Re: no exit status available when a command doesn't exist or fails in pipe to grep

[Ed Morton]

On 9/9/2021 5:04 PM, Neil R. Ormos wrote:
> Ed Morton wrote:
> > Neil R. Ormos wrote:
> > > the notion that we somehow don't know what "if
> > > the close was successful" means is a fiction.
> > We'll have to agree to disagree. If someone told
> > me to write a function foo() to release some
> > resource X and return success if foo() succeeded
> > my first question would be "what should it
> > return if X is already idle".
> The inability to close() something because it was already closed or never opened in the 
> first place is recognized as an "error" and the value returned by close() is 
> -1.  See the manual.
That's good that we agree it's an error but, since we're talking about 
POSIX compliance, see the POSIX spec for the value returned by close() 
(https://pubs.opengroup.org/onlinepubs/9699919799/utilities/awk.html#tag_20_06_13_14 
under "Input/Output and General Functions"):

   close(expression)
        Close the file or pipe opened by a print or printf statement or
   a call to getline with the same string-valued expression. The limit
   on the number of open expression arguments is
   implementation-defined. If the close was successful, the function
   shall return zero; otherwise, it shall return non-zero.

It doesn't say close() will return -1 on an error, it says if the call 
wasn't successful it'll return non-zero.

>   That is the "otherwise" condition in the POSIX spec.
>
> As demonstrated earlier, a non-zero return code from some piped command (or the 
> shell) does not defeat the operation of close().
>
> Reiterating: it's Arnold and Andy you have to convince.
I'm not trying to convince anyone, I made a suggestion they'll do 
whatever they think makes sense. I'm just responding to your emails to 
me about the subject, I was done with everything I had wanted to say 
when Arnold said "I will think about it.". I am done now though.

    Ed.

>