In the shell test framework, an error is either flagged by setting
the shell variable fail=1, or by invoking the fail_ function with an
error diagnostic.
* tests/init.sh (returns_): Change example code in the comment to
use '|| fail=1' properly.
---
ChangeLog | 9 +++++++++
tests/init.sh | 2 +-
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/ChangeLog b/ChangeLog
index 779e593a14..8e21997d6d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,12 @@
+2024-08-04 Bernhard Voelker <m...@bernhard-voelker.de>
+
+ test-framework-sh: Improve returns_ example.
+ In the shell test framework, an error is either flagged by setting
+ the shell variable fail=1, or by invoking the fail_ function with an
+ error diagnostic.
+ * tests/init.sh (returns_): Change example code in the comment to
+ use '|| fail=1' properly.
+
2024-08-03 Bruno Haible <br...@clisp.org>
gnulib-tool.py: Small refactoring.
diff --git a/tests/init.sh b/tests/init.sh
index d33ef61788..3bd1e3b9cd 100644
--- a/tests/init.sh
+++ b/tests/init.sh
@@ -595,7 +595,7 @@ fi
# I.e., just doing `command ... &&fail=1` will not catch
# a segfault in command for example. With this helper you
# instead check an explicit exit code like
-# returns_ 1 command ... || fail
+# returns_ 1 command ... || fail=1
returns_ () {
# Disable tracing so it doesn't interfere with stderr of the wrapped command
{ local is_tracing=`{ :; } 2>&1`; } 2>/dev/null
--
2.45.2
