Re: [dev] sbase: od: why avoid signed numbers?

From: FRIGN <>
Date: Tue, 27 Oct 2015 17:19:05 +0100

On Tue, 27 Oct 2015 11:06:07 -0400
Greg Reagle <> wrote:

Hey Greg,

> Hello. As you probably already know, the standard specifies that type
> 'd' means signed decimal. I am curious about why:
> "the 'd' parameter for the [-t] flag is interpreted as 'u'"

this decision was made because handling signs is a complex matter. We
dropped it as we kinda realized signed integers somehow aren't of much
use anyway. Prove me wrong if I am.
In the general case, I welcome anybody who'd like to take the challenge
of doing signed integers in od(1). We don't generally exclude it, but if
it brings too much complexity, we might not accept it.

> Given this difference, is od actually POSIX compliant?

Not in this regard obviously, but as you know, we sometimes move away
from POSIX where it seems to be necessary.



Received on Tue Oct 27 2015 - 17:19:05 CET

This archive was generated by hypermail 2.3.0 : Tue Oct 27 2015 - 17:24:09 CET