JuliaLang · KristofferC · Jun 17, 2025 · Nov 1, 2024 · Feb 18, 2025 · Feb 27, 2025
diff --git a/README.md b/README.md
@@ -43,7 +43,7 @@ and installing Julia, below.
 ## Resources
 
 - **Homepage:** <https://julialang.org>
-- **Binaries:** <https://julialang.org/downloads/>
+- **Install:** <https://julialang.org/install/>
 - **Source code:** <https://github.com/JuliaLang/julia>
 - **Documentation:** <https://docs.julialang.org>
 - **Packages:** <https://julialang.org/packages/>
@@ -64,17 +64,22 @@ helpful to start contributing to the Julia codebase.
 
 ## Binary Installation
 
-If you would rather not compile the latest Julia from source,
-platform-specific tarballs with pre-compiled binaries are also
-[available for download](https://julialang.org/downloads/). The
-downloads page also provides details on the
-[different tiers of support](https://julialang.org/downloads/#supported_platforms)
-for OS and platform combinations.
-
-If everything works correctly, you will see a Julia banner and an
-interactive prompt into which you can enter expressions for
-evaluation.  You can read about [getting
-started](https://docs.julialang.org/en/v1/manual/getting-started/) in the manual.
+The recommended way of installing Julia is to use `juliaup` which will install
+the latest stable `julia` for you and help keep it up to date. It can also let
+you install and run different Julia versions simultaneously. Instructions for
+this can be find [here](https://julialang.org/install/). If you want to manually
+download specific Julia binaries, you can find those on the [downloads
+page](https://julialang.org/downloads/). The downloads page also provides
+details on the [different tiers of
+support](https://julialang.org/downloads/#supported_platforms) for OS and
+platform combinations.
+
+If everything works correctly, you will get a `julia` program and when you run
+it in a terminal or command prompt, you will see a Julia banner and an
+interactive prompt into which you can enter expressions for evaluation. You can
+read about [getting
+started](https://docs.julialang.org/en/v1/manual/getting-started/) in the
+manual.
 
 **Note**: Although some system package managers provide Julia, such
 installations are neither maintained nor endorsed by the Julia

diff --git a/THIRDPARTY.md b/THIRDPARTY.md
@@ -6,7 +6,6 @@ for exceptions.
 - [crc32c.c](https://stackoverflow.com/questions/17645167/implementing-sse-4-2s-crc32c-in-software) (CRC-32c checksum code by Mark Adler) [[ZLib](https://opensource.org/licenses/Zlib)].
 - [LDC](https://github.com/ldc-developers/ldc/blob/master/LICENSE) (for ccall/cfunction ABI definitions) [BSD-3]. The portion of code that Julia uses from LDC is [BSD-3] licensed.
 - [LLVM](https://releases.llvm.org/3.9.0/LICENSE.TXT) (for parts of src/disasm.cpp) [UIUC]
-- [MINGW](https://sourceforge.net/p/mingw/mingw-org-wsl/ci/legacy/tree/mingwrt/mingwex/dirname.c) (for dirname implementation on Windows) [MIT]
 - [NetBSD](https://www.netbsd.org/about/redistribution.html) (for setjmp, longjmp, and strptime implementations on Windows) [BSD-3]
 - [Python](https://docs.python.org/3/license.html) (for strtod implementation on Windows) [PSF]
 - [FEMTOLISP](https://github.com/JeffBezanson/femtolisp) [BSD-3]

diff --git a/base/Base.jl b/base/Base.jl
@@ -330,10 +330,6 @@ using .Libc: getpid, gethostname, time, memcpy, memset, memmove, memcmp
 const libblas_name = "libblastrampoline" * (Sys.iswindows() ? "-5" : "")
 const liblapack_name = libblas_name
 
-# Logging
-include("logging.jl")
-using .CoreLogging
-
 # Concurrency
 include("linked_list.jl")
 include("condition.jl")
@@ -345,6 +341,10 @@ include("task.jl")
 include("threads_overloads.jl")
 include("weakkeydict.jl")
 
+# Logging
+include("logging.jl")
+using .CoreLogging
+
 include("env.jl")
 
