reported 17 mismatches between busybox and GNU coreutils

[Rob Landley]

On Tuesday 11 November 2008 22:00:46 Philip Guo wrote:
> > You opened your bugs as "standards compliance" without referencing a
> > standard.
>
> sorry again, this is my first time posting on this list and in the busybox
> bug database :)

Not trying to discourage you.  Testing is never a bad thing... :)

> > Beyond that, lots of the behavior in coreutils is sheer bloat.  I looked
> > at several of your things and you're asking us to make the code bigger,
> > in ways
> > not required by any standard, in order to provide features of negligible
> > utility.
>
> i totally agree, and perhaps i shouldn't have placed those records in the
> bug database in the first place, and instead simply emailed them as
> questions.

It's good that you reported them (thank you for that), and it's nice to look 
at them.  There are a few valid bugs hiding in there (not returning error for 
file not found is probably our bad).  But saying "no" is a valid option for a 
lot of it.

> i can see now that putting them in the database implies that i 
> thought they were bugs in busybox, which i didn't mean at all.  i thought
> it would be a good location to have discussions about whether those are
> actually bugs.

The mailing list is good for discussion.  The bug tracker is for tracking 
issues so they don't get forgotten if we aren't going to fix them 
immediately.

> > In fact, in some cases:
> >   http://busybox.net/bugs/view.php?id=6174
> >
> > You're asking us to implement undocumented behavior that looks like a bug
> > to
> > me.  You asked the coreutils to unset a variable, you gave it an insane
> > variable name (environment variable names shouldn't have = in them), and
> > it did something other than what you asked.  The man page doesn't suggest
> > env has the behavior you found, in fact it says it _won't_ do that.
>
> yep, i agree; that seems like a weird thing for coreutils to do, and i will
> report this issue to them; i'm fairly certain that busybox is correct in
> this case.

Yeah, so was I, but Denys went ahead and changed it anyway.

*shrug*

Rob