Module: OkJson

Extended by:
OkJson
Included in:
OkJson
Defined in:
lib/xplenty/kensa/okjson.rb,
lib/xplenty/kensa/okjson-support.rb

Overview

Some parts adapted from golang.org/src/pkg/json/decode.go and golang.org/src/pkg/utf8/utf8.go

Defined Under Namespace

Classes: Error

Constant Summary collapse

Utagx =

1000 0000

0x80
Utag2 =

1100 0000

0xc0
Utag3 =

1110 0000

0xe0
Utag4 =

1111 0000

0xf0
Utag5 =

1111 1000

0xF8
Umaskx =

0011 1111

0x3f
Umask2 =

0001 1111

0x1f
Umask3 =

0000 1111

0x0f
Umask4 =

0000 0111

0x07
Uchar1max = 
(1<<7) - 1
Uchar2max = 
(1<<11) - 1
Uchar3max = 
(1<<16) - 1
Ucharerr =

unicode “replacement char”

0xFFFD
Usurrself = 
0x10000
Usurr1 = 
0xd800
Usurr2 = 
0xdc00
Usurr3 = 
0xe000
Umax = 
0x10ffff
Spc = 


Unesc =



{?b=>?\b, ?f=>?\f, ?n=>?\n, ?r=>?\r, ?t=>?\t}


Hex =



'0123456789abcdef'





Instance Method Summary
collapse






Instance Method Details





#abbrev(s) ⇒ Object










250
251
252
253
254
255
256




# File 'lib/xplenty/kensa/okjson.rb', line 250

def abbrev(s)
  t = s[0,10]
  p = t['`']
  t = t[0,p] if p
  t = t + '...' if t.length < s.length
  '`' + t + '`'
end












#arrenc(a) ⇒ Object










441
442
443




# File 'lib/xplenty/kensa/okjson.rb', line 441

def arrenc(a)
  '[' + a.map{|x| valenc(x)}.join(',') + ']'
end












#arrparse(ts) ⇒ Object







Parses an “array” in the sense of RFC 4627. Returns the parsed value and any trailing tokens.















134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159




# File 'lib/xplenty/kensa/okjson.rb', line 134

def arrparse(ts)
  ts = eat('[', ts)
  arr = []

  if ts[0][0] == ']'
    return arr, ts[1..-1]
  end

  v, ts = valparse(ts)
  arr << v

  if ts[0][0] == ']'
    return arr, ts[1..-1]
  end

  loop do
    ts = eat(',', ts)

    v, ts = valparse(ts)
    arr << v

    if ts[0][0] == ']'
      return arr, ts[1..-1]
    end
  end
end












#decode(s) ⇒ Object







Decodes a json document in string s and returns the corresponding ruby value. String s must be valid UTF-8. If you have a string in some other encoding, convert it first.


String values in the resulting structure will be UTF-8.















42
43
44
45
46
47
48
49




# File 'lib/xplenty/kensa/okjson.rb', line 42

def decode(s)
  ts = lex(s)
  v, ts = textparse(ts)
  if ts.length > 0
    raise Error, 'trailing garbage'
  end
  v
end












#eat(typ, ts) ⇒ Object










162
163
164
165
166
167




# File 'lib/xplenty/kensa/okjson.rb', line 162

def eat(typ, ts)
  if ts[0][0] != typ
    raise Error, "expected #{typ} (got #{ts[0].inspect})"
  end
  ts[1..-1]
end












#encode(x) ⇒ Object







Encodes x into a json text. It may contain only Array, Hash, String, Numeric, true, false, nil. (Note, this list excludes Symbol.) X itself must be an Array or a Hash. No other value can be encoded, and an error will be raised if x contains any other value, such as Nan, Infinity, Symbol, and Proc, or if a Hash key is not a String. Strings contained in x must be valid UTF-8.















411
412
413
414
415
416
417
418




# File 'lib/xplenty/kensa/okjson.rb', line 411

def encode(x)
  case x
  when Hash    then objenc(x)
  when Array   then arrenc(x)
  else
    raise Error, 'root value must be an Array or a Hash'
  end
end












#encode_without_stringify ⇒ Object







Encodes x into a json text. It may contain only Array, Hash, String, Numeric, true, false, nil. (Note, this list excludes Symbol.) X itself must be an Array or a Hash. No other value can be encoded, and an error will be raised if x contains any other value, such as Nan, Infinity, Symbol, and Proc, or if a Hash key is not a String. Strings contained in x must be valid UTF-8.















