#38162

Committed 06 Aug 2025 08:25PM UTC coverage: 25.688% (-43.6%) from 69.336%

Build # #38162

Build Type

push

local

Committed by

web-flow

Commit Message

fix runtime cglobal builtin function implementation (#59210)

This had failed to be updated for the LazyLibrary changes to codegen.

Coverage Stats

12976 of 50513 relevant lines covered (25.69%)

676965.51 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

28.47

/base/char.jl

# This file is a part of Julia. License is MIT: https://julialang.org/license

import Core: AbstractChar, Char

"""
The `AbstractChar` type is the supertype of all character implementations
in Julia. A character represents a Unicode code point, and can be converted
to an integer via the [`codepoint`](@ref) function in order to obtain the
numerical value of the code point, or constructed from the same integer.
These numerical values determine how characters are compared with `<` and `==`,
for example.  New `T <: AbstractChar` types should define a `codepoint(::T)`
method and a `T(::UInt32)` constructor, at minimum.

A given `AbstractChar` subtype may be capable of representing only a subset
of Unicode, in which case conversion from an unsupported `UInt32` value
may throw an error. Conversely, the built-in [`Char`](@ref) type represents
a *superset* of Unicode (in order to losslessly encode invalid byte streams),
in which case conversion of a non-Unicode value *to* `UInt32` throws an error.
The [`isvalid`](@ref) function can be used to check which codepoints are
representable in a given `AbstractChar` type.

Internally, an `AbstractChar` type may use a variety of encodings.  Conversion
via `codepoint(char)` will not reveal this encoding because it always returns the
Unicode value of the character. `print(io, c)` of any `c::AbstractChar`
produces an encoding determined by `io` (UTF-8 for all built-in `IO`
types), via conversion to `Char` if necessary.

`write(io, c)`, in contrast, may emit an encoding depending on
`typeof(c)`, and `read(io, typeof(c))` should read the same encoding as `write`.
New `AbstractChar` types must provide their own implementations of
`write` and `read`.
"""
AbstractChar

"""
    Char(c::Union{Number,AbstractChar})

`Char` is a 32-bit [`AbstractChar`](@ref) type that is the default representation
of characters in Julia. `Char` is the type used for character literals like `'x'`
and it is also the element type of [`String`](@ref).

In order to losslessly represent arbitrary byte streams stored in a `String`,
a `Char` value may store information that cannot be converted to a Unicode
codepoint — converting such a `Char` to `UInt32` will throw an error.
The [`isvalid(c::Char)`](@ref) function can be used to query whether `c`
represents a valid Unicode character.
"""
Char

@constprop :aggressive (::Type{T})(x::Number) where {T<:AbstractChar} = T(UInt32(x))
@constprop :aggressive AbstractChar(x::Number) = Char(x)
@constprop :aggressive (::Type{T})(x::AbstractChar) where {T<:Union{Number,AbstractChar}} = T(codepoint(x))
@constprop :aggressive (::Type{T})(x::AbstractChar) where {T<:Union{Int32,Int64}} = codepoint(x) % T
(::Type{T})(x::T) where {T<:AbstractChar} = x

"""
    ncodeunits(c::Char)::Int

Return the number of code units required to encode a character as UTF-8.
This is the number of bytes which will be printed if the character is written
to an output stream, or `ncodeunits(string(c))` but computed efficiently.

!!! compat "Julia 1.1"
    This method requires at least Julia 1.1. In Julia 1.0 consider
    using `ncodeunits(string(c))`.
"""
function ncodeunits(c::Char)
    u = reinterpret(UInt32, c)
    # We care about how many trailing bytes are all zero
    # subtract that from the total number of bytes
    n_nonzero_bytes = sizeof(UInt32) - div(trailing_zeros(u), 0x8)
    # Take care of '\0', which has an all-zero bitpattern
    n_nonzero_bytes + iszero(u)
end

"""
    codepoint(c::AbstractChar)::Integer

Return the Unicode codepoint (an unsigned integer) corresponding
to the character `c` (or throw an exception if `c` does not represent
a valid character). For `Char`, this is a `UInt32` value, but
`AbstractChar` types that represent only a subset of Unicode may
return a different-sized integer (e.g. `UInt8`).
"""
function codepoint end

