• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

JuliaLang / julia / #37859

03 Aug 2024 02:29AM UTC coverage: 86.711% (-0.8%) from 87.527%
#37859

push

local

web-flow
Profile: close files when assembling heap snapshot (#55356)

28 of 28 new or added lines in 1 file covered. (100.0%)

1106 existing lines in 35 files now uncovered.

76980 of 88778 relevant lines covered (86.71%)

15372587.79 hits per line

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

95.85
/stdlib/Dates/src/parse.jl
1
# This file is a part of Julia. License is MIT: https://julialang.org/license
2

3
### Parsing utilities
4

5
_directives(::Type{DateFormat{S,T}}) where {S,T} = T.parameters
244✔
6

7
character_codes(df::Type{DateFormat{S,T}}) where {S,T} = character_codes(_directives(df))
166✔
8
function character_codes(directives::Core.SimpleVector)
244✔
9
    letters = sizehint!(Char[], length(directives))
244✔
10
    for (i, directive) in enumerate(directives)
488✔
11
        if directive <: DatePart
1,512✔
12
            letter = first(directive.parameters)
913✔
13
            push!(letters, letter)
913✔
14
        end
15
    end
2,780✔
16
    return letters
244✔
17
end
18

19
genvar(t::DataType) = Symbol(lowercase(string(nameof(t))))
1,938✔
20

21
"""
22
    tryparsenext_core(str::AbstractString, pos::Int, len::Int, df::DateFormat, raise=false)
23

24
Parse the string according to the directives within the `DateFormat`. Parsing will start at
25
character index `pos` and will stop when all directives are used or we have parsed up to
26
the end of the string, `len`. When a directive cannot be parsed the returned value
27
will be `nothing` if `raise` is false otherwise an exception will be thrown.
28

29
If successful, return a 3-element tuple `(values, pos, num_parsed)`:
30
* `values::Tuple`: A tuple which contains a value
31
  for each `DatePart` within the `DateFormat` in the order
32
  in which they occur. If the string ends before we finish parsing all the directives
33
  the missing values will be filled in with default values.
34
* `pos::Int`: The character index at which parsing stopped.
35
* `num_parsed::Int`: The number of values which were parsed and stored within `values`.
36
  Useful for distinguishing parsed values from default values.
37
"""
38
@generated function tryparsenext_core(str::AbstractString, pos::Int, len::Int,
237✔
39
                                      df::DateFormat, raise::Bool=false)
40
    directives = _directives(df)
78✔
41
    letters = character_codes(directives)
78✔
42

43
    tokens = Type[CONVERSION_SPECIFIERS[letter] for letter in letters]
78✔
44
    value_names = Symbol[genvar(t) for t in tokens]
78✔
45
    value_defaults = Tuple(CONVERSION_DEFAULTS[t] for t in tokens)
78✔
46

47
    # Pre-assign variables to defaults. Allows us to use `@goto done` without worrying about
48
    # unassigned variables.
49
    assign_defaults = Expr[]
78✔
50
    for (name, default) in zip(value_names, value_defaults)
156✔
51
        push!(assign_defaults, quote
295✔
52
            $name = $default
237✔
53
        end)
54
    end
512✔
55

56
    vi = 1
78✔
57
    parsers = Expr[]
78✔
58
    for i = 1:length(directives)
78✔
59
        if directives[i] <: DatePart
491✔
60
            name = value_names[vi]
295✔
61
            vi += 1
295✔
62
            push!(parsers, quote
295✔
63
                pos > len && @goto done
746✔
64
                let val = tryparsenext(directives[$i], str, pos, len, locale)
1,407✔
65
                    val === nothing && @goto error
743✔
66
                    $name, pos = val
732✔
67
                end
68
                num_parsed += 1
732✔
69
                directive_index += 1
732✔
70
            end)
71
        else
72
            push!(parsers, quote
196✔
73
                pos > len && @goto done
498✔
74
                let val = tryparsenext(directives[$i], str, pos, len, locale)
970✔
75
                    val === nothing && @goto error
478✔
76
                    delim, pos = val
463✔
77
                end
78
                directive_index += 1
463✔
79
            end)
80
        end
81
    end
491✔
82