19
20
21
22
23
24
25
26




# File 'lib/xplenty/kensa/okjson-support.rb', line 19

def encode(x)
  case x
  when Hash    then objenc(x)
  when Array   then arrenc(x)
  else
    raise Error, 'root value must be an Array or a Hash'
  end
end












#falsetok(s) ⇒ Object










224




# File 'lib/xplenty/kensa/okjson.rb', line 224

def falsetok(s); s[0,5] == 'false' && [:val, 'false', false] end












#hexdec4(s) ⇒ Object










361
362
363
364
365
366




# File 'lib/xplenty/kensa/okjson.rb', line 361

def hexdec4(s)
  if s.length != 4
    raise Error, 'short'
  end
  (nibble(s[0])<<12) | (nibble(s[1])<<8) | (nibble(s[2])<<4) | nibble(s[3])
end












#hexenc4(t, u) ⇒ Object










507
508
509
510
511
512




# File 'lib/xplenty/kensa/okjson.rb', line 507

def hexenc4(t, u)
  t.putc(Hex[(u>>12)&0xf])
  t.putc(Hex[(u>>8)&0xf])
  t.putc(Hex[(u>>4)&0xf])
  t.putc(Hex[u&0xf])
end












#keyenc(k) ⇒ Object










446
447
448
449
450
451
452




# File 'lib/xplenty/kensa/okjson.rb', line 446

def keyenc(k)
  case k
  when String then strenc(k)
  else
    raise Error, "Hash key is not a string: #{k.inspect}"
  end
end












#lex(s) ⇒ Object







Scans s and returns a list of json tokens, excluding white space (as defined in RFC 4627).















172
173
174
175
176
177
178
179
180
181
182
183
184
185




# File 'lib/xplenty/kensa/okjson.rb', line 172

def lex(s)
  ts = []
  while s.length > 0
    typ, lexeme, val = tok(s)
    if typ == nil
      raise Error, "invalid character at #{s[0,10].inspect}"
    end
    if typ != :space
      ts << [typ, lexeme, val]
    end
    s = s[lexeme.length..-1]
  end
  ts
end












#nibble(c) ⇒ Object










391
392
393
394
395
396
397
398
399




# File 'lib/xplenty/kensa/okjson.rb', line 391

def nibble(c)
  case true
  when ?0 <= c && c <= ?9 then c.ord - ?0.ord
  when ?a <= c && c <= ?z then c.ord - ?a.ord + 10
  when ?A <= c && c <= ?Z then c.ord - ?A.ord + 10
  else
    raise Error, "invalid hex code #{c}"
  end
end












#nulltok(s) ⇒ Object










222




# File 'lib/xplenty/kensa/okjson.rb', line 222

def nulltok(s);  s[0,4] == 'null'  && [:val, 'null',  nil]   end












#numenc(x) ⇒ Object










515
516
517
518
519
520




# File 'lib/xplenty/kensa/okjson.rb', line 515

def numenc(x)
  if ((x.nan? || x.infinite?) rescue false)
    raise Error, "Numeric cannot be represented: #{x}"
  end
  "#{x}"
end












#numtok(s) ⇒ Object










227
228
229
230
231
232
233
234
235
236
237
238




# File 'lib/xplenty/kensa/okjson.rb', line 227

def numtok(s)
  m = /-?([1-9][0-9]+|[0-9])([.][0-9]+)?([eE][+-]?[0-9]+)?/.match(s)
  if m && m.begin(0) == 0
    if m[3] && !m[2]
      [:val, m[0], Integer(m[1])*(10**Integer(m[3][1..-1]))]
    elsif m[2]
      [:val, m[0], Float(m[0])]
    else
      [:val, m[0], Integer(m[0])]
    end
  end
end












#objenc(x) ⇒ Object










436
437
438




# File 'lib/xplenty/kensa/okjson.rb', line 436

def objenc(x)
  '{' + x.map{|k,v| keyenc(k) + ':' + valenc(v)}.join(',') + '}'
end












#objparse(ts) ⇒ Object







Parses an “object” in the sense of RFC 4627. Returns the parsed value and any trailing tokens.