@constprop :aggressive codepoint(c::Char) = UInt32(c)

struct InvalidCharError{T<:AbstractChar} <: Exception
    char::T
end
struct CodePointError{T<:Integer} <: Exception
    code::T
end
@noinline throw_invalid_char(c::AbstractChar) = throw(InvalidCharError(c))
@noinline throw_code_point_err(u::Integer) = throw(CodePointError(u))

function ismalformed(c::Char)
    u = bitcast(UInt32, c)
    l1 = leading_ones(u) << 3
    t0 = trailing_zeros(u) & 56
    (l1 == 8) | (l1 + t0 > 32) |
    (((u & 0x00c0c0c0) ⊻ 0x00808080) >> t0 != 0)
end

@inline is_overlong_enc(u::UInt32) = (u >> 24 == 0xc0) | (u >> 24 == 0xc1) | (u >> 21 == 0x0704) | (u >> 20 == 0x0f08)

function isoverlong(c::Char)
    u = bitcast(UInt32, c)
    is_overlong_enc(u)
end

# fallback: other AbstractChar types, by default, are assumed
#           not to support malformed or overlong encodings.

"""
    ismalformed(c::AbstractChar)::Bool

Return `true` if `c` represents malformed (non-Unicode) data according to the
encoding used by `c`. Defaults to `false` for non-`Char` types.

See also [`show_invalid`](@ref).
"""
ismalformed(c::AbstractChar) = false

"""
    isoverlong(c::AbstractChar)::Bool

Return `true` if `c` represents an overlong UTF-8 sequence. Defaults
to `false` for non-`Char` types.

See also [`decode_overlong`](@ref) and [`show_invalid`](@ref).
"""
isoverlong(c::AbstractChar) = false

@constprop :aggressive function UInt32(c::Char)
    # TODO: use optimized inline LLVM
    u = bitcast(UInt32, c)
    u < 0x80000000 && return u >> 24
    l1 = leading_ones(u)
    t0 = trailing_zeros(u) & 56
    (l1 == 1) | (8l1 + t0 > 32) |
    ((((u & 0x00c0c0c0) ⊻ 0x00808080) >> t0 != 0) | is_overlong_enc(u)) &&
        throw_invalid_char(c)
    u &= 0xffffffff >> l1
    u >>= t0
    ((u & 0x0000007f) >> 0) | ((u & 0x00007f00) >> 2) |
    ((u & 0x007f0000) >> 4) | ((u & 0x7f000000) >> 6)
end

"""
    decode_overlong(c::AbstractChar)::Integer

When [`isoverlong(c)`](@ref) is `true`, `decode_overlong(c)` returns
the Unicode codepoint value of `c`. `AbstractChar` implementations
that support overlong encodings should implement `Base.decode_overlong`.
"""
function decode_overlong end

@constprop :aggressive function decode_overlong(c::Char)
    u = bitcast(UInt32, c)
    l1 = leading_ones(u)
    t0 = trailing_zeros(u) & 56
    u &= 0xffffffff >> l1
    u >>= t0
    ((u & 0x0000007f) >> 0) | ((u & 0x00007f00) >> 2) |
    ((u & 0x007f0000) >> 4) | ((u & 0x7f000000) >> 6)
end

@constprop :aggressive function Char(u::UInt32)
    u < 0x80 && return bitcast(Char, u << 24)
    u < 0x00200000 || throw_code_point_err(u)
    c = ((u << 0) & 0x0000003f) | ((u << 2) & 0x00003f00) |
        ((u << 4) & 0x003f0000) | ((u << 6) & 0x3f000000)
    c = u < 0x00000800 ? (c << 16) | 0xc0800000 :
        u < 0x00010000 ? (c << 08) | 0xe0808000 :
                         (c << 00) | 0xf0808080
    bitcast(Char, c)
end

@constprop :aggressive @noinline UInt32_cold(c::Char) = UInt32(c)
@constprop :aggressive function (T::Union{Type{Int8},Type{UInt8}})(c::Char)
    i = bitcast(Int32, c)
    i ≥ 0 ? ((i >>> 24) % T) : T(UInt32_cold(c))
end

