On Thu, 14 May 2026 at 09:04, Tomasz Kaminski <[email protected]> wrote: > > > > On Wed, May 13, 2026 at 10:01 PM Jonathan Wakely <[email protected]> wrote: >> >> libstdc++-v3/ChangeLog: >> >> * testsuite/std/time/clock/utc/leap_second_info-2.cc: Fix sign >> compare warning. Tweak comment. >> --- >> >> Tested x86_64-linux. Pushed to trunk. >> >> .../testsuite/std/time/clock/utc/leap_second_info-2.cc | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/libstdc++-v3/testsuite/std/time/clock/utc/leap_second_info-2.cc >> b/libstdc++-v3/testsuite/std/time/clock/utc/leap_second_info-2.cc >> index 090a28291c85..59634b34944f 100644 >> --- a/libstdc++-v3/testsuite/std/time/clock/utc/leap_second_info-2.cc >> +++ b/libstdc++-v3/testsuite/std/time/clock/utc/leap_second_info-2.cc >> @@ -52,7 +52,7 @@ test_future_leaps() >> std::ofstream("leapseconds") << ""; >> >> using std::chrono::years; >> - auto s = std::chrono::utc_seconds(1483228826s + years(100)); // 1 Jan 2117 >> + auto s = std::chrono::utc_seconds(1483228826s + years(100)); // 2117 >> auto lsi = get_leap_second_info(s); >> VERIFY( override_used ); // If this fails then XFAIL for the target. >> VERIFY( lsi.is_leap_second == false ); >> @@ -91,7 +91,7 @@ Leap 2120 Dec 30 23:59:59 - S >> // The file was not read again: >> VERIFY( ! override_used ); >> // The list in the tzdb contains the three fake leap seconds: >> - VERIFY( tzdb.leap_seconds.size() == hardcoded_count.count() + 3 ); >> + VERIFY( tzdb.leap_seconds.size() == size_t(hardcoded_count.count() + 3) ); > > I would preffer +3u, but it is fine as is.
I would prefer that too, but chrono::seconds::rep is int64_t and when you add 3u to that you still get int64_t, which is still signed and so still gives a warning when compared to size_t. 3ull would work. >> >> // And repeating the queries above gives the same results: >> lsi = get_leap_second_info(s); >> VERIFY( lsi.is_leap_second == false ); >> -- >> 2.54.0 >>
