Skip to content

Commit

Permalink
Store the capacity minus length as the trailing byte
Browse files Browse the repository at this point in the history 
in order for an inline string to be C-compatible (last byte is always a
null byte).
  • Loading branch information
@KristofferC
KristofferC committed Jul 23, 2024
1 parent bf772fb commit c51e12c
Showing 1 changed file with 19 additions and 17 deletions.
36 changes: 19 additions & 17 deletions src/InlineStrings.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -113,7 +113,9 @@ Base.widen(::Type{InlineString63}) = InlineString127
Base.widen(::Type{InlineString127}) = InlineString255
Base.widen(::Type{InlineString255}) = String

Base.ncodeunits(x::InlineString) = Int(Base.trunc_int(UInt8, x))
trailing_byte(::Type{T}, len) where {T <: InlineString} = UInt8(sizeof(T) - len - 1)

Base.ncodeunits(x::InlineString) = Core.sizeof(x) - Int(Base.trunc_int(UInt8, x)) - 1
Base.codeunit(::InlineString) = UInt8

Base.@propagate_inbounds function Base.codeunit(x::T, i::Int) where {T <: InlineString}
Expand Down Expand Up @@ -175,23 +177,22 @@ end

# add a codeunit to end of string method
function addcodeunit(x::T, b::UInt8) where {T <: InlineString}
len = Base.trunc_int(UInt8, x)
len = ncodeunits(x) % UInt8
sz = Base.trunc_int(UInt8, sizeof(T))
shf = Base.zext_int(Int16, max(0x01, sz - len - 0x01)) << 3
x = Base.or_int(x, Base.shl_int(Base.zext_int(T, b), shf))
return Base.add_int(x, Base.zext_int(T, 0x01)), (len + 0x01) >= sz
return Base.sub_int(x, Base.zext_int(T, 0x01)), (len + 0x01) >= sz
end

for T in (:InlineString1, :InlineString3, :InlineString7, :InlineString15, :InlineString31, :InlineString63, :InlineString127, :InlineString255)
@eval $T() = Base.zext_int($T, 0x00)

@eval $T() = Base.zext_int($T, trailing_byte($T, 0))
@eval function $T(x::AbstractString)
if typeof(x) === String && sizeof($T) <= sizeof(UInt)
len = sizeof(x)
len < sizeof($T) || stringtoolong($T, len)
y = GC.@preserve x unsafe_load(convert(Ptr{$T}, pointer(x)))
sz = 8 * (sizeof($T) - len)
return Base.or_int(Base.shl_int(Base.lshr_int(_bswap(y), sz), sz), Base.zext_int($T, UInt8(len)))
return Base.or_int(Base.shl_int(Base.lshr_int(_bswap(y), sz), sz), Base.zext_int($T, trailing_byte($T, len)))
else
len = ncodeunits(x)
len < sizeof($T) || stringtoolong($T, len)
Expand Down Expand Up @@ -220,7 +221,7 @@ for T in (:InlineString1, :InlineString3, :InlineString7, :InlineString15, :Inli
else
y = GC.@preserve buf unsafe_load(convert(Ptr{$T}, pointer(buf, pos)))
sz = 8 * (sizeof($T) - len)
return Base.or_int(Base.shl_int(Base.lshr_int(_bswap(y), sz), sz), Base.zext_int($T, UInt8(len)))
return Base.or_int(Base.shl_int(Base.lshr_int(_bswap(y), sz), sz), Base.zext_int($T, trailing_byte($T, len)))
end
end

Expand Down Expand Up @@ -254,11 +255,11 @@ for T in (:InlineString1, :InlineString3, :InlineString7, :InlineString15, :Inli
len = sizeof(x)
len > (sizeof($T) - 1) && stringtoolong($T, len)
y = Base.trunc_int($T, Base.lshr_int(x, 8 * (sizeof(S) - sizeof($T))))
return Base.add_int(y, Base.zext_int($T, UInt8(len)))
return Base.add_int(y, Base.zext_int($T, trailing_byte($T, len)))
else
# promoting smaller InlineString to larger
y = Base.shl_int(Base.zext_int($T, Base.lshr_int(x, 8)), 8 * (sizeof($T) - sizeof(S) + 1))
return Base.add_int(y, Base.zext_int($T, UInt8(sizeof(x))))
return Base.add_int(y, Base.zext_int($T, trailing_byte($T, sizeof(x))))
end
end
end
Expand Down Expand Up @@ -367,12 +368,13 @@ function Base.chop(s::InlineString; head::Integer = 0, tail::Integer = 1)
return _subinlinestring(s, i, j)
end


