Julia: Unsafe literal constants with powers (e.g. 10ˆ10 in 32 bit systems)

[mschauer]

I think this example might fail on 32 bit systems.

https://github.com/JuliaLang/julia/blob/df1ee8cb185bbf75999384ce882a89cd9ca8d3fd/base/math.jl#L533

[mschauer]

This is also an instance, right? https://github.com/JuliaLang/julia/blob/df1ee8cb185bbf75999384ce882a89cd9ca8d3fd/base/libc.jl#L387

[mschauer]

For reference, I looked for examples for example in

grep -r --include="*.jl" [^\.][0-9]\^[0-9] . | less

[mschauer]

https://github.com/JuliaLang/julia/blob/df1ee8cb185bbf75999384ce882a89cd9ca8d3fd/test/rounding.jl#L136

[mschauer]

https://github.com/JuliaLang/julia/blob/df1ee8cb185bbf75999384ce882a89cd9ca8d3fd/test/hashing.jl#L75

[mschauer]

... and I did this to convince you to reopen #5573 ;-) This problem permeates the entire ecosystem, https://pkg.julialang.org/docs/search?q=2%5C%5E%5B3-9%5D%5B0-9%5D%7C%5B%5E%5C.%5D%5B0-9%5D%7B2,%7D%5C%5E%5B0-9%5D%7B2,%7D

[StefanKarpinski]

As a somewhat broader point, maybe we should run doctests on both a 64- and 32-bit platform?

[StefanKarpinski]

Also, what would you like the behavior here to be?

[mschauer]

As a somewhat broader point, maybe we should run doctests on both a 64- and 32-bit platform?

A somewhat minimal requirement is to make passing 32-bit doctests mandatory for the release candidates?

[ararslan]

@staticfloat, could we just add an extra CI build for 32-bit doctests?

[mschauer]

Also, what would you like the behavior here to be?

I think I would like ^ with literal arguments use a checked version of power_by_squaring (which would make sense if there is no performance penalty due to constant propagation), but I cannot say if this realistic.

[KristofferC]

Lots of things print differently on 32bit so running the doctests there are not that easy.

[StefanKarpinski]

Even though the proximate issue is dealt with perhaps I should leave this open for a bit more discussion of the general problem.

[staticfloat]

There are a very large number of doctest errors when run on a 32bit machine, because many of the outputs expect e.g. ::Array{Int64,1}, when they instead get ::Array{Int32,1}. I'm not sure it's worth our time to try and generify this, although perhaps it is! I can create a doctest_linux32 builder trivially; but it will just add to the noise unless we go through and change a large proportion of the current docstrings.

[fredrikekre]

I'm not sure it's worth our time to try and generify this, although perhaps it is!

It's not.