bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition can yield, meaning multiple concurrent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition can yield, meaning multiple concurrent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition can yield, meaning multiple concurrent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition yield, meaning multiple concur- rent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition yield, meaning multiple concur- rent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition yield, meaning multiple concur- rent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition yield, meaning multiple concur- rent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition yield, meaning multiple concur- rent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition can yield, meaning multiple concurrent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points

bring x and y to a common type: f(x::T, y::T) where {T} = ... f(x, y) = f(promote(x, y)...) One risk with this design is the possibility that if there is no suitable promotion method converting x and For a more general way to make all such calls work sensibly, see Conversion and Promotion. At the risk of spoiling the suspense, we can reveal here that all it takes is the following outer method definition can yield, meaning multiple concurrent tasks may use the same buffer on a given thread, introducing risk of data races. Further, when more than one thread is available tasks may change thread at yield points