Re: signal(3) "new" signals warning

[Crystal Kolipe <kolipe.c%exoticsilicon.com@localhost>]

On Thu, Jul 24, 2025 at 04:38:29AM +0300, Valery Ushakov wrote:
> Out signal(3) man page screams at the reader with emphasis:
> 
>    Unlike previous signal facilities, the handler func() remains
>    installed after a signal has been delivered.
> 
> This paragraph and its supposed importance is probably a bit opaque
> for a modern reader.  It has not been relevant in ... 35-40 years?  It
> refers to the historic behavior (v7, 4BSD), where signal disposition
> was reset to SIG_DFL after a signal was handled.
> 
> I think it's time we reword this and move it to HISTORY, but my memory
> is hazy.  It seems 4.1BSD was transitional, and 4.2BSD had new signals
> as the default and introduced the above phrase in signal(3).
> Feedback, clarifications and proposed new wording are appreciated.

I'm pretty sure 4.1BSD had the old behaviour and 4.2BSD introduced the SOUSIG
flag to optionally re-enable it.  (I vaguely remember that it may also have
had another function which was later removed, although that may be incorrect.)