On 2020-Sep-03, Tom Lane wrote:
> As for the question of SIGQUIT handling, I see that postgres.c
> does "PG_SETMASK(&BlockSig)" immediately after applying the sigdelset
> change, so there probably isn't any harm in having the background
> processes do likewise. I wonder though why bgworkers are not
> applying the same policy.
It's quite likely that it's the way it is more by accident than because
I was thinking extremely carefully about signal handling when originally
writing that code. Some parts of that code I was copying from others'
patches, and I could easily have missed a detail like this. (I didn't
"git blame" to verify that these parts are mine, though).
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services