UCL / LatBo.jl

Lattice-Boltzmann implementation in Julia
Other
19 stars 3 forks source link

test fail on julia-0.5 #16

Closed nlooije closed 7 years ago

nlooije commented 7 years ago

Hi,

I was trying out your module and found that there are several errors while running Pkg.test("LatBo"):

ERROR: LoadError: LoadError: UndefVarError: FloatingPoint not defined while loading .\LatBo\src\LatBo.jl, in expression starting on line 8

FloatingPoint is no longer supported. Changing to AbstractFloat yields the next error:

ERROR: LoadError: LoadError: LoadError: syntax: "<" is not a unary operator while loading .\LatBo\src\indices/indices.jl, in expression starting on line 17

which is solved by changing from =< to <=. The next error has me stumped:

ERROR: LoadError: LoadError: LoadError: MethodError: no method matching start(::Type{Integer}) Closest candidates are: start(::SimpleVector) at essentials.jl:170 start(::Base.MethodList) at reflection.jl:258 start(::IntSet) at intset.jl:184 while loading C:\Users\niels.julia\v0.5\LatBo\src\indices/indices.jl, in expression starting on line 21

I don't have enough experience with Julia to fix this error. Can you update this for Julia-0.5, please?

mdavezac commented 7 years ago

I'll take a look. I think it comes from an older way specifying variational arguments.

index(kernel::Indexing, indices::(Integer...)) = index(kernel::Indexing, collect(indices))

should now read

index{I <: Integer}(kernel::Indexing, indices::Vararg{I}) = index(kernel::Indexing, collect(indices))

For what it is worth, I'm not sure I've chose the most efficient way to do Lattice Boltzmann in Julia, especially with respect to this tip.

mdavezac commented 7 years ago

Got the tests running locally. Now need to check the examples, at least pipe.ipynb

mdavezac commented 7 years ago

Done! The pipe and profiling notebooks both work in examples. You can run them using IJulia.

nlooije commented 7 years ago

Great, thx for fixing this!