 # functions defined in Random

diff --git a/base/abstractarray.jl b/base/abstractarray.jl
@@ -3335,12 +3335,19 @@ function ith_all(i, as)
 end
 
 function map_n!(f::F, dest::AbstractArray, As) where F
-    idxs1 = LinearIndices(As[1])
-    @boundscheck LinearIndices(dest) == idxs1 && all(x -> LinearIndices(x) == idxs1, As)
-    for i = idxs1
-        @inbounds I = ith_all(i, As)
-        val = f(I...)
-        @inbounds dest[i] = val
+    idxs = LinearIndices(dest)
+    if all(x -> LinearIndices(x) == idxs, As)
+        for i in idxs
+            @inbounds as = ith_all(i, As)
+            val = f(as...)
+            @inbounds dest[i] = val
+        end
+    else
+        for (i, Is...) in zip(eachindex(dest), map(eachindex, As)...)
+            as = ntuple(j->getindex(As[j], Is[j]), length(As))
+            val = f(as...)
+            dest[i] = val
+        end
     end
     return dest
 end
@@ -3536,7 +3543,31 @@ function _keepat!(a::AbstractVector, m::AbstractVector{Bool})
     deleteat!(a, j:lastindex(a))
 end
 
-## 1-d circshift ##
+"""
+    circshift!(a::AbstractVector, shift::Integer)
+
+Circularly shift, or rotate, the data in vector `a` by `shift` positions.
+
+# Examples
+
+```jldoctest
+julia> circshift!([1, 2, 3, 4, 5], 2)
+5-element Vector{Int64}:
+ 4
+ 5
+ 1
+ 2
+ 3
+
+julia> circshift!([1, 2, 3, 4, 5], -2)
+5-element Vector{Int64}:
+ 3
+ 4
+ 5
+ 1
+ 2
+```
+"""
 function circshift!(a::AbstractVector, shift::Integer)
     n = length(a)
     n == 0 && return

diff --git a/base/accumulate.jl b/base/accumulate.jl
@@ -5,12 +5,14 @@
 # it does double the number of operations compared to accumulate,
 # though for cheap operations like + this does not have much impact (20%)
 function _accumulate_pairwise!(op::Op, c::AbstractVector{T}, v::AbstractVector, s, i1, n)::T where {T,Op}
-    @inbounds if n < 128
-        s_ = v[i1]
-        c[i1] = op(s, s_)
+    if n < 128
+        @inbounds s_ = v[i1]
+        ci1 = op(s, s_)
+        @inbounds c[i1] = ci1
         for i = i1+1:i1+n-1
-            s_ = op(s_, v[i])
-            c[i] = op(s, s_)
+            s_ = op(s_, @inbounds(v[i]))
+            ci = op(s, s_)
+            @inbounds c[i] = ci
         end
     else
         n2 = n >> 1
@@ -26,7 +28,8 @@ function accumulate_pairwise!(op::Op, result::AbstractVector, v::AbstractVector)
     n = length(li)
     n == 0 && return result
     i1 = first(li)
-    @inbounds result[i1] = v1 = reduce_first(op,v[i1])
+    v1 = reduce_first(op, @inbounds(v[i1]))
+    @inbounds result[i1] = v1
     n == 1 && return result
     _accumulate_pairwise!(op, result, v, v1, i1+1, n-1)
     return result
@@ -379,16 +382,16 @@ function _accumulate!(op, B, A, dims::Integer, init::Union{Nothing, Some})
         # We can accumulate to a temporary variable, which allows
         # register usage and will be slightly faster
         ind1 = inds_t[1]
-        @inbounds for I in CartesianIndices(tail(inds_t))
+        for I in CartesianIndices(tail(inds_t))
             if init === nothing
-                tmp = reduce_first(op, A[first(ind1), I])
+                tmp = reduce_first(op, @inbounds(A[first(ind1), I]))
             else
-                tmp = op(something(init), A[first(ind1), I])
+                tmp = op(something(init), @inbounds(A[first(ind1), I]))
             end
-            B[first(ind1), I] = tmp
+            @inbounds B[first(ind1), I] = tmp
             for i_1 = first(ind1)+1:last(ind1)
-                tmp = op(tmp, A[i_1, I])
-                B[i_1, I] = tmp
+                tmp = op(tmp, @inbounds(A[i_1, I]))
+                @inbounds B[i_1, I] = tmp
             end
         end
     else
@@ -402,25 +405,31 @@ end
 @noinline function _accumulaten!(op, B, A, R1, ind, R2, init::Nothing)
     # Copy the initial element in each 1d vector along dimension `dim`
     ii = first(ind)
-    @inbounds for J in R2, I in R1
-        B[I, ii, J] = reduce_first(op, A[I, ii, J])
+    for J in R2, I in R1
+        tmp = reduce_first(op, @inbounds(A[I, ii, J]))
+        @inbounds B[I, ii, J] = tmp
     end
     # Accumulate
-    @inbounds for J in R2, i in first(ind)+1:last(ind), I in R1
-        B[I, i, J] = op(B[I, i-1, J], A[I, i, J])
+    for J in R2, i in first(ind)+1:last(ind), I in R1
+        @inbounds Bv, Av = B[I, i-1, J], A[I, i, J]
+        tmp = op(Bv, Av)
+        @inbounds B[I, i, J] = tmp
     end
     B
 end
 
