Backports 1.6-rc1

Make.inc

@@ -454,10 +454,6 @@ CXX_DISABLE_ASSERTION := -DJL_NDEBUG
 DISABLE_ASSERTIONS := -DNDEBUG -DJL_NDEBUG
 endif
 
-ifeq ($(LLVM_ASSERTIONS),0)
-CXX_DISABLE_ASSERTION += -DNDEBUG
-endif
-
 # Compiler specific stuff
 
 ifeq ($(USEMSVC), 1)
@@ -1043,7 +1039,7 @@ endif # SYSTEM_LLVM
 # Windows builds need a little help finding the LLVM libraries for llvm-config
 LLVM_CONFIG_PATH_FIX :=
 ifeq ($(OS),WINNT)
-LLVM_CONFIG_PATH_FIX := PATH="$(PATH):$(build_bindir)"
+LLVM_CONFIG_PATH_FIX := PATH="$(build_bindir):$(PATH)"
 endif
 
 ifeq ($(BUILD_OS),$(OS))
@@ -1200,12 +1196,6 @@ endef
 $(foreach proj,$(BB_PROJECTS),$(eval $(call SET_BB_DEFAULT,$(proj))))
 
 
-
-# Use the Assertions build
-BINARYBUILDER_LLVM_ASSERTS ?= 0
-
-
-
 # OS specific stuff
 
 # install_name_tool

NEWS.md

@@ -75,6 +75,7 @@ Multi-threading changes
 
 * Locks now automatically inhibit finalizers from running, to avoid deadlock ([#38487]).
 * New function `Base.Threads.foreach(f, channel::Channel)` for multithreaded `Channel` consumption ([#34543]).
+* There is no longer a restriction on the number of threads ([#36778]).
 
 Build system changes
 --------------------

base/Base.jl

@@ -394,37 +394,36 @@ in_sysimage(pkgid::PkgId) = pkgid in _sysimage_modules
 # Precompiles for Revise
 # TODO: move these to contrib/generate_precompile.jl
 # The problem is they don't work there
-let m = which(+, (Int, Int))
-    while true  # defeat interpreter heuristic to force compilation
-        delete!(push!(Set{Method}(), m), m)
-        copy(Core.Compiler.retrieve_code_info(Core.Compiler.specialize_method(m, [Int, Int], Core.svec())))
-
-        empty!(Set())
-        push!(push!(Set{Union{GlobalRef,Symbol}}(), :two), GlobalRef(Base, :two))
-        (setindex!(Dict{String,Base.PkgId}(), Base.PkgId(Base), "file.jl"))["file.jl"]
-        (setindex!(Dict{Symbol,Vector{Int}}(), [1], :two))[:two]
-        (setindex!(Dict{Base.PkgId,String}(), "file.jl", Base.PkgId(Base)))[Base.PkgId(Base)]
-        (setindex!(Dict{Union{GlobalRef,Symbol}, Vector{Int}}(), [1], :two))[:two]
-        (setindex!(IdDict{Type, Union{Missing, Vector{Tuple{LineNumberNode, Expr}}}}(), missing, Int))[Int]
-        Dict{Symbol, Union{Nothing, Bool, Symbol}}(:one => false)[:one]
-        Dict(Base => [:(1+1)])[Base]
-        Dict(:one => [1])[:one]
-        Dict("abc" => Set())["abc"]
-        pushfirst!([], sum)
-        get(Base.pkgorigins, Base.PkgId(Base), nothing)
-        sort!([1,2,3])
-        unique!([1,2,3])
-        cumsum([1,2,3])
-        append!(Int[], BitSet())
-        isempty(BitSet())
-        delete!(BitSet([1,2]), 3)
-        deleteat!(Int32[1,2,3], [1,3])
-        deleteat!(Any[1,2,3], [1,3])
-        Core.svec(1, 2) == Core.svec(3, 4)
-        any(t->t[1].line > 1, [(LineNumberNode(2,:none), :(1+1))])
-
-        break   # end defeat interpreter heuristic
-    end
+for match = _methods(+, (Int, Int), -1, get_world_counter())
+    m = match.method
+    delete!(push!(Set{Method}(), m), m)
+    copy(Core.Compiler.retrieve_code_info(Core.Compiler.specialize_method(match)))
+
+    empty!(Set())
+    push!(push!(Set{Union{GlobalRef,Symbol}}(), :two), GlobalRef(Base, :two))
+    (setindex!(Dict{String,Base.PkgId}(), Base.PkgId(Base), "file.jl"))["file.jl"]
+    (setindex!(Dict{Symbol,Vector{Int}}(), [1], :two))[:two]
+    (setindex!(Dict{Base.PkgId,String}(), "file.jl", Base.PkgId(Base)))[Base.PkgId(Base)]
+    (setindex!(Dict{Union{GlobalRef,Symbol}, Vector{Int}}(), [1], :two))[:two]
+    (setindex!(IdDict{Type, Union{Missing, Vector{Tuple{LineNumberNode, Expr}}}}(), missing, Int))[Int]
+    Dict{Symbol, Union{Nothing, Bool, Symbol}}(:one => false)[:one]
+    Dict(Base => [:(1+1)])[Base]
+    Dict(:one => [1])[:one]
+    Dict("abc" => Set())["abc"]
+    pushfirst!([], sum)
+    get(Base.pkgorigins, Base.PkgId(Base), nothing)
+    sort!([1,2,3])
+    unique!([1,2,3])
+    cumsum([1,2,3])
+    append!(Int[], BitSet())
+    isempty(BitSet())
+    delete!(BitSet([1,2]), 3)
+    deleteat!(Int32[1,2,3], [1,3])
+    deleteat!(Any[1,2,3], [1,3])
+    Core.svec(1, 2) == Core.svec(3, 4)
+    any(t->t[1].line > 1, [(LineNumberNode(2,:none), :(1+1))])
+
+    break   # only actually need to do this once
 end
 
