tests: Fix double expansion bugs in test code
Two test programs - check_path and overlay - define a CHECK() helper macro
in such a way that in the case of an error it will re-execute the checked
code fragment, instead of using the return value from the initial call.
This can lead to misreporting errors, because the code may fail in a
different way the second time around due to changes made during the first
failing call.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
diff --git a/tests/check_path.c b/tests/check_path.c
index 0d6a73b..f12f950 100644
--- a/tests/check_path.c
+++ b/tests/check_path.c
@@ -26,8 +26,9 @@
#define CHECK(code) \
{ \
- if (code) \
- FAIL(#code ": %s", fdt_strerror(code)); \
+ int err = (code); \
+ if (err) \
+ FAIL(#code ": %s", fdt_strerror(err)); \
}
/* 4k ought to be enough for anybody */
diff --git a/tests/overlay.c b/tests/overlay.c
index 961ed60..3093eec 100644
--- a/tests/overlay.c
+++ b/tests/overlay.c
@@ -27,8 +27,9 @@
#define CHECK(code) \
{ \
- if (code) \
- FAIL(#code ": %s", fdt_strerror(code)); \
+ int err = (code); \
+ if (err) \
+ FAIL(#code ": %s", fdt_strerror(err)); \
}
/* 4k ought to be enough for anybody */