Re: PIPESTATUS differs from $? for compound command

[Dale R. Worley]

Chet Ramey <chet.ramey@case.edu> writes:
> On 12/11/24 9:54 PM, Dale R. Worley wrote:
>> This needs to be corrected to note that compound commands that are not
>> one component of a multi-command pipeline do not cound as pipelines for
>> setting PIPESTATUS.  Exactly which compounds commands don't set
>> PIPESTATUS needs to be made clear as it's not at all obvious.
>
> I'll add something enumerating the commands that set PIPESTATUS.

Well, given that the documentation as written is that a solo compound
command is a pipeline and by implication sets PIPESTATUS, the edit is
actually "enumerating the commands that *don't* set PIPESTATUS if they
are the only command in a pipeline".

Dale