83
    return quote
78✔
84
        directives = df.tokens
237✔
85
        locale::DateLocale = df.locale
237✔
86

87
        num_parsed = 0
237✔
88
        directive_index = 1
237✔
89

90
        $(assign_defaults...)
91
        $(parsers...)
92

93
        pos > len || @goto error
188✔
94

95
        @label done
96
        return $(Expr(:tuple, value_names...)), pos, num_parsed
205✔
97

98
        @label error
99
        if raise
32✔
100
            if directive_index > length(directives)
31✔
101
                throw(ArgumentError("Found extra characters at the end of date time string"))
6✔
102
            else
103
                d = directives[directive_index]
25✔
104
                throw(ArgumentError("Unable to parse date time. Expected directive $d at char $pos"))
25✔
105
            end
106
        end
107
        return nothing
1✔
108
    end
109
end
110

111
"""
112
    tryparsenext_internal(::Type{<:TimeType}, str, pos, len, df::DateFormat, raise=false)
113

114
Parse the string according to the directives within the `DateFormat`. The specified `TimeType`
115
type determines the type of and order of tokens returned. If the given `DateFormat` or string
116
does not provide a required token a default value will be used. When the string cannot be
117
parsed the returned value will be `nothing` if `raise` is false otherwise an exception will
118
be thrown.
119

120
If successful, returns a 2-element tuple `(values, pos)`:
121
* `values::Tuple`: A tuple which contains a value
122
  for each token as specified by the passed in type.
123
* `pos::Int`: The character index at which parsing stopped.
124
"""
125
@generated function tryparsenext_internal(::Type{T}, str::AbstractString, pos::Int, len::Int,
417✔
126
                                          df::DateFormat, raise::Bool=false) where T<:TimeType
127
    letters = character_codes(df)
322✔
128

129
    tokens = Type[CONVERSION_SPECIFIERS[letter] for letter in letters]
161✔
130
    value_names = Symbol[genvar(t) for t in tokens]
161✔
131

132
    output_tokens = CONVERSION_TRANSLATIONS[T]
161✔
133
    output_names = Symbol[genvar(t) for t in output_tokens]
1,214✔
134
    output_defaults = Tuple(CONVERSION_DEFAULTS[t] for t in output_tokens)
161✔
135

136
    # Pre-assign output variables to defaults. Ensures that all output variables are
137
    # assigned as the value tuple returned from `tryparsenext_core` may not include all
138
    # of the required variables.
139
    assign_defaults = Expr[
1,214✔
140
        quote
141
            $name = $default
193✔
142
        end
143
        for (name, default) in zip(output_names, output_defaults)
144
    ]
145

146
    # Unpacks the value tuple returned by `tryparsenext_core` into separate variables.
147
    value_tuple = Expr(:tuple, value_names...)
161✔
148

149
    return quote
161✔
150
        val = tryparsenext_core(str, pos, len, df, raise)
224✔
151
        val === nothing && return nothing
194✔
152
        values, pos, num_parsed = val
193✔
153
        $(assign_defaults...)
154
        $value_tuple = values
193✔
155
        return $(Expr(:tuple, output_names...)), pos
193✔
156
    end
157
end
158

159
@inline function tryparsenext_sign(str::AbstractString, i::Int, len::Int)
160
    i > len && return nothing
198✔
161
    c, ii = iterate(str, i)::Tuple{Char, Int}
396✔
162
    if c == '+'
198✔
163
        return 1, ii
4✔
164
    elseif c == '-'
194✔
165
        return -1, ii
6✔
166
    else
167
        return nothing
188✔
168
    end
169
end
170

