IRC channel logs

2025-10-12.log

back to list of logs

<azeem>youpi: the (x86-64) Postgres buildfarm animal had one case of SIGABRT a while ago, where apparently a signal with a bogus signal number came in, see https://www.postgresql.org/message-id/flat/CA%2BhUKGKo6rToAxyObEnOnOiT3QEJuJFVkVNqikw0wVFHag%3D3Fg%40mail.gmail.com#3e171bcd1a51999d2c8ee39ea87c6248 
<azeem>the sub-thread starts at https://www.postgresql.org/message-id/flat/CA%2BhUKGJE_moUF74c97GkoP6RaknRMoeFOednXe2FyXnS_bOTFQ%40mail.gmail.com#da918a20abc4e4f076ad7b34afeb8f95 
<azeem>maybe that's a x86-64 problem, as it is much newer as a port?
<youpi>azeem: there could be some bug in the 64b version of delivering signals, yes
<youpi>28476608 is clearly some value that overflow somehow
<azeem>I'll try to reproduce it on 32bit - Alexander seems to be able to readily reproducei t