Re: Avoid resource leak (src/test/regress/pg_regress.c)

Em sex., 24 de out. de 2025 às 03:03, Michael Paquier <michael@paquier.xyz>
escreveu:

> On Thu, Oct 23, 2025 at 09:37:21PM -0300, Ranier Vilela wrote:
> > The function *config_sspi_auth* is responsible for
> > rewrite pg_hba.conf and pg_ident.conf to use SSPI authentication.
> >
> > Coverity complains that the struct addrinfo gai_result is leaked.
> > The variable is declared inside block and is not used
> > outside the block.
> >
> > So if the function WSAStartup is successful then the function getaddrinfo
> > allocates and fills the struct addrinfo.
> >
> > The memory must be released at the end of the block.
>
> Not sure that this one is worth caring about.  We have a bunch of
> allocations that we know would be freed once a binary exits.

Fortunately, this happens less and less.


>   This is
> just one of them, allocated in the context of what is a short-term
> execution.
>
In this case, I believe pg_regress is run, thousands of times, on hundreds
of computers.

best regards,
Ranier Vilela