171
@inline function tryparsenext_base10(str::AbstractString, i::Int, len::Int, min_width::Int=1, max_width::Int=0)
172
    i > len && return nothing
823✔
173
    min_pos = min_width <= 0 ? i : i + min_width - 1
1,453✔
174
    max_pos = max_width <= 0 ? len : min(i + max_width - 1, len)
845✔
175
    d::Int64 = 0
775✔
176
    @inbounds while i <= max_pos
775✔
177
        c, ii = iterate(str, i)::Tuple{Char, Int}
4,635✔
178
        if '0' <= c <= '9'
2,319✔
179
            d = d * 10 + (c - '0')
1,854✔
180
        else
181
            break
465✔
182
        end
183
        i = ii
1,854✔
184
    end
1,854✔
185
    if i <= min_pos
775✔
186
        return nothing
3✔
187
    else
188
        return d, i
772✔
189
    end
190
end
191

192
@inline function tryparsenext_word(str::AbstractString, i, len, locale, maxchars=0)
193
    word_start, word_end = i, 0
35✔
194
    max_pos = maxchars <= 0 ? len : min(len, nextind(str, i, maxchars-1))
41✔
195
    @inbounds while i <= max_pos
35✔
196
        c, ii = iterate(str, i)::Tuple{Char, Int}
275✔
197
        if isletter(c)
142✔
198
            word_end = i
111✔
199
        else
200
            break
28✔
201
        end
202
        i = ii
111✔
203
    end
111✔
204
    if word_end == 0
35✔
205
        return nothing
2✔
206
    else
207
        return SubString(str, word_start, word_end), i
33✔
208
    end
209
end
210

211
function Base.parse(::Type{DateTime}, s::AbstractString, df::typeof(ISODateTimeFormat))
28✔
212
    i, end_pos = firstindex(s), lastindex(s)
52✔
213
    i > end_pos && throw(ArgumentError("Cannot parse an empty string as a DateTime"))
28✔
214

215
    coefficient = 1
24✔
216
    local dy
217
    dm = dd = Int64(1)
24✔
218
    th = tm = ts = tms = Int64(0)
24✔
219

220
    # Optional sign
221
    let val = tryparsenext_sign(s, i, end_pos)
48✔
222
        if val !== nothing
24✔
223
            coefficient, i = val
4✔
224
        end
225
    end
226

227
    let val = tryparsenext_base10(s, i, end_pos, 1)
46✔
228
        val === nothing && @goto error
24✔
229
        dy, i = val
22✔
230
        i > end_pos && @goto done
22✔
231
    end
232

233
    c, i = iterate(s, i)::Tuple{Char, Int}
42✔
234
    c != '-' && @goto error
21✔
235
    i > end_pos && @goto done
21✔
236

237
    let val = tryparsenext_base10(s, i, end_pos, 1, 2)
42✔
238
        val === nothing && @goto error
21✔
239
        dm, i = val
21✔
240
        i > end_pos && @goto done
21✔
241
    end
242

243
    c, i = iterate(s, i)::Tuple{Char, Int}
42✔
244
    c != '-' && @goto error
21✔
245
    i > end_pos && @goto done
21✔
246

247
    let val = tryparsenext_base10(s, i, end_pos, 1, 2)
42✔
248
        val === nothing && @goto error
21✔
249
        dd, i = val
21✔
250
        i > end_pos && @goto done
21✔
251
    end
252

253
    c, i = iterate(s, i)::Tuple{Char, Int}
22✔
254
    c != 'T' && @goto error
11✔
255
    i > end_pos && @goto done
11✔
256

257
    let val = tryparsenext_base10(s, i, end_pos, 1, 2)
22✔
258
        val === nothing && @goto error
11✔
259
        th, i = val
11✔
260
        i > end_pos && @goto done
11✔
261
    end
262

263
    c, i = iterate(s, i)::Tuple{Char, Int}
22✔
264
    c != ':' && @goto error
11✔
265
    i > end_pos && @goto done
11✔
266

