[Rd] Failing lm-tests due to extra 0 in scientific notation?

[Prof Brian Ripley]

On 07/01/2015 06:24, Avraham Adler wrote:
> Hello.
>
> I've compiled R on Windows many times, and this is the first time I've
> seen this error. While running make check-all (and using
> testInstalledBasic("both")), the lm-tests routines fail, and, as far
> as I can tell, the diff is failing because in one file, answers are
> coming back like this "3.11e-004" while in the save file they are
> "3.11e-04". Every value is the same, outside the extra 0 in the
> scientific notation. I've never seen R put two 0s in a row like that
> before, and I cannot think of why that would happen. Is there a way to
> change that so that it passes the tests?

That is what the Windows runtime does, so it seems you did something  in 
your compilation that linked to the wrong printf function.  R on Windows 
should use that from the (modified version of) the trio library in the 
sources.

You failed to follow the posting guide: we do not even know the version 
of R nor if this is 32- or 64-bit nor the locale ....

But as one data point, a 64-bit build of current R-patched from SVN 
checked for me a couple of hours ago.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Emeritus Professor of Applied Statistics, University of Oxford
1 South Parks Road, Oxford OX1 3TG, UK