Roland Mainz wrote: >>> cksum(1) and specified by IEEE Std 1003.1-2008 is "Committed". >> Why are you explicitly referencing IEEE Std 1003.1-2008? Isn't Committed >> enough? > > That's intended to be a wave with an axe, following the original rule of > the ksh93-integration project "... we _first_ conform to POSIX and > _then_ add features on top of it and we're _NOT_ breaking the POSIX/SUS > test suites in _any_ case...". > Or short: We conform to IEEE Std 1003.1-2008. > Additionally the manpage describes lots of features beyond the POSIX/SUS > specificcation and therefore I added links to the IEEE Std 1003.1-2008 > manpages that people who need to write portable scripts can check which > features are defined by the standard and which are extensions from our > side.
Which is another way of saying not only is the output of cksum(1) Committed it is actually the output we expect not something else. Just Committed would mean that in theory the output could be something other than what 1003.1-2008 required. So it is good that Roland has been very specific here. -- Darren J Moffat