267
    let val = tryparsenext_base10(s, i, end_pos, 1, 2)
22✔
268
        val === nothing && @goto error
11✔
269
        tm, i = val
11✔
270
        i > end_pos && @goto done
11✔
271
    end
272

273
    c, i = iterate(s, i)::Tuple{Char, Int}
22✔
274
    c != ':' && @goto error
11✔
275
    i > end_pos && @goto done
11✔
276

277
    let val = tryparsenext_base10(s, i, end_pos, 1, 2)
22✔
278
        val === nothing && @goto error
11✔
279
        ts, i = val
11✔
280
        i > end_pos && @goto done
11✔
281
    end
282

UNCOV
283
    c, i = iterate(s, i)::Tuple{Char, Int}
×
UNCOV
284
    c != '.' && @goto error
×
UNCOV
285
    i > end_pos && @goto done
×
286

UNCOV
287
    let val = tryparsenext_base10(s, i, end_pos, 1, 3)
×
UNCOV
288
        val === nothing && @goto error
×
UNCOV
289
        tms, j = val
×
UNCOV
290
        tms *= 10 ^ (3 - (j - i))
×
UNCOV
291
        j > end_pos || @goto error
×
292
    end
293

294
    @label done
295
    return DateTime(dy * coefficient, dm, dd, th, tm, ts, tms)
22✔
296

297
    @label error
298
    throw(ArgumentError("Invalid DateTime string"))
2✔
299
end
300

301
function Base.parse(::Type{T}, str::AbstractString, df::DateFormat=default_format(T)) where T<:TimeType
244✔
302
    pos, len = firstindex(str), lastindex(str)
465✔
303
    pos > len && throw(ArgumentError("Cannot parse an empty string as a Date or Time"))
238✔
304
    val = tryparsenext_internal(T, str, pos, len, df, true)
412✔
305
    @assert val !== nothing
191✔
306
    values, endpos = val
191✔
307
    return T(values...)::T
191✔
308
end
309

310
function Base.tryparse(::Type{T}, str::AbstractString, df::DateFormat=default_format(T)) where T<:TimeType
13✔
311
    pos, len = firstindex(str), lastindex(str)
16✔
312
    pos > len && return nothing
9✔
313
    res = tryparsenext_internal(T, str, pos, len, df, false)
5✔
314
    res === nothing && return nothing
3✔
315
    values, endpos = res
2✔
316
    if validargs(T, values...) === nothing
2✔
317
        # TODO: validargs gets called twice, since it's called again in the T constructor
318
        return T(values...)::T
1✔
319
    end
320
    return nothing
1✔
321
end
322

323
"""
324
    parse_components(str::AbstractString, df::DateFormat) -> Array{Any}
325

326
Parse the string into its components according to the directives in the `DateFormat`.
327
Each component will be a distinct type, typically a subtype of Period. The order of the
328
components will match the order of the `DatePart` directives within the `DateFormat`. The
329
number of components may be less than the total number of `DatePart`.
330
"""
331
@generated function parse_components(str::AbstractString, df::DateFormat)
13✔
332
    letters = character_codes(df)
10✔
333
    tokens = Type[CONVERSION_SPECIFIERS[letter] for letter in letters]
5✔
334

335
    return quote
5✔
336
        pos, len = firstindex(str), lastindex(str)
26✔
337
        val = tryparsenext_core(str, pos, len, df, #=raise=#true)
13✔
338
        @assert val !== nothing
12✔
339
        values, pos, num_parsed = val
12✔
340
        types = $(Expr(:tuple, tokens...))
12✔
341
        result = Vector{Any}(undef, num_parsed)
24✔
342
        for (i, typ) in enumerate(types)
12✔
343
            i > num_parsed && break
53✔
344
            result[i] = typ(values[i])  # Constructing types takes most of the time
53✔
345
        end
53✔
346
        return result
12✔
347
    end
348
end
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2025 Coveralls, Inc