91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116




# File 'lib/xplenty/kensa/okjson.rb', line 91

def objparse(ts)
  ts = eat('{', ts)
  obj = {}

  if ts[0][0] == '}'
    return obj, ts[1..-1]
  end

  k, v, ts = pairparse(ts)
  obj[k] = v

  if ts[0][0] == '}'
    return obj, ts[1..-1]
  end

  loop do
    ts = eat(',', ts)

    k, v, ts = pairparse(ts)
    obj[k] = v

    if ts[0][0] == '}'
      return obj, ts[1..-1]
    end
  end
end












#pairparse(ts) ⇒ Object







Parses a “member” in the sense of RFC 4627. Returns the parsed values and any trailing tokens.















121
122
123
124
125
126
127
128
129




# File 'lib/xplenty/kensa/okjson.rb', line 121

def pairparse(ts)
  (typ, _, k), ts = ts[0], ts[1..-1]
  if typ != :str
    raise Error, "unexpected #{k.inspect}"
  end
  ts = eat(':', ts)
  v, ts = valparse(ts)
  [k, v, ts]
end












#strenc(s) ⇒ Object










455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490




# File 'lib/xplenty/kensa/okjson.rb', line 455

def strenc(s)
  t = StringIO.new
  t.putc(?")
  r = 0

  # In ruby >= 1.9, s[r] is a codepoint, not a byte.
  rubydoesenc = s.class.method_defined?(:encoding)

  while r < s.length
    case s[r]
    when ?"  then t.print('\\"')
    when ?\\ then t.print('\\\\')
    when ?\b then t.print('\\b')
    when ?\f then t.print('\\f')
    when ?\n then t.print('\\n')
    when ?\r then t.print('\\r')
    when ?\t then t.print('\\t')
    else
      c = s[r]
      case true
      when Spc <= c && c <= ?~
        t.putc(c)
      when rubydoesenc
        u = c.ord
        surrenc(t, u)
      else
        u, size = uchardec(s, r)
        r += size - 1 # we add one more at the bottom of the loop
        surrenc(t, u)
      end
    end
    r += 1
  end
  t.putc(?")
  t.string
end












#strtok(s) ⇒ Object










241
242
243
244
245
246
247




# File 'lib/xplenty/kensa/okjson.rb', line 241

def strtok(s)
  m = /"([^"\\]|\\["\/\\bfnrt]|\\u[0-9a-fA-F]{4})*"/.match(s)
  if ! m
    raise Error, "invalid string literal at #{abbrev(s)}"
  end
  [:str, m[0], unquote(m[0])]
end












#subst(u1, u2) ⇒ Object










369
370
371
372
373
374




# File 'lib/xplenty/kensa/okjson.rb', line 369

def subst(u1, u2)
  if Usurr1 <= u1 && u1 < Usurr2 && Usurr2 <= u2 && u2 < Usurr3
    return ((u1-Usurr1)<<10) | (u2-Usurr2) + Usurrself
  end
  return Ucharerr
end












#surrenc(t, u) ⇒ Object










493
494
495
496
497
498
499
500
501
502
503
504




# File 'lib/xplenty/kensa/okjson.rb', line 493

def surrenc(t, u)
  if u < 0x10000
    t.print('\\u')
    hexenc4(t, u)
  else
    u1, u2 = unsubst(u)
    t.print('\\u')
    hexenc4(t, u1)
    t.print('\\u')
    hexenc4(t, u2)
  end
end












#surrogate?(u) ⇒ Boolean













Returns:


    

  • 
(Boolean)

    • 










386
387
388




# File 'lib/xplenty/kensa/okjson.rb', line 386

def surrogate?(u)
  Usurr1 <= u && u < Usurr3
end












#textparse(ts) ⇒ Object







Parses a “json text” in the sense of RFC 4627. Returns the parsed value and any trailing tokens. Note: this is almost the same as valparse, except that it does not accept atomic values.















