Currently in several places of the Julia Manual random number generation is used in examples. We require explicit using Random for those examples to work (now it will throw deprecation warning but in Julia 1.0 those examples will throw an error).
This means that novice Julia users when trying to copy-paste an example will not understand why it fails.
There are three possible solutions:
using Random in the examples where needed to make them work stand alone;My personal preference would be option 2 (adding using Random where needed to the examples - the downside is adds a bit of noise to the examples), but maybe there are other opinions.
I will make a PR with the appropriate changes when I we have an opinion.
+1 for option 2, using Random. We did that recently for an example that required Sockets.
Currently in several places of the Julia Manual random number generation is used in examples.
Hmm... I didn't find any of those, only calls to rand(...) which works without using Random.
@fredrikekre I thought that only rand(...) is exported. Now I have checked that also randn(...) can be used without using Random which I did not know (and it is used in several places). So this leaves us with only one line
julia> r = randjump(MersenneTwister(1), big(10)^20, nthreads());
in parallel-computing.md which uses non-exported methods. I will add using Random to this example then.
Most helpful comment
@fredrikekre I thought that only
rand(...)is exported. Now I have checked that alsorandn(...)can be used withoutusing Randomwhich I did not know (and it is used in several places). So this leaves us with only one linein parallel-computing.md which uses non-exported methods. I will add
using Randomto this example then.