@constprop :aggressive @noinline Char_cold(b::UInt32) = Char(b)
@constprop :aggressive function Char(b::Union{Int8,UInt8})
    0 ≤ b ≤ 0x7f ? bitcast(Char, (b % UInt32) << 24) : Char_cold(UInt32(b))
end

convert(::Type{AbstractChar}, x::Number) = Char(x) # default to Char
convert(::Type{T}, x::Number) where {T<:AbstractChar} = T(x)::T
convert(::Type{T}, x::AbstractChar) where {T<:Number} = T(x)::T
convert(::Type{T}, c::AbstractChar) where {T<:AbstractChar} = T(c)::T
convert(::Type{T}, c::T) where {T<:AbstractChar} = c

rem(x::AbstractChar, ::Type{T}) where {T<:Number} = rem(codepoint(x), T)

typemax(::Type{Char}) = bitcast(Char, typemax(UInt32))
typemin(::Type{Char}) = bitcast(Char, typemin(UInt32))

size(c::AbstractChar) = ()
size(c::AbstractChar, d::Integer) = d < 1 ? throw(BoundsError()) : 1
ndims(c::AbstractChar) = 0
ndims(::Type{<:AbstractChar}) = 0
length(c::AbstractChar) = 1
IteratorSize(::Type{<:AbstractChar}) = HasShape{0}()
firstindex(c::AbstractChar) = 1
lastindex(c::AbstractChar) = 1
getindex(c::AbstractChar) = c
getindex(c::AbstractChar, i::Integer) = i == 1 ? c : throw(BoundsError())
getindex(c::AbstractChar, I::Integer...) = all(x -> x == 1, I) ? c : throw(BoundsError())
first(c::AbstractChar) = c
last(c::AbstractChar) = c
eltype(::Type{T}) where {T<:AbstractChar} = T

iterate(c::AbstractChar, done=false) = done ? nothing : (c, true)
isempty(c::AbstractChar) = false
in(x::AbstractChar, y::AbstractChar) = x == y

==(x::Char, y::Char) = bitcast(UInt32, x) == bitcast(UInt32, y)
isless(x::Char, y::Char) = bitcast(UInt32, x) < bitcast(UInt32, y)
hash(x::Char, h::UInt) =
    hash_finalizer(((bitcast(UInt32, x) + UInt64(0xd4d64234)) << 32) ⊻ UInt64(h)) % UInt

# fallbacks:
isless(x::AbstractChar, y::AbstractChar) = isless(Char(x), Char(y))
==(x::AbstractChar, y::AbstractChar) = Char(x) == Char(y)
hash(x::AbstractChar, h::UInt) = hash(Char(x), h)
widen(::Type{T}) where {T<:AbstractChar} = T

@inline -(x::AbstractChar, y::AbstractChar) = Int(x) - Int(y)
@inline function -(x::T, y::Integer) where {T<:AbstractChar}
    if x isa Char
        u = Int32((bitcast(UInt32, x) >> 24) % Int8)
        if u >= 0 # inline the runtime fast path
            z = u - y
            return 0 <= z < 0x80 ? bitcast(Char, (z % UInt32) << 24) : Char(UInt32(z))
        end
    end
    return T(Int32(x) - Int32(y))
end
@inline function +(x::T, y::Integer) where {T<:AbstractChar}
    if x isa Char
        u = Int32((bitcast(UInt32, x) >> 24) % Int8)
        if u >= 0 # inline the runtime fast path
            z = u + y
            return 0 <= z < 0x80 ? bitcast(Char, (z % UInt32) << 24) : Char(UInt32(z))
        end
    end
    return T(Int32(x) + Int32(y))
end
@inline +(x::Integer, y::AbstractChar) = y + x

