Hmmm, I hadn't paid attention to leap seconds.
The problem I have is, I can only try musl on machines where the system
clock is UTC (and I can't change it). The machine I have in TAI-10 does
not have a suitable gcc for musl: it's Aboriginal Linux uClibc-based gcc,
which prevents the musl-gcc wrapper from working.
But I tried the test you linked on a UTC machine, and the results were
fishy indeed, whether the musl version was static or dynamoc.
$ TZ=CEST ./localtime-musl ; TZ=CEST ./localtime-glibc
Local: 06:13:34
Global: 06:13:34 TUU
Local: 06:13:34
Global: 06:13:34
$ TZ=:right/Europe/Paris ./localtime-musl ; TZ=:right/Europe/Paris ./localtime-glibc
Local: 08:14:10
Global: 06:14:10 TUU
Local: 08:13:45
Global: 06:13:45
What's that 'TUU' in the first place ? doesn't look good to me. Also, in the
right/ timezone test, musl and glibc are clearly giving different results - with a
25 second difference, which screams "leap second table problem". A further test
shows that glibc gives the right time: TZ=:right/Europe/Paris is 25 seconds
earlier than TZ=CEST, which is correct.
I will do more testing, then file a report on the musl mailing-list.
--
Laurent
Received on Sat Oct 26 2013 - 06:23:23 UTC