Discussion:
float: fix LDBL_MIN value on FreeBSD/x86
(too old to reply)
Bruno Haible
2017-08-17 19:33:01 UTC
Permalink
Raw Message
On FreeBSD/x86, I'm seeing this test failure:


FAIL: test-float
================

../../gltests/test-float.c:334: assertion 'm == x' failed
FAIL test-float (exit status: 134)


It's caused by gnulib itself. Here's the fix:


2017-08-17 Bruno Haible <***@clisp.org>

float: Fix LDBL_MIN value on FreeBSD/x86.
* lib/float.in.h (LDBL_MIN) [__FreeBSD__]: Add more precision.

diff --git a/lib/float.in.h b/lib/float.in.h
index b51aafe..8105d11 100644
--- a/lib/float.in.h
+++ b/lib/float.in.h
@@ -81,7 +81,7 @@
# define LDBL_MAX_EXP 16384
/* Minimum positive normalized number. */
# undef LDBL_MIN
-# define LDBL_MIN 3.3621031431120935E-4932L /* = 0x1p-16382L */
+# define LDBL_MIN 3.362103143112093506262677817321752E-4932L /* = 0x1p-16382L */
/* Maximum representable finite number. */
# undef LDBL_MAX
/* LDBL_MAX is represented as { 0xFFFFFFFF, 0xFFFFFFFF, 32766 }.

Loading...