Re: Is there a memory leak in commit 8561e48?

On Mon, Apr 30, 2018 at 05:10:14PM -0400, Peter Eisentraut wrote:
> The memory leak can be fixed by adding a pfree().
>
> The example you show can be fixed by doing SPI cleanup in both
> transaction abort and exception return to main loop, similar to other
> cases that now have to handle these separately. Patch attached.

I have looked at the patch for some time and tested it, and the approach
looks good to me.

> +/*
> + * Clean up SPI state. Called on transaction end (of non-SPI-internal
> + * transactions) and when returning to the main loop on error.
> + */
> +void
> +SPICleanup(void)
> +{
> + if (_SPI_stack)
> + pfree(_SPI_stack);
> + _SPI_stack = NULL;
> + _SPI_stack_depth = 0;
> + _SPI_current = NULL;
> + _SPI_connected = -1;
> + SPI_processed = 0;
> + SPI_lastoid = InvalidOid;
> + SPI_tuptable = NULL;
> +}
If _SPI_stack is NULL, a quick exit path would make sense?
--
Michael