Unable to make BigFloat with different precision

[dpsanders]

julia> x = big"1.0";

julia> precision(x)
256

julia> y = BigFloat(x, 100);

julia> precision(y)
256

This should make a BigFloat with precision 100.

[simonbyrne]

So the question is: should BigFloat(x::BigFloat) be able to give a result with different precision than x? What about convert(::Type{BigFloat}, x::BigFloat)?

[dpsanders]

I seem to have missed your comment, sorry.

I would say that following through the current policy of having a global precision for BigFloat construction means that indeed BigFloat(x::BigFloat) should convert x to a BigFloat with the current global precision.

I do see that convert(::Type{BigFloat}, x::BigFloat) is questionable, though...

[StefanKarpinski]

Yet again, I find myself feeling that BigFloat should have a precision parameter.

[dpsanders]

Yes!