 if is_primary_base_module

base/abstractarray.jl

@@ -1193,7 +1193,7 @@ function _getindex(::IndexLinear, A::AbstractArray, I::Vararg{Int,M}) where M
 end
 _to_linear_index(A::AbstractArray, i::Integer) = i
 _to_linear_index(A::AbstractVector, i::Integer, I::Integer...) = i
-_to_linear_index(A::AbstractArray) = 1
+_to_linear_index(A::AbstractArray) = first(LinearIndices(A))
 _to_linear_index(A::AbstractArray, I::Integer...) = (@_inline_meta; _sub2ind(A, I...))
 
 ## IndexCartesian Scalar indexing: Canonical method is full dimensionality of Ints
@@ -1481,7 +1481,7 @@ vcat(V::AbstractVector{T}...) where {T} = typed_vcat(T, V...)
 # but that solution currently fails (see #27188 and #27224)
 AbstractVecOrTuple{T} = Union{AbstractVector{<:T}, Tuple{Vararg{T}}}
 
-_typed_vcat_similar(V, T, n) = similar(V[1], T, n)
+_typed_vcat_similar(V, ::Type{T}, n) where T = similar(V[1], T, n)
 _typed_vcat(::Type{T}, V::AbstractVecOrTuple{AbstractVector}) where T =
     _typed_vcat!(_typed_vcat_similar(V, T, mapreduce(length, +, V)), V)
 
@@ -1577,9 +1577,10 @@ cat_size(A::AbstractArray, d) = size(A, d)
 cat_indices(A, d) = OneTo(1)
 cat_indices(A::AbstractArray, d) = axes(A, d)
 
-cat_similar(A, T, shape) = Array{T}(undef, shape)
-cat_similar(A::AbstractArray, T, shape) = similar(A, T, shape)
+cat_similar(A, ::Type{T}, shape) where T = Array{T}(undef, shape)
+cat_similar(A::AbstractArray, ::Type{T}, shape) where T = similar(A, T, shape)
 
+# These are for backwards compatibility (even though internal)
 cat_shape(dims, shape::Tuple{Vararg{Int}}) = shape
 function cat_shape(dims, shapes::Tuple)
     out_shape = ()
@@ -1588,6 +1589,11 @@ function cat_shape(dims, shapes::Tuple)
     end
     return out_shape
 end
+# The new way to compute the shape (more inferrable than combining cat_size & cat_shape, due to Varargs + issue#36454)
+cat_size_shape(dims) = ntuple(zero, Val(length(dims)))
+@inline cat_size_shape(dims, X, tail...) = _cat_size_shape(dims, _cshp(1, dims, (), cat_size(X)), tail...)
+_cat_size_shape(dims, shape) = shape
+@inline _cat_size_shape(dims, shape, X, tail...) = _cat_size_shape(dims, _cshp(1, dims, shape, cat_size(X)), tail...)
 