# `print` should output UTF-8 by default for all AbstractChar types.
# (Packages may implement other IO subtypes to specify different encodings.)
# In contrast, `write(io, c)` outputs a `c` in an encoding determined by typeof(c).
print(io::IO, c::Char) = (write(io, c); nothing)
print(io::IO, c::AbstractChar) = print(io, Char(c)) # fallback: convert to output UTF-8

const hex_chars = UInt8['0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
                        'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i',
                        'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r',
                        's', 't', 'u', 'v', 'w', 'x', 'y', 'z']

function show_invalid(io::IO, c::Char)
    write(io, 0x27)
    u = bitcast(UInt32, c)
    while true
        a = hex_chars[((u >> 28) & 0xf) + 1]
        b = hex_chars[((u >> 24) & 0xf) + 1]
        write(io, 0x5c, UInt8('x'), a, b)
        (u <<= 8) == 0 && break
    end
    write(io, 0x27)
end

"""
    show_invalid(io::IO, c::AbstractChar)

Called by `show(io, c)` when [`isoverlong(c)`](@ref) or
[`ismalformed(c)`](@ref) return `true`.   Subclasses
of `AbstractChar` should define `Base.show_invalid` methods
if they support storing invalid character data.
"""
show_invalid

# show c to io, assuming UTF-8 encoded output
function show(io::IO, c::AbstractChar)
    if c <= '\\'
        b = c == '\0' ? 0x30 :
            c == '\a' ? 0x61 :
            c == '\b' ? 0x62 :
            c == '\t' ? 0x74 :
            c == '\n' ? 0x6e :
            c == '\v' ? 0x76 :
            c == '\f' ? 0x66 :
            c == '\r' ? 0x72 :
            c == '\e' ? 0x65 :
            c == '\'' ? 0x27 :
            c == '\\' ? 0x5c : 0xff
        if b != 0xff
            write(io, 0x27, 0x5c, b, 0x27)
            return
        end
    end
    if isoverlong(c) || ismalformed(c)
        show_invalid(io, c)
    elseif isprint(c)
        write(io, 0x27)
        print(io, c) # use print, not write, to use UTF-8 for any AbstractChar
        write(io, 0x27)
    else # unprintable, well-formed, non-overlong Unicode
        u = codepoint(c)
        write(io, 0x27, 0x5c, u <= 0x7f ? 0x78 : u <= 0xffff ? 0x75 : 0x55)
        d = max(2, 8 - (leading_zeros(u) >> 2))
        while 0 < d
            write(io, hex_chars[((u >> ((d -= 1) << 2)) & 0xf) + 1])
        end
        write(io, 0x27)
    end
    return
end

function show(io::IO, ::MIME"text/plain", c::T) where {T<:AbstractChar}
    show(io, c)
    get(io, :compact, false)::Bool && return
    if !ismalformed(c)
        print(io, ": ")
        if isoverlong(c)
            print(io, "[overlong] ")
            u = decode_overlong(c)
            c = T(u)
        else
            u = codepoint(c)
        end
        h = uppercase(string(u, base = 16, pad = 4))
        print(io, (isascii(c) ? "ASCII/" : ""), "Unicode U+", h)
    else
        print(io, ": Malformed UTF-8")
    end
    abr = Unicode.category_abbrev(c)
    str = Unicode.category_string(c)
    print(io, " (category ", abr, ": ", str, ")")
end