56
57
58
59
60
61
62
63
64
65
66
67
68




# File 'lib/xplenty/kensa/okjson.rb', line 56

def textparse(ts)
  if ts.length < 0
    raise Error, 'empty'
  end

  typ, _, val = ts[0]
  case typ
  when '{' then objparse(ts)
  when '[' then arrparse(ts)
  else
    raise Error, "unexpected #{val.inspect}"
  end
end












#tok(s) ⇒ Object







Scans the first token in s and returns a 3-element list, or nil if s does not begin with a valid token.


The first list element is one of ‘‘’, ‘:’, ‘,’, ‘[’, ‘]’, :val, :str, and :space.


The second element is the lexeme.


The third element is the value of the token for :val and :str, otherwise it is the lexeme.















201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219




# File 'lib/xplenty/kensa/okjson.rb', line 201

def tok(s)
  case s[0]
  when ?{  then ['{', s[0,1], s[0,1]]
  when ?}  then ['}', s[0,1], s[0,1]]
  when ?:  then [':', s[0,1], s[0,1]]
  when ?,  then [',', s[0,1], s[0,1]]
  when ?[  then ['[', s[0,1], s[0,1]]
  when ?]  then [']', s[0,1], s[0,1]]
  when ?n  then nulltok(s)
  when ?t  then truetok(s)
  when ?f  then falsetok(s)
  when ?"  then strtok(s)
  when Spc then [:space, s[0,1], s[0,1]]
  when ?\t then [:space, s[0,1], s[0,1]]
  when ?\n then [:space, s[0,1], s[0,1]]
  when ?\r then [:space, s[0,1], s[0,1]]
  else          numtok(s)
  end
end












#truetok(s) ⇒ Object










223




# File 'lib/xplenty/kensa/okjson.rb', line 223

def truetok(s);  s[0,4] == 'true'  && [:val, 'true',  true]  end












#uchardec(s, i) ⇒ Object







Decodes unicode character u from UTF-8 bytes in string s at position i. Returns u and the number of bytes read.















526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577




# File 'lib/xplenty/kensa/okjson.rb', line 526

def uchardec(s, i)
  n = s.length - i
  return [Ucharerr, 1] if n < 1

  c0 = s[i].ord

  # 1-byte, 7-bit sequence?
  if c0 < Utagx
    return [c0, 1]
  end

  # unexpected continuation byte?
  return [Ucharerr, 1] if c0 < Utag2

  # need continuation byte
  return [Ucharerr, 1] if n < 2
  c1 = s[i+1].ord
  return [Ucharerr, 1] if c1 < Utagx || Utag2 <= c1

  # 2-byte, 11-bit sequence?
  if c0 < Utag3
    u = (c0&Umask2)<<6 | (c1&Umaskx)
    return [Ucharerr, 1] if u <= Uchar1max
    return [u, 2]
  end

  # need second continuation byte
  return [Ucharerr, 1] if n < 3
  c2 = s[i+2].ord
  return [Ucharerr, 1] if c2 < Utagx || Utag2 <= c2

  # 3-byte, 16-bit sequence?
  if c0 < Utag4
    u = (c0&Umask3)<<12 | (c1&Umaskx)<<6 | (c2&Umaskx)
    return [Ucharerr, 1] if u <= Uchar2max
    return [u, 3]
  end

  # need third continuation byte
  return [Ucharerr, 1] if n < 4
  c3 = s[i+3].ord
  return [Ucharerr, 1] if c3 < Utagx || Utag2 <= c3

  # 4-byte, 21-bit sequence?
  if c0 < Utag5
    u = (c0&Umask4)<<18 | (c1&Umaskx)<<12 | (c2&Umaskx)<<6 | (c3&Umaskx)
    return [Ucharerr, 1] if u <= Uchar3max
    return [u, 4]
  end

  return [Ucharerr, 1]
end












#ucharenc(a, i, u) ⇒ Object







Encodes unicode character u as UTF-8 bytes in string a at position i. Returns the number of bytes written.