 _cshp(ndim::Int, ::Tuple{}, ::Tuple{}, ::Tuple{}) = ()
 _cshp(ndim::Int, ::Tuple{}, ::Tuple{}, nshape) = nshape
@@ -1631,36 +1637,37 @@ _cat(dims, X...) = cat_t(promote_eltypeof(X...), X...; dims=dims)
 @inline cat_t(::Type{T}, X...; dims) where {T} = _cat_t(dims, T, X...)
 @inline function _cat_t(dims, ::Type{T}, X...) where {T}
     catdims = dims2cat(dims)
-    shape = cat_shape(catdims, map(cat_size, X))
+    shape = cat_size_shape(catdims, X...)
     A = cat_similar(X[1], T, shape)
     if count(!iszero, catdims)::Int > 1
         fill!(A, zero(T))
     end
     return __cat(A, shape, catdims, X...)
 end
 
-function __cat(A, shape::NTuple{M}, catdims, X...) where M
-    N = M::Int
-    offsets = zeros(Int, N)
-    inds = Vector{UnitRange{Int}}(undef, N)
-    concat = copyto!(zeros(Bool, N), catdims)
-    for x in X
-        for i = 1:N
-            if concat[i]
-                inds[i] = offsets[i] .+ cat_indices(x, i)
-                offsets[i] += cat_size(x, i)
-            else
-                inds[i] = 1:shape[i]
-            end
-        end
-        I::NTuple{N, UnitRange{Int}} = (inds...,)
-        if x isa AbstractArray
-            A[I...] = x
-        else
-            fill!(view(A, I...), x)
-        end
+# Why isn't this called `__cat!`?
+__cat(A, shape, catdims, X...) = __cat_offset!(A, shape, catdims, ntuple(zero, length(shape)), X...)
+
+function __cat_offset!(A, shape, catdims, offsets, x, X...)
+    # splitting the "work" on x from X... may reduce latency (fewer costly specializations)
+    newoffsets = __cat_offset1!(A, shape, catdims, offsets, x)
+    return __cat_offset!(A, shape, catdims, newoffsets, X...)
+end
+__cat_offset!(A, shape, catdims, offsets) = A
+
+function __cat_offset1!(A, shape, catdims, offsets, x)
+    inds = ntuple(length(offsets)) do i
+        (i <= length(catdims) && catdims[i]) ? offsets[i] .+ cat_indices(x, i) : 1:shape[i]
+    end
+    if x isa AbstractArray
+        A[inds...] = x
+    else
+        fill!(view(A, inds...), x)
+    end
+    newoffsets = ntuple(length(offsets)) do i
+        (i <= length(catdims) && catdims[i]) ? offsets[i] + cat_size(x, i) : offsets[i]
     end
-    return A
+    return newoffsets
 end
 
 """

base/binaryplatforms.jl

@@ -992,7 +992,7 @@ only available in macOS `v"10.11"` and later, or an artifact can state that it r
 a libstdc++ that is at least `v"3.4.22"`, etc...
 """
 function platforms_match(a::AbstractPlatform, b::AbstractPlatform)
-    for k in union(keys(tags(a)), keys(tags(b)))
+    for k in union(keys(tags(a)::Dict{String,String}), keys(tags(b)::Dict{String,String}))
         ak = get(tags(a), k, nothing)
         bk = get(tags(b), k, nothing)
 

base/bitarray.jl

@@ -1394,7 +1394,7 @@ function bitcount(Bc::Vector{UInt64}; init::T=0) where {T}
     return n
 end
 
-count(B::BitArray; init=0) = bitcount(B.chunks; init)
+_count(::typeof(identity), B::BitArray, ::Colon, init) = bitcount(B.chunks; init)
 
 function unsafe_bitfindnext(Bc::Vector{UInt64}, start::Int)
     chunk_start = _div64(start-1)+1

base/broadcast.jl

@@ -978,8 +978,10 @@ preprocess_args(dest, args::Tuple{}) = ()
         end
     end
     bc′ = preprocess(dest, bc)
-    @simd for I in eachindex(bc′)
-        @inbounds dest[I] = bc′[I]
+    # Performance may vary depending on whether `@inbounds` is placed outside the
+    # for loop or not. (cf. https://github.com/JuliaLang/julia/issues/38086)
+    @inbounds @simd for I in eachindex(bc′)
+        dest[I] = bc′[I]
     end
     return dest
 end