 @noinline function _accumulaten!(op, B, A, R1, ind, R2, init::Some)
     # Copy the initial element in each 1d vector along dimension `dim`
     ii = first(ind)
-    @inbounds for J in R2, I in R1
-        B[I, ii, J] = op(something(init), A[I, ii, J])
+    for J in R2, I in R1
+        tmp = op(something(init), @inbounds(A[I, ii, J]))
+        @inbounds B[I, ii, J] = tmp
     end
     # Accumulate
-    @inbounds for J in R2, i in first(ind)+1:last(ind), I in R1
-        B[I, i, J] = op(B[I, i-1, J], A[I, i, J])
+    for J in R2, i in first(ind)+1:last(ind), I in R1
+        @inbounds Bv, Av = B[I, i-1, J], A[I, i, J]
+        tmp = op(Bv, Av)
+        @inbounds B[I, i, J] = tmp
     end
     B
 end
@@ -434,10 +443,10 @@ function _accumulate1!(op, B, v1, A::AbstractVector, dim::Integer)
     cur_val = v1
     B[i1] = cur_val
     next = iterate(inds, state)
-    @inbounds while next !== nothing
+    while next !== nothing
         (i, state) = next
-        cur_val = op(cur_val, A[i])
-        B[i] = cur_val
+        cur_val = op(cur_val, @inbounds(A[i]))
+        @inbounds B[i] = cur_val
         next = iterate(inds, state)
     end
     return B

diff --git a/base/arraymath.jl b/base/arraymath.jl
@@ -72,12 +72,12 @@ _reverse!(A::AbstractArray{<:Any,N}, ::Colon) where {N} = _reverse!(A, ntuple(id
 _reverse!(A, dim::Integer) = _reverse!(A, (Int(dim),))
 _reverse!(A, dims::NTuple{M,Integer}) where {M} = _reverse!(A, Int.(dims))
 function _reverse!(A::AbstractArray{<:Any,N}, dims::NTuple{M,Int}) where {N,M}
+    dims === () && return A # nothing to reverse
     dimrev = ntuple(k -> k in dims, Val{N}()) # boolean tuple indicating reversed dims
 
     if N < M || M != sum(dimrev)
         throw(ArgumentError("invalid dimensions $dims in reverse!"))
     end
-    M == 0 && return A # nothing to reverse
 
     # swapping loop only needs to traverse ≈half of the array
     halfsz = ntuple(k -> k == dims[1] ? size(A,k) ÷ 2 : size(A,k), Val{N}())

diff --git a/base/channels.jl b/base/channels.jl
@@ -61,7 +61,7 @@ Channel(sz=0) = Channel{Any}(sz)
 """
     Channel{T=Any}(func::Function, size=0; taskref=nothing, spawn=false, threadpool=nothing)
 
-Create a new task from `func`, bind it to a new channel of type
+Create a new task from `func`, [`bind`](@ref) it to a new channel of type
 `T` and size `size`, and schedule the task, all in a single call.
 The channel is automatically closed when the task terminates.
 

