Test suite error on 32bit systems

[tillea]

Hi,
the Debian packaged data.table received a bug report when running the test suite on 32bit systems. We are running CI tests on several release architectures. If you look at our tracker page you see Regression links (marked in red color) for the 32 bit architectures. If you follow these links for instance for i386 you will find the full regression log which ends in:

Error in test.data.table() : 
   1 error out of 8547. Search tests/tests.Rraw for test number 1702.4.

The armhf/armel logs are schowing the same error. Please let me know if you need more information.
Kind regards, Andreas.

[jangorecki]

Thank you for reporting. Warning was emitted from base R.

[MichaelChirico]

For posterity, here's the source in base R:

https://github.com/r-devel/r-svn/blob/505c30228bfb5a671440d519e1b3cf3382e50ff1/src/main/datetime.c#L509-L529

it specifically notes that this is a 32-bit issue. The culprit test should not really be affected so agree with suppressing the warning.

It's also a recent change in R: r-devel/r-svn@ee888dd

[jangorecki]

If test will actually be affected it will fail due to inequality check in test() anyway. No need for failing due to base R warning about possibly inaccurate date.