Re: Re: musl bugs found through gnulib

[Rich Felker <dalias@aerifal.cx>]

On Mon, Jun 18, 2012 at 05:26:32PM +0200, Szabolcs Nagy wrote:
> * Rich Felker <dalias@aerifal.cx> [2012-06-18 10:55:45 -0400]:
> > > there's only one broken program out there, which is gnu coreutils' "od".
> > > if someone wants to use coreutils and cares about this uncommon
> > > issue, he/she can simply use a patch for this specific program.
> > 
> > The problem is that gnulib is potentially replacing printf in many
> > programs (I haven't checked this; would you care to check?) over this
> > stupid issue that only affects one broken program, resulting in
> > massive bloat for users, and potentially breaking things.
> > 
> 
> hm, what's the issue with od?
> is it about printing random binary data as long double?

Yes.

> if od invokes undefined behaviour then it's a coreutils bug

Yes.

> it sounds bad that based on a broken test gnulib may replace
> a correct printf implementation with its own broken one
> just to make a broken coreutils tool to work so now we
> need a libc workaround for this issue..

Coreutils should at least be changed to call isnanl() in the data
before calling printf. Then they can get by with just replacing isnanl
(ugly but relatively cheap and non-intrusive) instead of replacing all
of printf.

Rich