wait() spec (SIGCHLD ignored)

[Michael Kerrisk <mtk.lists@xxxxxxxxx>]

(I'm not sure if I'm misreading the text, or whether this perhaps 
warrants an Aardvark.)

In both POSIX.1-2001 and the current revision, there is the following 
text for wait():

==
If the calling process has SA_NOCLDWAIT set or has SIGCHLD set to 
SIG_IGN, and the process
has no unwaited-for children that were transformed into zombie 
processes, the calling thread
shall block until all of the children of the process containing the 
calling thread terminate, and
wait() and waitpid() shall fail and set errno to [ECHILD].
==

If I read this right, then the standard is saying that in the following 
scenario where a process has created 2 children, and

a) One child still executing
b) One child already terminated

Then the above standards para does *not* apply ("and the process has no 
unwaited-for children that were transformed into zombie processes"), and 
instead the wait() call should return the status of the child that has 
already terminated.  Is that a correct interpretation.

Notwithstanding that, in this circumstance, on Linux (2.6.25), FreeBSD 
(6.2), and Solaris (8), a waitpid() call does wait until child a) has 
already terminated, and then the call fails with ECHILD.  It looks as 
though the standard deviates from existing practice on this point (or 
did I miss something?).

Cheers,

Michael