Noah Misch

I've reviewed this. Note that the linked issues say this began with PostgreSQL 15 commit postgres/postgres@4eb2176318d, but it was Windows-only until postgres/postgres@e2f65f4255. I do recommend merging this PR as-is, but...

This achieves a lot for a one-line change. @eulerto (since no other person has merged a pglogical commit in the last eight months), would you merge this? If not, what...

@petere I see you are the other of the two people who have committed to pglogical in the last 2y. This pull request achieves a lot for a one-line change....

+1. Also, that section contains URL https://www.2ndquadrant.com/en/resources/pglogical/, which reaches "Page not found".

@eulerto @petere @mwanner This is just adding a CHECK_FOR_INTERRUPTS() so DROP DATABASE does not hang. For 11 months, the pull request has been waiting attention from someone with repository write...

@eulerto This is still just adding one CHECK_FOR_INTERRUPTS(), and it's been waiting 14 months for someone with repository write access. Adding a CHECK_FOR_INTERRUPTS() is one of the safest changes doable...

Interesting. I agree it's a bug, though I'm not thinking of a way it would harm a program other than a program going out of its way to be harmed....

> Perhaps a similar patch can be applied at line 3594 in the result method? Agreed; probably like this: ``` diff --git a/lib/IPC/Run.pm b/lib/IPC/Run.pm index 77863cb..7b7555e 100644 --- a/lib/IPC/Run.pm +++...

I was able to reproduce it by ignoring `SIGCHLD`: ``` use IPC::Run 'start'; use strict; use warnings; $SIG{CHLD} = 'IGNORE'; my $h = start ['cat'], 'pipe', \*OUT, '2>pipe', \*ERR or...

That did not succeed. The SIGTERM is sometimes lost if it arrives between the start and end of the child's execve(). kdump excerpt: ``` 2024-05-19T23:31:23.1388650Z 7858 7858 perl CALL execve(0x7315a6eec320,0x7315a6eec340,0x7315a8aae000)...