smartypants.go (view raw)
1//
2// Blackfriday Markdown Processor
3// Available at http://github.com/russross/blackfriday
4//
5// Copyright © 2011 Russ Ross <russ@russross.com>.
6// Distributed under the Simplified BSD License.
7// See README.md for details.
8//
9
10//
11//
12// SmartyPants rendering
13//
14//
15
16package blackfriday
17
18import (
19 "bytes"
20 "io"
21)
22
23// SPRenderer is a struct containing state of a Smartypants renderer.
24type SPRenderer struct {
25 inSingleQuote bool
26 inDoubleQuote bool
27 callbacks [256]smartCallback
28}
29
30func wordBoundary(c byte) bool {
31 return c == 0 || isspace(c) || ispunct(c)
32}
33
34func tolower(c byte) byte {
35 if c >= 'A' && c <= 'Z' {
36 return c - 'A' + 'a'
37 }
38 return c
39}
40
41func isdigit(c byte) bool {
42 return c >= '0' && c <= '9'
43}
44
45func smartQuoteHelper(out *bytes.Buffer, previousChar byte, nextChar byte, quote byte, isOpen *bool, addNBSP bool) bool {
46 // edge of the buffer is likely to be a tag that we don't get to see,
47 // so we treat it like text sometimes
48
49 // enumerate all sixteen possibilities for (previousChar, nextChar)
50 // each can be one of {0, space, punct, other}
51 switch {
52 case previousChar == 0 && nextChar == 0:
53 // context is not any help here, so toggle
54 *isOpen = !*isOpen
55 case isspace(previousChar) && nextChar == 0:
56 // [ "] might be [ "<code>foo...]
57 *isOpen = true
58 case ispunct(previousChar) && nextChar == 0:
59 // [!"] hmm... could be [Run!"] or [("<code>...]
60 *isOpen = false
61 case /* isnormal(previousChar) && */ nextChar == 0:
62 // [a"] is probably a close
63 *isOpen = false
64 case previousChar == 0 && isspace(nextChar):
65 // [" ] might be [...foo</code>" ]
66 *isOpen = false
67 case isspace(previousChar) && isspace(nextChar):
68 // [ " ] context is not any help here, so toggle
69 *isOpen = !*isOpen
70 case ispunct(previousChar) && isspace(nextChar):
71 // [!" ] is probably a close
72 *isOpen = false
73 case /* isnormal(previousChar) && */ isspace(nextChar):
74 // [a" ] this is one of the easy cases
75 *isOpen = false
76 case previousChar == 0 && ispunct(nextChar):
77 // ["!] hmm... could be ["$1.95] or [</code>"!...]
78 *isOpen = false
79 case isspace(previousChar) && ispunct(nextChar):
80 // [ "!] looks more like [ "$1.95]
81 *isOpen = true
82 case ispunct(previousChar) && ispunct(nextChar):
83 // [!"!] context is not any help here, so toggle
84 *isOpen = !*isOpen
85 case /* isnormal(previousChar) && */ ispunct(nextChar):
86 // [a"!] is probably a close
87 *isOpen = false
88 case previousChar == 0 /* && isnormal(nextChar) */ :
89 // ["a] is probably an open
90 *isOpen = true
91 case isspace(previousChar) /* && isnormal(nextChar) */ :
92 // [ "a] this is one of the easy cases
93 *isOpen = true
94 case ispunct(previousChar) /* && isnormal(nextChar) */ :
95 // [!"a] is probably an open
96 *isOpen = true
97 default:
98 // [a'b] maybe a contraction?
99 *isOpen = false
100 }
101
102 // Note that with the limited lookahead, this non-breaking
103 // space will also be appended to single double quotes.
