Paul Eggert wrote:
> I installed the attached patch
Thanks. Stating which conditions produce undefined behaviour goes a long way
at avoiding programmer pitfalls.
> It's OK for
> behavior to be undefined when the base is bad, just as it's OK for it to
> be undefined if the string is bad (e.g., when nptr == (char *) &errno).
Strictly speaking, yes. What I meant to avoid are programmer mistakes
that can reasonably likely occur.
> I didn't adjust the recently-added Gnulib test cases even though they're
> testing what is now documented to be undefined behavior
Yes, this test now merely serves documentation purposes.
Bruno
2024-07-26 Bruno Haible <br...@clisp.org>
xstrtol, xstrtoll tests: Improve comment.
* tests/test-xstrtol.c (main): Clarify that the new test exercises
undefined behaviour.
diff --git a/tests/test-xstrtol.c b/tests/test-xstrtol.c
index 13ee24219b..2b2b51a47e 100644
--- a/tests/test-xstrtol.c
+++ b/tests/test-xstrtol.c
@@ -67,7 +67,8 @@ main (int argc, char **argv)
{
/* Miscellaneous test cases. */
- /* Test an invalid base. Reported by Alejandro Colomar. */
+ /* Test an invalid base (undefined behaviour, as documented in
xstrtol.h).
+ Reported by Alejandro Colomar. */
{
const char input[] = "k";
char *endp = NULL;
