On 07/04/2026 10:26, Bruno Haible via GNU coreutils General Discussion wrote:
Today's CI shows a test failure of numfmt/numfmt on macOS and on OpenBSD.
I pushed the attached to avoid this false failure
on systems without long double support.
thanks,
Padraig
From 5b0cb4379abfe328bc6a3838fbc59ebf6b39b720 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?P=C3=A1draig=20Brady?= <[email protected]>
Date: Tue, 7 Apr 2026 21:44:56 +0100
Subject: [PATCH] tests: numfmt: avoid false failure on systems without long
double
* tests/numfmt/numfmt.pl: Move recently added test that depends
on long double support to the appropriately guarded set.
Also reduce the value to be in the definitely safe long double range.
Reported by Bruno Haible.
---
tests/numfmt/numfmt.pl | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/tests/numfmt/numfmt.pl b/tests/numfmt/numfmt.pl
index b78aef14b..c6ed3f073 100755
--- a/tests/numfmt/numfmt.pl
+++ b/tests/numfmt/numfmt.pl
@@ -898,11 +898,6 @@ my @Tests =
"Try '$prog --help' for more information.\n"},
{EXIT => 1}],
- # Large integers beyond 2^53 must be preserved exactly
- # (no f64 round-trip).
- ['large-int-precision', '--from=iec 9153396227555392131',
- {OUT => "9153396227555392131"}],
-
# Scientific notation ('1e9') is not accepted as numeric input;
# it is reported as an invalid suffix.
['reject-sci-notation', '1e9',
@@ -962,6 +957,11 @@ my @Limit_Tests =
['large-1','1000000000000000', {OUT=>"1000000000000000"}],
# 18 digits is OK
['large-2','1000000000000000000', {OUT=>"1000000000000000000"}],
+
+ # Large integers beyond 2^53 must be preserved exactly
+ # (no double round-trip).
+ ['large-2a', '915339622755539213', {OUT => "915339622755539213"}],
+
# 19 digits is too much (without output scaling)
['large-3','10000000000000000000',
{ERR => "$prog: value too large to be printed: '1e+19' " .
--
2.53.0
