Andy Dienes
Andy Dienes
this creates a 1-arg method `Base.hash(:: AbstractExpressionNode{T})` morespecific than the default `hash(::Any)` fallback, and since the default seed is not `0` anymore this leads to hash inconsistencies and bugs.
basically title. might be useful to know if `using Foo` will have to do a bunch of precompilation work
https://github.com/JuliaData/CSV.jl/blob/04ec1cf3a7b0f069c573ba7c9da351026427532b/src/utils.jl#L581 this line defines `hash` on `PointerString` using `memhash`, [which will be removed in 1.13](https://github.com/JuliaLang/julia/issues/59151) note that there are some [in-progress plans to simplify the migration path](https://github.com/JuliaLang/julia/issues/59151#issuecomment-3141676024) but I just...
https://github.com/sl-solution/InMemoryDatasets.jl/blob/b02dc59da12a92f41114b8cbee94b677f15f584f/src/characters/characters.jl#L137 it's used here; ref https://github.com/JuliaLang/julia/issues/59151 note that there are some [in-progress plans to simplify the migration path](https://github.com/JuliaLang/julia/issues/59151#issuecomment-3141676024) but I just wanted to give advance warning
it appears in some test here https://github.com/JuliaDebug/JuliaInterpreter.jl/blob/b429971de7dabadf0b9f5d097d882d6edb803ca0/test/interpret.jl#L570 although not in any source. ref: https://github.com/JuliaLang/julia/issues/59151
avoids https://github.com/JuliaLang/julia/issues/33712 this is probably not a bottleneck anywhere, but may as well I guess I saw JuliaSyntax.jl has a parallel implementation of some of the code from this repo....
would be paired with https://github.com/JuliaLang/julia/pull/59025 to make `setindex!` a bit more coherent