Jishnu Bhattacharya
Jishnu Bhattacharya
In https://github.com/JuliaApproximation/ApproxFunBase.jl/pull/506, I am trying to move `Statistics` and `SparseArrays` to weak deps loaded through extensions, and I test downstream packages against this PR. However, I [obtained the following warning](https://github.com/JuliaApproximation/ApproxFunBase.jl/actions/runs/5677433632/job/15385743586?pr=506)...
Given that many of the constants defined here are reciprocals of each other, we may specialize `inv` and `literal_pow` to improve floating-point accuracy. On main ```julia ulia> sqrt2π*invsqrt2π 1.0000000000000002 julia>...
Splitting the large method into two appears to reduce the TTFX for both the methods, and shifts part of the latency to the other call. First, the TTFX results (each...
This makes auto-completing `Base.OneTo` easier, as there's no ambiguity with this internal union. The leading underscore may also signal that the name is internal to `Base`.
I suspect this might have been discussed before, but I couldn't find it. ```julia julia> VERSION v"1.10.4" julia> Base.unsafe_convert(Ptr{Float64}, reinterpret(Float64, ComplexF64[1.0])) Ptr{Float64} @0x00007fc192837e20 ``` but on v"1.11.0-beta2" ```julia julia> Base.unsafe_convert(Ptr{Float64},...
Currently, `zeros` and `ones` allocate an array and call the in-place `fill!` to populate it. This PR changes it to call `fill` instead, so that immutable arrays that extend `fill`...
Seen in https://buildkite.com/julialang/julia-master/builds/37688#01904e79-f6f9-48a4-95ca-413c60c24c64 ```julia Error in testset Sockets: Error During Test at /cache/build/tester-amdci5-15/julialang/julia-master/julia-5646da32a4/share/julia/stdlib/v1.12/Sockets/test/runtests.jl:640 Got exception outside of a @test IOError: listen: address already in use (EADDRINUSE) Stacktrace: [1] uv_error @...
Typically, one can't construct `0//0`, and it's not a valid `Rational`, so this issue doesn't show up in standard use cases. However, `0//0` pops up occasionally in arrays constructed with...
For monotonic ranges, `findfirst` and `findlast` with `==(val)` as the predicate should be identical, as each value appears only once in the range. Since `findfirst` is specialized for some ranges,...
See https://github.com/JuliaApproximation/ApproxFunOrthogonalPolynomials.jl/pull/350 and https://github.com/JuliaApproximation/ApproxFunOrthogonalPolynomials.jl/pull/349, which are basically the same change, but two PRs are created as the package is listed both under `[deps]` and `[weakdeps]`. The project.toml: https://github.com/JuliaApproximation/ApproxFunOrthogonalPolynomials.jl/blob/c91c8d53ff565e95ac86949c0477fc698dcc72af/Project.toml