Re: Possible bug in uname command

[Bob Proulx]

Alfred M. Szmidt wrote:
>    I don't think it will break scripts because legacy operating
>    systems don't support those options either.
                                         ^^^^^^

> If you consider GNU a legacy operating system, sure.  Recall, GNU
> coreutils is for GNU, not non-GNU systems.

Notice that I said "either".

Does Hurd support the interface to supply 'uname -p' information?  The
implication from reading this is your statement is that it does not.
In which case 'uname -p' on GNU Hurd is not useful either.

> How about those options get disabled if POSIXLY_CORRECT is defined
> instead?  That makes sense, and doesn't break scripts.

Can you give an example of scripts that would use 'uname -p' and why
they are not already broken now?  Because basically when dealing with
a command as non-portable as uname you should take these differences
into consideration now already.  The only portable way to use uname is
without any options and then work from there after you know what
system you are on.  Look at config.guess for lots of examples of uname
issues.  Basically uname is awful.

Bob