From owner-freebsd-bugs  Mon Nov 24 01:11:20 1997
Return-Path: <owner-freebsd-bugs>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id BAA18584
          for bugs-outgoing; Mon, 24 Nov 1997 01:11:20 -0800 (PST)
          (envelope-from owner-freebsd-bugs)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id BAA18576;
          Mon, 24 Nov 1997 01:11:11 -0800 (PST)
          (envelope-from bde@zeta.org.au)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.7/8.6.9) id UAA06647; Mon, 24 Nov 1997 20:01:52 +1100
Date: Mon, 24 Nov 1997 20:01:52 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199711240901.UAA06647@godzilla.zeta.org.au>
To: cagney@tpgi.com.au, freebsd-bugs@FreeBSD.ORG, joerg@FreeBSD.ORG
Subject: Re: bin/5135
Sender: owner-freebsd-bugs@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

>Synopsis: scanf accepts %Lg but treates it as %g
>
>State-Changed-From-To: open-closed
>State-Changed-By: joerg
>State-Changed-When: Mon Nov 24 08:49:18 MET 1997
>State-Changed-Why: 
>
>Fixed by Bruce in rev 1.12 of src/lib/libc/stdio/vfscanf.c a couple of
>days ago (also merged to 2.2-stable).

I even reported it to bug-gdb.

Of course, the simple fix of reading long doubles as doubles doesn't
actually work if the input has more than double precision or is too
large for a double.  The corresponding simplistic conversions for
printf causes more problems:
---
#include <stdio.h>

main()
{
    fprintf(stderr, "%.20Lg\n", 12345678901234567890.0L);
    fprintf(stderr, "%.20Lg\n", 12345678901234567890.0e1234L);
}
---
Here the first printf loses precision and the second printf causes
SIGFPE for overflow (unless overflow exceptions are masked; then
then it loses precision and the result is "Inf").

Bruce