possibly lost: 0 bytes in 0 blocks
==3254266== still reachable: 2,403 bytes in 22 blocks
==3254266== suppressed: 0 bytes in 0 blocks

It seems that actually the pointer arrays *are* a big chunk of
the leakage, because the individual strings get freed in the
output loops! That's a bit ugly but I feel no need to change it.

regards, tom lane