# `i`, `j` must be `isvalid` string indexes
@inline function _subinlinestring(s::T, i::Integer, j::Integer) where {T <: InlineString}
new_n = max(0, nextind(s, j) - i) # new ncodeunits
jx = nextind(s, j) - 1 # last codeunit to keep
s = clear_n_bytes(s, sizeof(typeof(s)) - jx)
return Base.or_int(Base.shl_int(s, (i - 1) * 8), _oftype(typeof(s), new_n))
return Base.or_int(Base.shl_int(s, (i - 1) * 8), _oftype(typeof(s), trailing_byte(T, new_n)))
end

Base.getindex(s::InlineString, r::AbstractUnitRange{<:Integer}) = getindex(s, Int(first(r)):Int(last(r)))
Expand Down Expand Up @@ -420,7 +422,7 @@ end
i = min(n + 1, max(nextind(s, firstindex(s), lprefix), 1))
s = clear_n_bytes(s, 1) # clear out the length bits
s = Base.shl_int(s, (i - 1) * 8) # clear out prefix
return Base.or_int(s, _oftype(typeof(s), new_n))
return Base.or_int(s, _oftype(typeof(s), trailing_byte(typeof(s), new_n)))
end

throw_strip_argument_error() =
Expand Down Expand Up @@ -465,7 +467,7 @@ _chopsuffix(s::InlineString, suffix::AbstractString) = _chopsuffix(s, ncodeunits
n = ncodeunits(s)
new_n = n - nsuffix
s = clear_n_bytes(s, sizeof(typeof(s)) - new_n)
return Base.or_int(s, _oftype(typeof(s), new_n))
return Base.or_int(s, _oftype(typeof(s), trailing_byte(typeof(s), new_n)))
end

function Base.rstrip(f, s::InlineString)
Expand All @@ -488,16 +490,16 @@ function Base.chomp(s::InlineString)
if i < 1 || codeunit(s, i) != 0x0a
return s
elseif i < 2 || codeunit(s, i - 1) != 0x0d
return Base.or_int(clear_n_bytes(s, sizeof(typeof(s)) - i + 1), _oftype(typeof(s), len - 1))
return Base.or_int(clear_n_bytes(s, sizeof(typeof(s)) - i + 1), _oftype(typeof(s), trailing_byte(typeof(s), len - 1)))
else
return Base.or_int(clear_n_bytes(s, sizeof(typeof(s)) - i + 2), _oftype(typeof(s), len - 2))
return Base.or_int(clear_n_bytes(s, sizeof(typeof(s)) - i + 2), _oftype(typeof(s), trailing_byte(typeof(s), len - 2)))
end
end

function Base.first(s::T, n::Integer) where {T <: InlineString}
newlen = nextind(s, min(lastindex(s), nextind(s, 0, n))) - 1
i = sizeof(T) - newlen
return Base.or_int(clear_n_bytes(s, i), _oftype(typeof(s), newlen))
return Base.or_int(clear_n_bytes(s, i), _oftype(typeof(s), trailing_byte(T, newlen)))
end

function Base.last(s::T, n::Integer) where {T <: InlineString}
Expand All @@ -507,7 +509,7 @@ function Base.last(s::T, n::Integer) where {T <: InlineString}
newlen = nc - i
# clear out the length bits before shifting left
s = clear_n_bytes(s, 1)
return Base.or_int(Base.shl_int(s, (i - 1) * 8), _oftype(typeof(s), newlen))
return Base.or_int(Base.shl_int(s, (i - 1) * 8), _oftype(typeof(s), trailing_byte(T, newlen)))
end

Base.reverse(x::String1) = x
Expand Down Expand Up @@ -582,7 +584,7 @@ function _string(a::Ta, b::Tb) where {Ta <: SmallInlineStrings, Tb <: SmallInlin
# Remove length byte (lshr), grow to new size (zext), move chars forward (shl).
a2 = Base.shl_int(Base.zext_int(T, Base.lshr_int(a, 8)), 8 * (sizeof(T) - sizeof(Ta) + 1))
b2 = Base.shl_int(Base.zext_int(T, Base.lshr_int(b, 8)), 8 * (sizeof(T) - sizeof(Tb) + 1 - len_a))
lb = _oftype(T, len_a + len_b) # new length byte
lb = _oftype(T, trailing_byte(T, len_a + len_b)) # new length byte
return Base.or_int(Base.or_int(a2, b2), lb)
end

Expand Down

0 comments on commit c51e12c

Please sign in to comment.