diff --git a/base/cmd.jl b/base/cmd.jl
@@ -497,7 +497,7 @@ julia> run(cm)
 Process(`echo 1`, ProcessExited(0))
 ```
 """
-macro cmd(str)
+macro cmd(str::String)
     cmd_ex = shell_parse(str, special=shell_special, filename=String(__source__.file))[1]
     return :(cmd_gen($(esc(cmd_ex))))
 end
diff --git a/base/compiler/abstractinterpretation.jl b/base/compiler/abstractinterpretation.jl
@@ -3065,7 +3065,7 @@ function typeinf_local(interp::AbstractInterpreter, frame::InferenceState)
                     rt = abstract_eval_value(interp, stmt.val, currstate, frame)
                     rt = widenreturn(rt, BestguessInfo(interp, bestguess, nargs, slottypes, currstate))
                     # narrow representation of bestguess slightly to prepare for tmerge with rt
-                    if rt isa InterConditional && bestguess isa Const
+                    if rt isa InterConditional && bestguess isa Const && bestguess.val isa Bool
                         let slot_id = rt.slot
                             old_id_type = slottypes[slot_id]
                             if bestguess.val === true && rt.elsetype !== Bottom
@@ -3074,6 +3074,15 @@ function typeinf_local(interp::AbstractInterpreter, frame::InferenceState)
                                 bestguess = InterConditional(slot_id, Bottom, old_id_type)
                             end
                         end
+                    # or narrow representation of rt slightly to prepare for tmerge with bestguess
+                    elseif bestguess isa InterConditional && rt isa Const && rt.val isa Bool
+                        slot_id = bestguess.slot
+                        old_id_type = widenconditional(slottypes[slot_id])
+                        if rt.val === true && bestguess.elsetype !== Bottom
+                            rt = InterConditional(slot_id, old_id_type, Bottom)
+                        elseif rt.val === false && bestguess.thentype !== Bottom
+                            rt = InterConditional(slot_id, Bottom, old_id_type)
+                        end
                     end
                     # copy limitations to return value
                     if !isempty(frame.pclimitations)

diff --git a/base/compiler/typeinfer.jl b/base/compiler/typeinfer.jl
@@ -406,7 +406,7 @@ function cache_result!(interp::AbstractInterpreter, result::InferenceResult)
         code_cache(interp)[linfo] = ci = CodeInstance(interp, result, inferred_result, valid_worlds)
         if track_newly_inferred[]
             m = linfo.def
-            if isa(m, Method) && m.module != Core
+            if isa(m, Method)
                 ccall(:jl_push_newly_inferred, Cvoid, (Any,), ci)
             end
         end

diff --git a/base/compiler/typelattice.jl b/base/compiler/typelattice.jl
@@ -427,6 +427,13 @@ end
         end
         a = Bool
     elseif isa(b, ConditionalT)
+        if isa(a, Const) && isa(a.val, Bool)
+           if (a.val === true && b.thentype === Any && b.elsetype === Bottom) ||
+              (a.val === false && b.elsetype === Any && b.thentype === Bottom)
+               # this Conditional contains distinctly no lattice information, and is simply an alternative representation of the Const Bool used for internal tracking purposes
+               return true
+           end
+        end
         return false
     end
     return ⊑(widenlattice(lattice), a, b)

diff --git a/base/generator.jl b/base/generator.jl
@@ -97,7 +97,7 @@ IteratorSize(::Type{Any}) = SizeUnknown()
 
 IteratorSize(::Type{<:Tuple}) = HasLength()
 IteratorSize(::Type{<:AbstractArray{<:Any,N}})  where {N} = HasShape{N}()
-IteratorSize(::Type{Generator{I,F}}) where {I,F} = IteratorSize(I)
+IteratorSize(::Type{<:Generator{I}}) where {I} = (@isdefined I) ? IteratorSize(I) : SizeUnknown()
 
 haslength(iter) = IteratorSize(iter) isa Union{HasShape, HasLength}
 

diff --git a/base/int.jl b/base/int.jl
@@ -286,8 +286,14 @@ function mod(x::T, y::T) where T<:Integer
     y == -1 && return T(0)   # avoid potential overflow in fld
     return x - fld(x, y) * y
 end
-mod(x::BitSigned, y::Unsigned) = rem(y + unsigned(rem(x, y)), y)
-mod(x::Unsigned, y::Signed) = rem(y + signed(rem(x, y)), y)
+function mod(x::BitSigned, y::Unsigned)
+    remval = rem(x, y) # correct iff  remval>=0
+    return unsigned(remval + (remval<zero(remval))*y)
+end
+function mod(x::Unsigned, y::Signed)
+    remval =  signed(rem(x, y)) #remval>0 so correct iff y>0 or remval==0
+    return remval + (!iszero(remval) && y<zero(y))*y
+end
 mod(x::T, y::T) where {T<:Unsigned} = rem(x, y)
 
 # Don't promote integers for div/rem/mod since there is no danger of overflow,
@@ -711,7 +717,7 @@ ERROR: LoadError: ArgumentError: invalid base 10 digit '.' in "123456789123.4"
 [...]
 ```
 """
-macro int128_str(s)
+macro int128_str(s::String)
     return parse(Int128, s)
 end
 
@@ -731,7 +737,7 @@ ERROR: LoadError: ArgumentError: invalid base 10 digit '-' in "-123456789123"
 [...]
 ```
 """
-macro uint128_str(s)
+macro uint128_str(s::String)
     return parse(UInt128, s)
 end
 
@@ -755,7 +761,7 @@ ERROR: ArgumentError: invalid number format _ for BigInt or BigFloat
 [...]
 ```
 """
-macro big_str(s)
+macro big_str(s::String)
     message = "invalid number format $s for BigInt or BigFloat"
     throw_error =  :(throw(ArgumentError($message)))
     if '_' in s