[PATCH] Updates on busybox POSIX compliance

Rob Landley rob at landley.net
Fri Jun 19 20:44:29 UTC 2009


On Friday 19 June 2009 14:35:12 Denys Vlasenko wrote:
> On Friday 19 June 2009 10:16, Rob Landley wrote:
> > No, the entire point of these features is to serve the needs of users out
> > there, most of which don't really care what the gnu tools do.  They just
> > care that their scripts run.
>
> Last I checked, there are 3.14159269 zillions of desktop users.

20x as many of which use Macs as use Linux.

Focusing on Linux and focusing on the desktop is pretty much a contradiction 
in terms these days, we've actually _lost_ desktop market share in the past 
year.  (Still doing fine in the server and embedded niches, though.)  If you 
really want desktop users, port the sucker to the macintosh.

> Don't know about you, but I think whenever some of them install
> busybox on their machine it is a win for the project.
>
> Our ls has -S option. It is not in the POSIX. Why do we have it?

Because it makes users' scripts run?

And it's easy to test without resorting to comparing with some other specific 
implementation, which was my point.  Something like:

  mkdir sub
  cd sub || exit 1
  echo -ne "500 abc\n" "100 def\n" "1000 ghi\n" | while read SIZE NAME
  do
    dd if=/dev/zero bs=$SIZE count=1 of=$NAME 2>/dev/null
  done

  testing "ls -1" "ls -1" "abc\ndef\nghi\n" "" ""
  testing "ls -S1" "ls -S1" "ghi\nabc\ndef\n" "" ""

  cd ..
  rm -rf sub

Pondering that, testing that the output when going to a tty is all on one line 
and the output when going to a pipe isn't could get tricky, you'd need to set 
up a pty to do it right and I dunno how to do that from the command line off 
the top of my head...  I'd have to look into it.

> it will be even more effective in alienating a huge potential
> user base than emails on some mailing list.

I'm sure there are plenty of more effective ways we could drive users away if 
we decided that was a goal.

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds


More information about the busybox mailing list