1	# This file is a part of Julia. License is MIT: https://julialang.org/license
2
3	import Core: AbstractChar, Char
4
5	"""
6	The `AbstractChar` type is the supertype of all character implementations
7	in Julia. A character represents a Unicode code point, and can be converted
8	to an integer via the [`codepoint`](@ref) function in order to obtain the
9	numerical value of the code point, or constructed from the same integer.
10	These numerical values determine how characters are compared with `<` and `==`,
11	for example. New `T <: AbstractChar` types should define a `codepoint(::T)`
12	method and a `T(::UInt32)` constructor, at minimum.
13
14	A given `AbstractChar` subtype may be capable of representing only a subset
15	of Unicode, in which case conversion from an unsupported `UInt32` value
16	may throw an error. Conversely, the built-in [`Char`](@ref) type represents
17	a superset of Unicode (in order to losslessly encode invalid byte streams),
18	in which case conversion of a non-Unicode value to `UInt32` throws an error.
19	The [`isvalid`](@ref) function can be used to check which codepoints are
20	representable in a given `AbstractChar` type.
21
22	Internally, an `AbstractChar` type may use a variety of encodings. Conversion
23	via `codepoint(char)` will not reveal this encoding because it always returns the
24	Unicode value of the character. `print(io, c)` of any `c::AbstractChar`
25	produces an encoding determined by `io` (UTF-8 for all built-in `IO`
26	types), via conversion to `Char` if necessary.
27
28	`write(io, c)`, in contrast, may emit an encoding depending on
29	`typeof(c)`, and `read(io, typeof(c))` should read the same encoding as `write`.
30	New `AbstractChar` types must provide their own implementations of
31	`write` and `read`.
32	"""
33	AbstractChar
34
35	"""
36	Char(c::Union{Number,AbstractChar})
37
38	`Char` is a 32-bit [`AbstractChar`](@ref) type that is the default representation
39	of characters in Julia. `Char` is the type used for character literals like `'x'`
40	and it is also the element type of [`String`](@ref).
41
42	In order to losslessly represent arbitrary byte streams stored in a `String`,
43	a `Char` value may store information that cannot be converted to a Unicode
44	codepoint — converting such a `Char` to `UInt32` will throw an error.
45	The [`isvalid(c::Char)`](@ref) function can be used to query whether `c`
46	represents a valid Unicode character.
47	"""
48	Char
49
50	@constprop :aggressive (::Type{T})(x::Number) where {T<:AbstractChar} = T(UInt32(x))	2✔
51	@constprop :aggressive AbstractChar(x::Number) = Char(x)	×
52	@constprop :aggressive (::Type{T})(x::AbstractChar) where {T<:Union{Number,AbstractChar}} = T(codepoint(x))	×
53	@constprop :aggressive (::Type{T})(x::AbstractChar) where {T<:Union{Int32,Int64}} = codepoint(x) % T	15,946✔
54	(::Type{T})(x::T) where {T<:AbstractChar} = x	28✔
55
56	"""
57	ncodeunits(c::Char)::Int
58
59	Return the number of code units required to encode a character as UTF-8.
60	This is the number of bytes which will be printed if the character is written
61	to an output stream, or `ncodeunits(string(c))` but computed efficiently.
62
63	!!! compat "Julia 1.1"
64	This method requires at least Julia 1.1. In Julia 1.0 consider
65	using `ncodeunits(string(c))`.
66	"""
67	function ncodeunits(c::Char)	6✔
68	u = reinterpret(UInt32, c)	22,157✔
69	# We care about how many trailing bytes are all zero