104 if addNBSP && !*isOpen {
105 out.WriteString(" ")
106 }
107
108 out.WriteByte('&')
109 if *isOpen {
110 out.WriteByte('l')
111 } else {
112 out.WriteByte('r')
113 }
114 out.WriteByte(quote)
115 out.WriteString("quo;")
116
117 if addNBSP && *isOpen {
118 out.WriteString(" ")
119 }
120
121 return true
122}
123
124func (r *SPRenderer) smartSingleQuote(out *bytes.Buffer, previousChar byte, text []byte) int {
125 if len(text) >= 2 {
126 t1 := tolower(text[1])
127
128 if t1 == '\'' {
129 nextChar := byte(0)
130 if len(text) >= 3 {
131 nextChar = text[2]
132 }
133 if smartQuoteHelper(out, previousChar, nextChar, 'd', &r.inDoubleQuote, false) {
134 return 1
135 }
136 }
137
138 if (t1 == 's' || t1 == 't' || t1 == 'm' || t1 == 'd') && (len(text) < 3 || wordBoundary(text[2])) {
139 out.WriteString("’")
140 return 0
141 }
142
143 if len(text) >= 3 {
144 t2 := tolower(text[2])
145
146 if ((t1 == 'r' && t2 == 'e') || (t1 == 'l' && t2 == 'l') || (t1 == 'v' && t2 == 'e')) &&
147 (len(text) < 4 || wordBoundary(text[3])) {
148 out.WriteString("’")
149 return 0
150 }
151 }
152 }
153
154 nextChar := byte(0)
155 if len(text) > 1 {
156 nextChar = text[1]
157 }
158 if smartQuoteHelper(out, previousChar, nextChar, 's', &r.inSingleQuote, false) {
159 return 0
160 }
161
162 out.WriteByte(text[0])
163 return 0
164}
165
166func (r *SPRenderer) smartParens(out *bytes.Buffer, previousChar byte, text []byte) int {
167 if len(text) >= 3 {
168 t1 := tolower(text[1])
169 t2 := tolower(text[2])
170
171 if t1 == 'c' && t2 == ')' {
172 out.WriteString("©")
173 return 2
174 }
175
176 if t1 == 'r' && t2 == ')' {
177 out.WriteString("®")
178 return 2
179 }
180
181 if len(text) >= 4 && t1 == 't' && t2 == 'm' && text[3] == ')' {
182 out.WriteString("™")
183 return 3
184 }
185 }
186
187 out.WriteByte(text[0])
188 return 0
189}
190
191func (r *SPRenderer) smartDash(out *bytes.Buffer, previousChar byte, text []byte) int {
192 if len(text) >= 2 {
193 if text[1] == '-' {
194 out.WriteString("—")
195 return 1
196 }
197
198 if wordBoundary(previousChar) && wordBoundary(text[1]) {
199 out.WriteString("–")
200 return 0
201 }
202 }
203
204 out.WriteByte(text[0])
205 return 0
206}
207
208func (r *SPRenderer) smartDashLatex(out *bytes.Buffer, previousChar byte, text []byte) int {
209 if len(text) >= 3 && text[1] == '-' && text[2] == '-' {
210 out.WriteString("—")
211 return 2
212 }
213 if len(text) >= 2 && text[1] == '-' {
214 out.WriteString("–")
215 return 1
216 }
217
218 out.WriteByte(text[0])
219 return 0
220}
221
222func (r *SPRenderer) smartAmpVariant(out *bytes.Buffer, previousChar byte, text []byte, quote byte, addNBSP bool) int {
223 if bytes.HasPrefix(text, []byte(""")) {
224 nextChar := byte(0)
225 if len(text) >= 7 {
226 nextChar = text[6]
227 }
228 if smartQuoteHelper(out, previousChar, nextChar, quote, &r.inDoubleQuote, addNBSP) {
229 return 5
230 }
231 }
232
233 if bytes.HasPrefix(text, []byte("�")) {
234 return 3
235 }
236
237 out.WriteByte('&')
238 return 0
239}
240
241func (r *SPRenderer) smartAmp(angledQuotes, addNBSP bool) func(*bytes.Buffer, byte, []byte) int {
242 var quote byte = 'd'