337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358




# File 'lib/xplenty/kensa/okjson.rb', line 337

def ucharenc(a, i, u)
  case true
  when u <= Uchar1max
    a[i] = (u & 0xff).chr
    1
  when u <= Uchar2max
    a[i+0] = (Utag2 | ((u>>6)&0xff)).chr
    a[i+1] = (Utagx | (u&Umaskx)).chr
    2
  when u <= Uchar3max
    a[i+0] = (Utag3 | ((u>>12)&0xff)).chr
    a[i+1] = (Utagx | ((u>>6)&Umaskx)).chr
    a[i+2] = (Utagx | (u&Umaskx)).chr
    3
  else
    a[i+0] = (Utag4 | ((u>>18)&0xff)).chr
    a[i+1] = (Utagx | ((u>>12)&Umaskx)).chr
    a[i+2] = (Utagx | ((u>>6)&Umaskx)).chr
    a[i+3] = (Utagx | (u&Umaskx)).chr
    4
  end
end












#unquote(q) ⇒ Object







Converts a quoted json string literal q into a UTF-8-encoded string. The rules are different than for Ruby, so we cannot use eval. Unquote will raise an error if q contains control characters.















262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331




# File 'lib/xplenty/kensa/okjson.rb', line 262

def unquote(q)
  q = q[1...-1]
  a = q.dup # allocate a big enough string
  rubydoesenc = false
  # In ruby >= 1.9, a[w] is a codepoint, not a byte.
  if a.class.method_defined?(:force_encoding)
    a.force_encoding('UTF-8')
    rubydoesenc = true
  end
  r, w = 0, 0
  while r < q.length
    c = q[r]
    case true
    when c == ?\\
      r += 1
      if r >= q.length
        raise Error, "string literal ends with a \"\\\": \"#{q}\""
      end

      case q[r]
      when ?",?\\,?/,?'
        a[w] = q[r]
        r += 1
        w += 1
      when ?b,?f,?n,?r,?t
        a[w] = Unesc[q[r]]
        r += 1
        w += 1
      when ?u
        r += 1
        uchar = begin
          hexdec4(q[r,4])
        rescue RuntimeError => e
          raise Error, "invalid escape sequence \\u#{q[r,4]}: #{e}"
        end
        r += 4
        if surrogate? uchar
          if q.length >= r+6
            uchar1 = hexdec4(q[r+2,4])
            uchar = subst(uchar, uchar1)
            if uchar != Ucharerr
              # A valid pair; consume.
              r += 6
            end
          end
        end
        if rubydoesenc
          a[w] = '' << uchar
          w += 1
        else
          w += ucharenc(a, w, uchar)
        end
      else
        raise Error, "invalid escape char #{q[r]} in \"#{q}\""
      end
    when c == ?", c < Spc
      raise Error, "invalid character in string literal \"#{q}\""
    else
      # Copy anything else byte-for-byte.
      # Valid UTF-8 will remain valid UTF-8.
      # Invalid UTF-8 will remain invalid UTF-8.
      # In ruby >= 1.9, c is a codepoint, not a byte,
      # in which case this is still what we want.
      a[w] = c
      r += 1
      w += 1
    end
  end
  a[0,w]
end












#unsubst(u) ⇒ Object










377
378
379
380
381
382
383




# File 'lib/xplenty/kensa/okjson.rb', line 377

def unsubst(u)
  if u < Usurrself || u > Umax || surrogate?(u)
    return Ucharerr, Ucharerr
  end
  u -= Usurrself
  [Usurr1 + ((u>>10)&0x3ff), Usurr2 + (u&0x3ff)]
end












#valenc(x) ⇒ Object










421
422
423
424
425
426
427
428
429
430
431
432
433




# File 'lib/xplenty/kensa/okjson.rb', line 421

def valenc(x)
  case x
  when Hash    then objenc(x)
  when Array   then arrenc(x)
  when String  then strenc(x)
  when Numeric then numenc(x)
  when true    then "true"
  when false   then "false"
  when nil     then "null"
  else
    raise Error, "cannot encode #{x.class}: #{x.inspect}"
  end
end












#valparse(ts) ⇒ Object







Parses a “value” in the sense of RFC 4627. Returns the parsed value and any trailing tokens.















73
74
75
76
77
78
79
80
81
82
83
84
85
86




# File 'lib/xplenty/kensa/okjson.rb', line 73

def valparse(ts)
  if ts.length < 0
    raise Error, 'empty'
  end

  typ, _, val = ts[0]
  case typ
  when '{' then objparse(ts)
  when '[' then arrparse(ts)
  when :val,:str then [val, ts[1..-1]]
  else
    raise Error, "unexpected #{val.inspect}"
  end
end