Devrim GÜNDÜZ wrote
>> but yeah, trying to build it on a modern Linux system looks
>> a bit shaky. If you do go this route, I'd recommend making sure it
>> passes its regression tests before you risk letting it touch your real
>> data.
>>
>
> geometry and horology tests are failing here. regression.diffs is
> attached for reference.
>
same here on CentOS 5.x, except the geometry test had fewer failures
logged, only the "six | box" one, and the differences were
.000000000000001 rounding kinda things, like -0.535533905932738 came out
-0.535533905932737.
the horology tests seem to be all about PST vs PDT and may well be
because Mar 15 is now PDT when it didn't used to be?
I'm guessing this is good enough to dump his data from.