[systemd-devel] Antw: Re: Antw: systemd prerelease 243-rc2

[Reindl Harald]

Am 26.08.19 um 09:14 schrieb Ulrich Windl:
>>> I doubt it's increasing robustness for any existing application as
>>> pid_traditionally was 16 bit. I don't know if some applications try to
>>> sprintf() a pid into a char[6], but if they do, it might cause an 
>> application
>>> failure...
>>
>>   What kind of tradition would that be?  Could you please point the
>> specific standard and implentation?  Everywhere I look pid_t is
>> "signed integer type" which is implemented as 32 bits.
> 
> pid_t is quite new, also. In System V a pid_t was simply an int, but no PID
> ever exceeded 2^15.
> That observation could have lead programmers to some false expectations.
> Also Using pid_t does not help you to get the correct sprintf buffer size
> unless you look how pid_t is actually defined, which -- in turn -- may be
> platform-dependent.

frankly can you stop making a fuss about such a simple default change
whcih was done here long ago in sysctl.conf on 50 systems from desktops
to all sort of servers with not a single issue?

yeah, programmers could have made wrong expectations, then they need to
fix them, you don't get this change on your SLES for years and in the
meantime all Fedora users will help clean it up

the only issue i remember was some cli tool which didn't get rioght it's
column display and was already fixed