70	# subtract that from the total number of bytes
71	n_nonzero_bytes = sizeof(UInt32) - div(trailing_zeros(u), 0x8)	22,157✔
72	# Take care of '\0', which has an all-zero bitpattern
73	n_nonzero_bytes + iszero(u)	22,157✔
74	end
75
76	"""
77	codepoint(c::AbstractChar)::Integer
78
79	Return the Unicode codepoint (an unsigned integer) corresponding
80	to the character `c` (or throw an exception if `c` does not represent
81	a valid character). For `Char`, this is a `UInt32` value, but
82	`AbstractChar` types that represent only a subset of Unicode may
83	return a different-sized integer (e.g. `UInt8`).
84	"""
85	function codepoint end
86
87	@constprop :aggressive codepoint(c::Char) = UInt32(c)	47,724✔
88
89	struct InvalidCharError{T<:AbstractChar} <: Exception
90	char::T	×
91	end
92	struct CodePointError{T<:Integer} <: Exception
93	code::T	×
94	end
95	@noinline throw_invalid_char(c::AbstractChar) = throw(InvalidCharError(c))	×
96	@noinline throw_code_point_err(u::Integer) = throw(CodePointError(u))	×
97
98	function ismalformed(c::Char)
99	u = bitcast(UInt32, c)	202,892✔
100	l1 = leading_ones(u) << 3	202,892✔
101	t0 = trailing_zeros(u) & 56	202,892✔
102	(l1 == 8) \| (l1 + t0 > 32) \|	202,892✔
103	(((u & 0x00c0c0c0) ⊻ 0x00808080) >> t0 != 0)
104	end
105
106	@inline is_overlong_enc(u::UInt32) = (u >> 24 == 0xc0) \| (u >> 24 == 0xc1) \| (u >> 21 == 0x0704) \| (u >> 20 == 0x0f08)	44,067✔
107
108	function isoverlong(c::Char)
109	u = bitcast(UInt32, c)	44,053✔
110	is_overlong_enc(u)	44,053✔
111	end
112
113	# fallback: other AbstractChar types, by default, are assumed
114	# not to support malformed or overlong encodings.
115
116	"""
117	ismalformed(c::AbstractChar)::Bool
118
119	Return `true` if `c` represents malformed (non-Unicode) data according to the
120	encoding used by `c`. Defaults to `false` for non-`Char` types.
121
122	See also [`show_invalid`](@ref).
123	"""
124	ismalformed(c::AbstractChar) = false	×
125
126	"""
127	isoverlong(c::AbstractChar)::Bool
128
129	Return `true` if `c` represents an overlong UTF-8 sequence. Defaults
130	to `false` for non-`Char` types.
131
132	See also [`decode_overlong`](@ref) and [`show_invalid`](@ref).
133	"""
134	isoverlong(c::AbstractChar) = false	×
135
136	@constprop :aggressive function UInt32(c::Char)	2✔
137	# TODO: use optimized inline LLVM
138	u = bitcast(UInt32, c)	208,277✔
139	u < 0x80000000 && return u >> 24	210,787✔
140	l1 = leading_ones(u)	14✔
141	t0 = trailing_zeros(u) & 56	14✔
142	(l1 == 1) \| (8l1 + t0 > 32) \|	14✔
143	((((u & 0x00c0c0c0) ⊻ 0x00808080) >> t0 != 0) \| is_overlong_enc(u)) &&
144	throw_invalid_char(c)
145	u &= 0xffffffff >> l1	14✔
146	u >>= t0	14✔
147	((u & 0x0000007f) >> 0) \| ((u & 0x00007f00) >> 2) \|	14✔
148	((u & 0x007f0000) >> 4) \| ((u & 0x7f000000) >> 6)
149	end
150
151	"""
152	decode_overlong(c::AbstractChar)::Integer
153
154	When [`isoverlong(c)`](@ref) is `true`, `decode_overlong(c)` returns
155	the Unicode codepoint value of `c`. `AbstractChar` implementations
156	that support overlong encodings should implement `Base.decode_overlong`.
157	"""
158	function decode_overlong end
159
160	@constprop :aggressive function decode_overlong(c::Char)	×
161	u = bitcast(UInt32, c)	×
162	l1 = leading_ones(u)	×
163	t0 = trailing_zeros(u) & 56	×
164	u &= 0xffffffff >> l1	×
165	u >>= t0	×
166	((u & 0x0000007f) >> 0) \| ((u & 0x00007f00) >> 2) \|	×