243 if angledQuotes {
244 quote = 'a'
245 }
246
247 return func(out *bytes.Buffer, previousChar byte, text []byte) int {
248 return r.smartAmpVariant(out, previousChar, text, quote, addNBSP)
249 }
250}
251
252func (r *SPRenderer) smartPeriod(out *bytes.Buffer, previousChar byte, text []byte) int {
253 if len(text) >= 3 && text[1] == '.' && text[2] == '.' {
254 out.WriteString("…")
255 return 2
256 }
257
258 if len(text) >= 5 && text[1] == ' ' && text[2] == '.' && text[3] == ' ' && text[4] == '.' {
259 out.WriteString("…")
260 return 4
261 }
262
263 out.WriteByte(text[0])
264 return 0
265}
266
267func (r *SPRenderer) smartBacktick(out *bytes.Buffer, previousChar byte, text []byte) int {
268 if len(text) >= 2 && text[1] == '`' {
269 nextChar := byte(0)
270 if len(text) >= 3 {
271 nextChar = text[2]
272 }
273 if smartQuoteHelper(out, previousChar, nextChar, 'd', &r.inDoubleQuote, false) {
274 return 1
275 }
276 }
277
278 out.WriteByte(text[0])
279 return 0
280}
281
282func (r *SPRenderer) smartNumberGeneric(out *bytes.Buffer, previousChar byte, text []byte) int {
283 if wordBoundary(previousChar) && previousChar != '/' && len(text) >= 3 {
284 // is it of the form digits/digits(word boundary)?, i.e., \d+/\d+\b
285 // note: check for regular slash (/) or fraction slash (⁄, 0x2044, or 0xe2 81 84 in utf-8)
286 // and avoid changing dates like 1/23/2005 into fractions.
287 numEnd := 0
288 for len(text) > numEnd && isdigit(text[numEnd]) {
289 numEnd++
290 }
291 if numEnd == 0 {
292 out.WriteByte(text[0])
293 return 0
294 }
295 denStart := numEnd + 1
296 if len(text) > numEnd+3 && text[numEnd] == 0xe2 && text[numEnd+1] == 0x81 && text[numEnd+2] == 0x84 {
297 denStart = numEnd + 3
298 } else if len(text) < numEnd+2 || text[numEnd] != '/' {
299 out.WriteByte(text[0])
300 return 0
301 }
302 denEnd := denStart
303 for len(text) > denEnd && isdigit(text[denEnd]) {
304 denEnd++
305 }
306 if denEnd == denStart {
307 out.WriteByte(text[0])
308 return 0
309 }
310 if len(text) == denEnd || wordBoundary(text[denEnd]) && text[denEnd] != '/' {
311 out.WriteString("<sup>")
312 out.Write(text[:numEnd])
313 out.WriteString("</sup>⁄<sub>")
314 out.Write(text[denStart:denEnd])
315 out.WriteString("</sub>")
316 return denEnd - 1
317 }
318 }
319
320 out.WriteByte(text[0])
321 return 0
322}
323
324func (r *SPRenderer) smartNumber(out *bytes.Buffer, previousChar byte, text []byte) int {
325 if wordBoundary(previousChar) && previousChar != '/' && len(text) >= 3 {
326 if text[0] == '1' && text[1] == '/' && text[2] == '2' {
327 if len(text) < 4 || wordBoundary(text[3]) && text[3] != '/' {
328 out.WriteString("½")
329 return 2
330 }
331 }
332
333 if text[0] == '1' && text[1] == '/' && text[2] == '4' {
334 if len(text) < 4 || wordBoundary(text[3]) && text[3] != '/' || (len(text) >= 5 && tolower(text[3]) == 't' && tolower(text[4]) == 'h') {
335 out.WriteString("¼")
336 return 2
337 }
338 }
339
340 if text[0] == '3' && text[1] == '/' && text[2] == '4' {
341 if len(text) < 4 || wordBoundary(text[3]) && text[3] != '/' || (len(text) >= 6 && tolower(text[3]) == 't' && tolower(text[4]) == 'h' && tolower(text[5]) == 's') {
342 out.WriteString("¾")
343 return 2
344 }
345 }
346 }
347
348 out.WriteByte(text[0])
349 return 0
350}
351
352func (r *SPRenderer) smartDoubleQuoteVariant(out *bytes.Buffer, previousChar byte, text []byte, quote byte) int {
353 nextChar := byte(0)
354 if len(text) > 1 {
355 nextChar = text[1]
356 }
357 if !smartQuoteHelper(out, previousChar, nextChar, quote, &r.inDoubleQuote, false) {
358 out.WriteString(""")
359 }
360
361 return 0
362}
363
364func (r *SPRenderer) smartDoubleQuote(out *bytes.Buffer, previousChar byte, text []byte) int {
365 return r.smartDoubleQuoteVariant(out, previousChar, text, 'd')
366}
367
368func (r *SPRenderer) smartAngledDoubleQuote(out *bytes.Buffer, previousChar byte, text []byte) int {
369 return r.smartDoubleQuoteVariant(out, previousChar, text, 'a')
370}
371
372func (r *SPRenderer) smartLeftAngle(out *bytes.Buffer, previousChar byte, text []byte) int {
373 i := 0
374
375 for i < len(text) && text[i] != '>' {
376 i++
377 }
378
379 out.Write(text[:i+1])
380 return i
381}
382
383type smartCallback func(out *bytes.Buffer, previousChar byte, text []byte) int
384
385// NewSmartypantsRenderer constructs a Smartypants renderer object.
386func NewSmartypantsRenderer(flags HTMLFlags) *SPRenderer {
387 var (
388 r SPRenderer
389
390 smartAmpAngled = r.smartAmp(true, false)
391 smartAmpAngledNBSP = r.smartAmp(true, true)
392 smartAmpRegular = r.smartAmp(false, false)
393 smartAmpRegularNBSP = r.smartAmp(false, true)
394
395 addNBSP = flags&SmartypantsQuotesNBSP != 0
396 )
397
398 if flags&SmartypantsAngledQuotes == 0 {
399 r.callbacks['"'] = r.smartDoubleQuote
400 if !addNBSP {
401 r.callbacks['&'] = smartAmpRegular
402 } else {
403 r.callbacks['&'] = smartAmpRegularNBSP
404 }
405 } else {
406 r.callbacks['"'] = r.smartAngledDoubleQuote
407 if !addNBSP {
408 r.callbacks['&'] = smartAmpAngled
409 } else {
410 r.callbacks['&'] = smartAmpAngledNBSP
411 }
412 }
413 r.callbacks['\''] = r.smartSingleQuote
414 r.callbacks['('] = r.smartParens
415 if flags&SmartypantsDashes != 0 {
416 if flags&SmartypantsLatexDashes == 0 {
417 r.callbacks['-'] = r.smartDash
418 } else {
419 r.callbacks['-'] = r.smartDashLatex
420 }
421 }
422 r.callbacks['.'] = r.smartPeriod
423 if flags&SmartypantsFractions == 0 {
424 r.callbacks['1'] = r.smartNumber
425 r.callbacks['3'] = r.smartNumber
426 } else {
427 for ch := '1'; ch <= '9'; ch++ {
428 r.callbacks[ch] = r.smartNumberGeneric
429 }
430 }
431 r.callbacks['<'] = r.smartLeftAngle
432 r.callbacks['`'] = r.smartBacktick
433 return &r
434}
435
436// Process is the entry point of the Smartypants renderer.
437func (r *SPRenderer) Process(w io.Writer, text []byte) {
438 mark := 0
439 for i := 0; i < len(text); i++ {
440 if action := r.callbacks[text[i]]; action != nil {
441 if i > mark {
442 w.Write(text[mark:i])
443 }
444 previousChar := byte(0)
445 if i > 0 {
446 previousChar = text[i-1]
447 }
448 var tmp bytes.Buffer
449 i += action(&tmp, previousChar, text[i:])
450 w.Write(tmp.Bytes())
451 mark = i + 1
452 }
453 }
454 if mark < len(text) {
455 w.Write(text[mark:])
456 }
457}