[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [CTTE #770789] IEC units in df output

On Mon, Dec 15, 2014 at 09:57:34PM +0100, Adam Borowski wrote:
> On Mon, Dec 15, 2014 at 10:38:33AM -0800, Don Armstrong wrote:
> > ==== RESOLUTION ====

> > In 770789, the Technical Committee was asked to override the decision
> > of upstream and the maintainer of df to not include i in the units
> > output when asked for IEC output (2^10).

> > The CTTE declines to override the decision of the maintainer and
> > upstream.

> > ==== END OF RESOLUTION ====

> Thank you guys!  The -ebi- plague is horrible and needs to be fought!
> You could have responded using some of Ted Ts'o arguments and quotations
> from #757831.

This is not a statement by the TC about the appropriateness of these prefix
conventions, one way or the other.  It is only a statement that the TC is
not overriding the maintainer (and upstream) regarding this decision.  I
don't think it's appropriate for the maintainer of the Debian package to
diverge from upstream on something like this; and furthermore I don't think
it's reasonable for the TC to overrule the maintainer's decision to follow
upstream.  That doesn't mean I think the current behavior is /correct/, only
that I don't think a TC override is an appropriate method of getting this
behavior changed.

If this is going to be changed, it should be done by getting consensus
upstream about how to change the behavior - not by making such a core tool
as df behave inconsistently between Debian and other distributions.

While my gut reaction to the IEC convention when I first became aware of it
was one of distaste, traditional behavior *is* ambiguous, and this ambiguity
(including behavior that differs between applications) is confusing.  Ubuntu
ratified a units policy several years ago that I believe strikes the right


For a tool such as df, there would be some benefit to being consistent with
this policy, by showing Ki/Mi/Gi instead of K/M/G in the output where
suitable.  But there is also a cost to the extra space used up in the line
by changing from a one-character suffix to a two-character suffix in each of
the columns.  I think it's appropriate for the Debian maintainer and the
coreutils upstream to do their own analysis of the cost/benefit tradeoff
here, without the TC presuming to meddle.

Steve Langasek                   Give me a lever long enough and a Free OS
Debian Developer                   to set it on, and I can move the world.
Ubuntu Developer                                    http://www.debian.org/
slangasek@ubuntu.com                                     vorlon@debian.org

Attachment: signature.asc
Description: Digital signature

Reply to: