summaryrefslogtreecommitdiff
path: root/lib/regex/interp.myr
blob: 2261de08991e0c102e698dd0a2c117494507fd8d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
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
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
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
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
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
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
use std

use "types"

pkg regex =
	/* regex execution */
	const exec	: (re : regex#, str : byte[:] -> std.option(byte[:][:]))
	const search	: (re : regex#, str : byte[:] -> std.option(byte[:][:]))

	/* regex execution returning indexes */
	const iexec	: (re : regex#, str : byte[:] -> std.option((std.size, std.size)[:]))
	const isearch	: (re : regex#, str : byte[:] -> std.option((std.size, std.size)[:]))

	/* substitution */
	const sub	: (re : regex#, str : byte[:], subst : byte[:][:] -> std.option(byte[:]))
	const sbsub	: (sb : std.strbuf#, re : regex#, str : byte[:], subst : byte[:][:] -> bool)
	const suball	: (re : regex#, str : byte[:], subst : byte[:][:] -> byte[:])
	const sbsuball	: (sb : std.strbuf#, re : regex#, str : byte[:], subst : byte[:][:] -> void)

	const matchfree	: (pat : byte[:][:] -> void)
;;

/* Ugly: for performance. std.option() should be used instead when unions get faster. */
const Zthr = (0 : rethread#)

const exec = {re, str
	var thr, m

	thr = run(re, str, 0, true)
	m = getmatches(re, thr)
	cleanup(re)
	-> m
}

const iexec = {re, str
	var thr, m

	thr = run(re, str, 0, true)
	m = getidxmatches(re, thr)
	cleanup(re)
	-> m
}

const search = {re, str
	var thr
	var m

	m = `std.None
	for var i = 0; i < str.len; i++
		thr = run(re, str[i:], 0, false)
		m = getmatches(re, thr)
		match m
		| `std.Some _:	break
		| `std.None:	/* nothing */
		;;
		cleanup(re)
	;;
	->  m
}

const isearch = {re, str
	var thr
	var m

	m = `std.None
	for var i = 0; i < str.len; i++
		thr = run(re, str[i:], 0, false)
		m = getidxmatches(re, thr)
		match m
		| `std.Some _:	break
		| `std.None:	/* nothing */
		;;
		cleanup(re)
	;;
	->  m
}

const sub = {re, str, subst
	var sb

	sb = std.mksb()
	if !sbsub(sb, re, str, subst)
		-> `std.None
	else
		-> `std.Some std.sbfin(sb)
	;;
}

const sbsub = {sb, re, str, subst
	var thr, m

	/* we always have m[0] as the full match */
	if re.nmatch != subst.len + 1
		-> false
	;;

	thr = run(re, str, 0, true)
	if thr == Zthr
		m = false
	else
		m = dosubst(sb, re, thr, str, subst)
	;;
	cleanup(re)
	-> m
}

const suball = {re, str, subst
	var sb

	sb = std.mksb()
	sbsuball(sb, re, str, subst)
	-> std.sbfin(sb)
}

const sbsuball = {sb, re, str, subst
	var thr, len, s, i

	/* we always have m[0] as the full match */
	if re.nmatch != subst.len + 1
		-> void
	;;

	i = 0
	while i < str.len
		thr = run(re, str[i:], 0, false)
		if thr == Zthr
			std.sbputb(sb, str[i])
			i++
		else
			len = thr.mend[0]
			s = str[i:len + i]
			dosubst(sb, re, thr, s, subst)
			i += len
		;;
		cleanup(re)
	;;
}


const dosubst = {sb, re, thr, str, subst
	var off

	off = 0
	for var i = 1; i < re.nmatch; i++
		if thr.mstart[i] != -1 && thr.mend[i] != -1
			std.sbputs(sb, str[off:thr.mstart[i]])
			std.sbputs(sb, subst[i - 1])
			off = thr.mend[i]
		;;
	;;
	std.sbputs(sb, str[off:])
	thrfree(re, thr)
	-> true
}

const cleanup = {re
	var thr, next

	for thr = re.runq; thr != Zthr; thr = next
		next = thr.next
		thrfree(re, thr)
	;;
	for thr = re.expired; thr != Zthr; thr = next
		next = thr.next
		thrfree(re, thr)
	;;
	re.runq = Zthr
	re.expired = Zthr
	re.nexttid = 0
	re.nthr = 0
}

const matchfree = {m
	std.slfree(m)
}

const getmatches = {re, thr
	var ret

	if thr == Zthr
		-> `std.None
	;;
	ret = std.slalloc(re.nmatch)
	for var i = 0; i < re.nmatch; i++
		if thr.mstart[i] != -1 && thr.mend[i] != -1
			ret[i] = re.str[thr.mstart[i]:thr.mend[i]]
		else
			ret[i] = [][:]
		;;
	;;
	thrfree(re, thr)
	-> `std.Some ret
}

const getidxmatches = {re, thr
	var ret

	if thr == Zthr
		-> `std.None
	;;
	ret = std.slalloc(re.nmatch)
	for var i = 0; i < re.nmatch; i++
		if thr.mstart[i] != -1 && thr.mend[i] != -1
			ret[i] = (thr.mstart[i], thr.mend[i])
		else
			ret[i] = (-1, -1)
		;;
	;;
	thrfree(re, thr)
	-> `std.Some ret
}

/* returns a matching thread, or Zthr if no threads matched */
const run = {re, str, idx, wholestr
	var bestmatch
	var consumed
	var states
	var thr
	var ip

	re.str = str
	re.strp = 0

	bestmatch = Zthr
	states = std.mkbs()
	re.runq = mkthread(re, 0)
	if re.debug
		/* The last run could have left things here, since we need this info after the run */
		for bs : re.traces
			std.bsfree(bs)
		;;
		std.slfree(re.traces)
		re.traces = [][:]
		std.slpush(&re.traces, std.mkbs())
	;;
	re.runq.mstart = std.slalloc(re.nmatch)
	re.runq.mend = std.slalloc(re.nmatch)
	for var i = 0; i < re.nmatch; i++
		re.runq.mstart[i] = -1
		re.runq.mend[i] = -1
	;;
	while re.nthr > 0
		while re.runq != Zthr
			/* set up the next thread */
			thr = re.runq
			re.runq = thr.next

			trace(re, thr, "\nrunning tid={}, ip={}, s[{}]={}\n", \
				thr.tid, thr.ip, re.strp, std.decode(re.str[re.strp:]))
			ip = thr.ip
			consumed = step(re, thr, -1)
			while !consumed
				consumed = step(re, thr, ip)
			;;

			if std.bshas(states, thr.ip)
				die(re, thr, "there can be only one")
			;;

			if thr.dead
				thrfree(re, thr)
			elif thr.matched
				trace(re, thr, "new bestmatch\n")
				if bestmatch != Zthr
					thrfree(re, bestmatch)
				;;

				if re.strp == re.str.len
					bestmatch = thr
					goto done
				elif !wholestr
					bestmatch = thr
				;;
			elif !thr.matched
				std.bsput(states, thr.ip)
				if re.expired == Zthr
					re.expired = thr
				;;
				if re.expiredtail != Zthr
					re.expiredtail.next = thr
				;;
				re.expiredtail = thr
				thr.next = Zthr

			;;
		;;
		std.bsclear(states)
		trace(re, Zthr, "switch\n")
		re.runq = re.expired
		re.expired = Zthr
		re.expiredtail = Zthr
		re.strp++
	;;
:done
	std.bsfree(states)
	-> bestmatch
}

/* 
 Steps forward one instruction. Returns true if a byte of input was
 consumed, false otherwise.
*/
const step = {re, thr, curip
	var str
	var mstart
	var mend

	str = re.str
	match re.prog[thr.ip]
	/* Char matching. Consume exactly one byte from the string. */
	| `Ibyte b:
		trace(re, thr, "\t{}:\tByte {} ({})\n", thr.ip, b, (b : char))
		if !within(re, str)
			die(re, thr, "end of string")
		elif b != str[re.strp]
			die(re, thr, "not right char")
		else
			hit(re, thr)
			thr.ip++
			trace(re, thr, "\t\tmatched {} with {}\n", b, str[re.strp])
		;;
	| `Irange (start, end):
		trace(re, thr, "\t{}:\tRange ({}, {}) /* {} - {} */\n", thr.ip, start, end, (start : char), (end : char))
		if !within(re, str) || start > str[re.strp] || end < str[re.strp]
			die(re, thr, "bad range")
		else
			hit(re, thr)
			thr.ip++
		;;
	/*
	  Non-consuming. All of these return false, and expect step to be
	  called again until exactly one byte is consumed from the string.
	 */
	| `Ibol:
		trace(re, thr, "\t{}:\tBol\n", thr.ip)
		if re.strp == 0 || str[re.strp - 1] == ('\n' : byte)
			hit(re, thr)
			thr.ip++
			-> false
		else
			die(re, thr, "not beginning of line")
		;;
	| `Ieol:
		trace(re, thr, "\t{}:\tEol\n", thr.ip)
		if re.strp == str.len || str[re.strp] == ('\n' : byte)
			hit(re, thr)
			thr.ip++
			-> false
		else
			die(re, thr, "not end of line")
		;;
	/* check for word characters */
	| `Ibow:
		trace(re, thr, "\t{}:\tBow\n", thr.ip)
		if iswordchar(str[re.strp:]) && (re.strp == 0 || !iswordchar(prevchar(str, re.strp)))
			hit(re, thr)
			thr.ip++
			-> false
		else
			die(re, thr, "not beginning of word")
		;;
	| `Ieow:
		trace(re, thr, "\t{}:\tEow\n", thr.ip)
		if re.strp == str.len && iswordchar(prevchar(str, re.strp))
			hit(re, thr)
			thr.ip++
			-> false
		elif re.strp > 0 && !iswordchar(str[re.strp:]) && iswordchar(prevchar(str, re.strp))
			hit(re, thr)
			thr.ip++
			-> false
		else
			die(re, thr, "not end of word")
		;;
	| `Ilbra m:
		trace(re, thr, "\t{}:\tLbra {}\n", thr.ip, m)
		trace(re, thr, "\t\tmatch start = {}\n", re.strp)
		thr.mstart[m] = re.strp
		hit(re, thr)
		thr.ip++
		-> false
	| `Irbra m:
		trace(re, thr, "\t{}:\tRbra {}\n", thr.ip, m)
		thr.mend[m] = re.strp
		hit(re, thr)
		thr.ip++
		-> false
	| `Ifork (lip, rip):
		trace(re, thr, "\t{}:\tFork ({}, {})\n", thr.ip, lip, rip)
		mstart = std.sldup(thr.mstart)
		mend = std.sldup(thr.mend)
		hit(re, thr)
		fork(re, rip, curip, mstart, mend)
		if re.debug
			std.slpush(&re.traces, std.bsdup(re.traces[thr.tid]))
		;;
		thr.ip = lip
		-> false
	| `Ijmp ip:
		trace(re, thr, "\t{}:\tJmp {}\n", thr.ip, ip)
		hit(re, thr)
		thr.ip = ip
		-> false
	| `Imatch id:
		trace(re, thr, "\t{}:\tMatch\n", thr.ip)
		re.lastthr = thr.tid
		finish(re, thr)
		-> true
	;;
	-> true
}

const fork = {re, ip, curip, mstart, mend
	var thr

	if ip == curip /* loop detection */
		-> void
	;;
	thr = mkthread(re, ip)
	thr.next = re.runq
	thr.mstart = mstart
	thr.mend = mend
	re.runq = thr
}

const die = {re, thr, msg
        /*
 	  we can have die called on a thread
	  multiple times, eg, if it has a bad
	  range *and* end in a state that another
	  thread is in. We should only decrement
	  the number of threads for that once.
	 */
	trace(re, thr, "\t\tdie {}: {}\n", thr.tid, msg)
        if !thr.dead
		re.nthr--
	;;
	re.lastip = thr.ip
	re.lastthr = thr.tid
	thr.dead = true
}

const finish = {re, thr
	trace(re, thr, "finish {}\n", thr.tid)
	thr.matched = true
	re.nthr--
}

const mkthread = {re, ip
	var thr : rethread#

	thr = std.alloc()

	thr.next = Zthr

	thr.ip = ip
	thr.tid = re.nexttid++
	thr.dead = false
	thr.matched = false

	thr.mstart = [][:]
	thr.mend = [][:]

	re.nthr++

	-> thr
}

const thrfree = {re, thr
	trace(re, thr, "\t\tcleanup {}\n", thr.tid)
	std.slfree(thr.mstart)
	std.slfree(thr.mend)
	std.free(thr)
}

const within = {re, str 
	-> re.strp < str.len
}

const trace : (re : regex#, thr : rethread#, msg : byte[:], args : ... -> void) = {re, thr, msg, args
	var ap

	if re.trace && thr != Zthr
		ap = std.vastart(&args)
		std.putv(msg, &ap)
		std.put("\t{}\n", re.pat)
		std.put("\t")
		for var i = 0; i < re.pcidx[thr.ip] - 1; i++
			std.put(" ")
		;;
		std.put("^\n")
	;;
}

const hit = {re, thr
        if re.debug
                std.bsput(re.traces[thr.tid], thr.ip)
        ;;
}

/* must be called with i >= 1 */
const prevchar = {s, i
	std.assert(i != 0, "prevchar must be called with i >= 1\n")
	i--
	while i != 0 && s[i] >= 0x80
		i--
	;;
	-> s[i:]
}

const iswordchar = {s
	var c

	c = std.decode(s)
	-> std.isalpha(c) || std.isdigit(c) || c == '_'
}