On Tue, Sep 27, 2022 at 01:38:26AM +0000, kuroda.hayato@fujitsu.com wrote:
> Maybe you mentioned about sigint_interrupt_enabled,
> and it also seems to be modified in the signal handler.
Yeah, at least as of the cancel callback psql_cancel_callback() that
handle_sigint() would call on SIGINT as this is set by psql. So it
does not seem right to use a boolean rather than a sig_atomic_t in
this case, as well.
--
Michael