167	((u & 0x007f0000) >> 4) \| ((u & 0x7f000000) >> 6)
168	end
169
170	@constprop :aggressive function Char(u::UInt32)
171	u < 0x80 && return bitcast(Char, u << 24)	2✔
172	u < 0x00200000 \|\| throw_code_point_err(u)	×
173	c = ((u << 0) & 0x0000003f) \| ((u << 2) & 0x00003f00) \|	×
174	((u << 4) & 0x003f0000) \| ((u << 6) & 0x3f000000)
175	c = u < 0x00000800 ? (c << 16) \| 0xc0800000 :	×
176	u < 0x00010000 ? (c << 08) \| 0xe0808000 :
177	(c << 00) \| 0xf0808080
178	bitcast(Char, c)	×
179	end
180
181	@constprop :aggressive @noinline UInt32_cold(c::Char) = UInt32(c)	×
182	@constprop :aggressive function (T::Union{Type{Int8},Type{UInt8}})(c::Char)	436✔
183	i = bitcast(Int32, c)	742✔
184	i ≥ 0 ? ((i >>> 24) % T) : T(UInt32_cold(c))	742✔
185	end
186
187	@constprop :aggressive @noinline Char_cold(b::UInt32) = Char(b)	×
188	@constprop :aggressive function Char(b::Union{Int8,UInt8})
189	0 ≤ b ≤ 0x7f ? bitcast(Char, (b % UInt32) << 24) : Char_cold(UInt32(b))	137,150✔
190	end
191
192	convert(::Type{AbstractChar}, x::Number) = Char(x) # default to Char	×
193	convert(::Type{T}, x::Number) where {T<:AbstractChar} = T(x)::T	×
194	convert(::Type{T}, x::AbstractChar) where {T<:Number} = T(x)::T	3,867✔
195	convert(::Type{T}, c::AbstractChar) where {T<:AbstractChar} = T(c)::T	×
196	convert(::Type{T}, c::T) where {T<:AbstractChar} = c	×
197
198	rem(x::AbstractChar, ::Type{T}) where {T<:Number} = rem(codepoint(x), T)	15,889✔
199
200	typemax(::Type{Char}) = bitcast(Char, typemax(UInt32))	×
201	typemin(::Type{Char}) = bitcast(Char, typemin(UInt32))	×
202
203	size(c::AbstractChar) = ()	×
204	size(c::AbstractChar, d::Integer) = d < 1 ? throw(BoundsError()) : 1	×
205	ndims(c::AbstractChar) = 0	×
206	ndims(::Type{<:AbstractChar}) = 0	×
207	length(c::AbstractChar) = 1	×
208	IteratorSize(::Type{<:AbstractChar}) = HasShape{0}()	×
209	firstindex(c::AbstractChar) = 1	×
210	lastindex(c::AbstractChar) = 1	×
211	getindex(c::AbstractChar) = c	×
212	getindex(c::AbstractChar, i::Integer) = i == 1 ? c : throw(BoundsError())	×
213	getindex(c::AbstractChar, I::Integer...) = all(x -> x == 1, I) ? c : throw(BoundsError())	×
214	first(c::AbstractChar) = c	×
215	last(c::AbstractChar) = c	×
216	eltype(::Type{T}) where {T<:AbstractChar} = T	×
217
218	iterate(c::AbstractChar, done=false) = done ? nothing : (c, true)	×
219	isempty(c::AbstractChar) = false	×
220	in(x::AbstractChar, y::AbstractChar) = x == y	20,420✔
221
222	==(x::Char, y::Char) = bitcast(UInt32, x) == bitcast(UInt32, y)	10,455,914✔
223	isless(x::Char, y::Char) = bitcast(UInt32, x) < bitcast(UInt32, y)	2,508,955✔
224	hash(x::Char, h::UInt) =	3✔
225	hash_finalizer(((bitcast(UInt32, x) + UInt64(0xd4d64234)) << 32) ⊻ UInt64(h)) % UInt
226
227	# fallbacks:
228	isless(x::AbstractChar, y::AbstractChar) = isless(Char(x), Char(y))	×
229	==(x::AbstractChar, y::AbstractChar) = Char(x) == Char(y)	×
230	hash(x::AbstractChar, h::UInt) = hash(Char(x), h)	×
231	widen(::Type{T}) where {T<:AbstractChar} = T	×
232
233	@inline -(x::AbstractChar, y::AbstractChar) = Int(x) - Int(y)	15,890✔
234	@inline function -(x::T, y::Integer) where {T<:AbstractChar}
235	if x isa Char	×
236	u = Int32((bitcast(UInt32, x) >> 24) % Int8)	6✔
237	if u >= 0 # inline the runtime fast path	6✔
238	z = u - y	6✔
239	return 0 <= z < 0x80 ? bitcast(Char, (z % UInt32) << 24) : Char(UInt32(z))	6✔
240	end
241	end
242	return T(Int32(x) - Int32(y))	×
243	end
244	@inline function +(x::T, y::Integer) where {T<:AbstractChar}
245	if x isa Char	×
246	u = Int32((bitcast(UInt32, x) >> 24) % Int8)	×
247	if u >= 0 # inline the runtime fast path	×
248	z = u + y	×
249	return 0 <= z < 0x80 ? bitcast(Char, (z % UInt32) << 24) : Char(UInt32(z))	×
250	end
251	end
252	return T(Int32(x) + Int32(y))	×
253	end
254	@inline +(x::Integer, y::AbstractChar) = y + x	×
255
256	# `print` should output UTF-8 by default for all AbstractChar types.
257	# (Packages may implement other IO subtypes to specify different encodings.)
258	# In contrast, `write(io, c)` outputs a `c` in an encoding determined by typeof(c).
259	print(io::IO, c::Char) = (write(io, c); nothing)	344,599✔
260	print(io::IO, c::AbstractChar) = print(io, Char(c)) # fallback: convert to output UTF-8	×
261
262	const hex_chars = UInt8['0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
263	'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i',
264	'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r',
265	's', 't', 'u', 'v', 'w', 'x', 'y', 'z']
266
267	function show_invalid(io::IO, c::Char)	×
268	write(io, 0x27)	×
269	u = bitcast(UInt32, c)	×
270	while true	×
271	a = hex_chars[((u >> 28) & 0xf) + 1]	×
272	b = hex_chars[((u >> 24) & 0xf) + 1]	×
273	write(io, 0x5c, UInt8('x'), a, b)	×
274	(u <<= 8) == 0 && break	×
275	end	×
276	write(io, 0x27)	×
277	end
278
279	"""
280	show_invalid(io::IO, c::AbstractChar)
281
282	Called by `show(io, c)` when [`isoverlong(c)`](@ref) or
283	[`ismalformed(c)`](@ref) return `true`. Subclasses
284	of `AbstractChar` should define `Base.show_invalid` methods
285	if they support storing invalid character data.
286	"""
287	show_invalid
288
289	# show c to io, assuming UTF-8 encoded output
290	function show(io::IO, c::AbstractChar)	×
291	if c <= '\\'	×
292	b = c == '\0' ? 0x30 :	×
293	c == '\a' ? 0x61 :
294	c == '\b' ? 0x62 :
295	c == '\t' ? 0x74 :
296	c == '\n' ? 0x6e :
297	c == '\v' ? 0x76 :
298	c == '\f' ? 0x66 :
299	c == '\r' ? 0x72 :
300	c == '\e' ? 0x65 :
301	c == '\'' ? 0x27 :
302	c == '\\' ? 0x5c : 0xff
303	if b != 0xff	×
304	write(io, 0x27, 0x5c, b, 0x27)	×
305	return	×
306	end
307	end
308	if isoverlong(c) \|\| ismalformed(c)	×
309	show_invalid(io, c)	×
310	elseif isprint(c)	×
311	write(io, 0x27)	×
312	print(io, c) # use print, not write, to use UTF-8 for any AbstractChar	×
313	write(io, 0x27)	×
314	else # unprintable, well-formed, non-overlong Unicode
315	u = codepoint(c)	×
316	write(io, 0x27, 0x5c, u <= 0x7f ? 0x78 : u <= 0xffff ? 0x75 : 0x55)	×
317	d = max(2, 8 - (leading_zeros(u) >> 2))	×
318	while 0 < d	×
319	write(io, hex_chars[((u >> ((d -= 1) << 2)) & 0xf) + 1])	×
320	end	×
321	write(io, 0x27)	×
322	end
323	return	×
324	end
325
326	function show(io::IO, ::MIME"text/plain", c::T) where {T<:AbstractChar}	×
327	show(io, c)	×
328	get(io, :compact, false)::Bool && return	×
329	if !ismalformed(c)	×
330	print(io, ": ")	×
331	if isoverlong(c)	×
332	print(io, "[overlong] ")	×
333	u = decode_overlong(c)	×
334	c = T(u)	×
335	else
336	u = codepoint(c)	×
337	end
338	h = uppercase(string(u, base = 16, pad = 4))	×
339	print(io, (isascii(c) ? "ASCII/" : ""), "Unicode U+", h)	×
340	else
341	print(io, ": Malformed UTF-8")	×
342	end
343	abr = Unicode.category_abbrev(c)	×
344	str = Unicode.category_string(c)	×
345	print(io, " (category ", abr, ": ", str, ")")	×
346	end

JuliaLang / julia / #38162

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous