1 /*
2  * Copyright (C) 2010 The Android Open Source Project
3  *
4  * Licensed under the Apache License, Version 2.0 (the "License");
5  * you may not use this file except in compliance with the License.
6  * You may obtain a copy of the License at
7  *
8  *      http://www.apache.org/licenses/LICENSE-2.0
9  *
10  * Unless required by applicable law or agreed to in writing, software
11  * distributed under the License is distributed on an "AS IS" BASIS,
12  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13  * See the License for the specific language governing permissions and
14  * limitations under the License.
15  */
16 
17 package com.android.emailcommon.utility;
18 
19 import com.google.common.annotations.VisibleForTesting;
20 
21 import android.graphics.Color;
22 import android.text.Spannable;
23 import android.text.SpannableString;
24 import android.text.SpannableStringBuilder;
25 import android.text.TextUtils;
26 import android.text.style.BackgroundColorSpan;
27 
28 import java.io.IOException;
29 import java.util.ArrayList;
30 import java.util.HashMap;
31 import java.util.Map;
32 import java.util.StringTokenizer;
33 
34 public class TextUtilities {
35     // Highlight color is yellow, as in other apps.
36     // TODO Push for this to be a global (style-related?) constant
37     public static final int HIGHLIGHT_COLOR_INT = Color.YELLOW;
38     // We AND off the "alpha" from the color (i.e. 0xFFFFFF00 -> 0x00FFFF00)
39     /*package*/ static final String HIGHLIGHT_COLOR_STRING =
40         '#' + Integer.toHexString(HIGHLIGHT_COLOR_INT & 0x00FFFFFF);
41 
42     // This is how many chars we'll allow in a snippet
43     private static final int MAX_SNIPPET_LENGTH = 200;
44     // For some reason, isWhitespace() returns false with the following...
45     /*package*/ static final char NON_BREAKING_SPACE_CHARACTER = (char)160;
46 
47     // Tags whose content must be stripped as well
48     static final String[] STRIP_TAGS =
49         new String[] {"title", "script", "style", "applet", "head"};
50     // The number of characters we peel off for testing against STRIP_TAGS; this should be the
51     // maximum size of the strings in STRIP_TAGS
52     static final int MAX_STRIP_TAG_LENGTH = 6;
53 
54     static final Map<String, Character> ESCAPE_STRINGS;
55     static {
56         // HTML character entity references as defined in HTML 4
57         // see http://www.w3.org/TR/REC-html40/sgml/entities.html
58         ESCAPE_STRINGS = new HashMap<String, Character>(252);
59 
60         ESCAPE_STRINGS.put("&nbsp", '\u00A0');
61         ESCAPE_STRINGS.put("&iexcl", '\u00A1');
62         ESCAPE_STRINGS.put("&cent", '\u00A2');
63         ESCAPE_STRINGS.put("&pound", '\u00A3');
64         ESCAPE_STRINGS.put("&curren", '\u00A4');
65         ESCAPE_STRINGS.put("&yen", '\u00A5');
66         ESCAPE_STRINGS.put("&brvbar", '\u00A6');
67         ESCAPE_STRINGS.put("&sect", '\u00A7');
68         ESCAPE_STRINGS.put("&uml", '\u00A8');
69         ESCAPE_STRINGS.put("&copy", '\u00A9');
70         ESCAPE_STRINGS.put("&ordf", '\u00AA');
71         ESCAPE_STRINGS.put("&laquo", '\u00AB');
72         ESCAPE_STRINGS.put("&not", '\u00AC');
73         ESCAPE_STRINGS.put("&shy", '\u00AD');
74         ESCAPE_STRINGS.put("&reg", '\u00AE');
75         ESCAPE_STRINGS.put("&macr", '\u00AF');
76         ESCAPE_STRINGS.put("&deg", '\u00B0');
77         ESCAPE_STRINGS.put("&plusmn", '\u00B1');
78         ESCAPE_STRINGS.put("&sup2", '\u00B2');
79         ESCAPE_STRINGS.put("&sup3", '\u00B3');
80         ESCAPE_STRINGS.put("&acute", '\u00B4');
81         ESCAPE_STRINGS.put("&micro", '\u00B5');
82         ESCAPE_STRINGS.put("&para", '\u00B6');
83         ESCAPE_STRINGS.put("&middot", '\u00B7');
84         ESCAPE_STRINGS.put("&cedil", '\u00B8');
85         ESCAPE_STRINGS.put("&sup1", '\u00B9');
86         ESCAPE_STRINGS.put("&ordm", '\u00BA');
87         ESCAPE_STRINGS.put("&raquo", '\u00BB');
88         ESCAPE_STRINGS.put("&frac14", '\u00BC');
89         ESCAPE_STRINGS.put("&frac12", '\u00BD');
90         ESCAPE_STRINGS.put("&frac34", '\u00BE');
91         ESCAPE_STRINGS.put("&iquest", '\u00BF');
92         ESCAPE_STRINGS.put("&Agrave", '\u00C0');
93         ESCAPE_STRINGS.put("&Aacute", '\u00C1');
94         ESCAPE_STRINGS.put("&Acirc", '\u00C2');
95         ESCAPE_STRINGS.put("&Atilde", '\u00C3');
96         ESCAPE_STRINGS.put("&Auml", '\u00C4');
97         ESCAPE_STRINGS.put("&Aring", '\u00C5');
98         ESCAPE_STRINGS.put("&AElig", '\u00C6');
99         ESCAPE_STRINGS.put("&Ccedil", '\u00C7');
100         ESCAPE_STRINGS.put("&Egrave", '\u00C8');
101         ESCAPE_STRINGS.put("&Eacute", '\u00C9');
102         ESCAPE_STRINGS.put("&Ecirc", '\u00CA');
103         ESCAPE_STRINGS.put("&Euml", '\u00CB');
104         ESCAPE_STRINGS.put("&Igrave", '\u00CC');
105         ESCAPE_STRINGS.put("&Iacute", '\u00CD');
106         ESCAPE_STRINGS.put("&Icirc", '\u00CE');
107         ESCAPE_STRINGS.put("&Iuml", '\u00CF');
108         ESCAPE_STRINGS.put("&ETH", '\u00D0');
109         ESCAPE_STRINGS.put("&Ntilde", '\u00D1');
110         ESCAPE_STRINGS.put("&Ograve", '\u00D2');
111         ESCAPE_STRINGS.put("&Oacute", '\u00D3');
112         ESCAPE_STRINGS.put("&Ocirc", '\u00D4');
113         ESCAPE_STRINGS.put("&Otilde", '\u00D5');
114         ESCAPE_STRINGS.put("&Ouml", '\u00D6');
115         ESCAPE_STRINGS.put("&times", '\u00D7');
116         ESCAPE_STRINGS.put("&Oslash", '\u00D8');
117         ESCAPE_STRINGS.put("&Ugrave", '\u00D9');
118         ESCAPE_STRINGS.put("&Uacute", '\u00DA');
119         ESCAPE_STRINGS.put("&Ucirc", '\u00DB');
120         ESCAPE_STRINGS.put("&Uuml", '\u00DC');
121         ESCAPE_STRINGS.put("&Yacute", '\u00DD');
122         ESCAPE_STRINGS.put("&THORN", '\u00DE');
123         ESCAPE_STRINGS.put("&szlig", '\u00DF');
124         ESCAPE_STRINGS.put("&agrave", '\u00E0');
125         ESCAPE_STRINGS.put("&aacute", '\u00E1');
126         ESCAPE_STRINGS.put("&acirc", '\u00E2');
127         ESCAPE_STRINGS.put("&atilde", '\u00E3');
128         ESCAPE_STRINGS.put("&auml", '\u00E4');
129         ESCAPE_STRINGS.put("&aring", '\u00E5');
130         ESCAPE_STRINGS.put("&aelig", '\u00E6');
131         ESCAPE_STRINGS.put("&ccedil", '\u00E7');
132         ESCAPE_STRINGS.put("&egrave", '\u00E8');
133         ESCAPE_STRINGS.put("&eacute", '\u00E9');
134         ESCAPE_STRINGS.put("&ecirc", '\u00EA');
135         ESCAPE_STRINGS.put("&euml", '\u00EB');
136         ESCAPE_STRINGS.put("&igrave", '\u00EC');
137         ESCAPE_STRINGS.put("&iacute", '\u00ED');
138         ESCAPE_STRINGS.put("&icirc", '\u00EE');
139         ESCAPE_STRINGS.put("&iuml", '\u00EF');
140         ESCAPE_STRINGS.put("&eth", '\u00F0');
141         ESCAPE_STRINGS.put("&ntilde", '\u00F1');
142         ESCAPE_STRINGS.put("&ograve", '\u00F2');
143         ESCAPE_STRINGS.put("&oacute", '\u00F3');
144         ESCAPE_STRINGS.put("&ocirc", '\u00F4');
145         ESCAPE_STRINGS.put("&otilde", '\u00F5');
146         ESCAPE_STRINGS.put("&ouml", '\u00F6');
147         ESCAPE_STRINGS.put("&divide", '\u00F7');
148         ESCAPE_STRINGS.put("&oslash", '\u00F8');
149         ESCAPE_STRINGS.put("&ugrave", '\u00F9');
150         ESCAPE_STRINGS.put("&uacute", '\u00FA');
151         ESCAPE_STRINGS.put("&ucirc", '\u00FB');
152         ESCAPE_STRINGS.put("&uuml", '\u00FC');
153         ESCAPE_STRINGS.put("&yacute", '\u00FD');
154         ESCAPE_STRINGS.put("&thorn", '\u00FE');
155         ESCAPE_STRINGS.put("&yuml", '\u00FF');
156         ESCAPE_STRINGS.put("&fnof", '\u0192');
157         ESCAPE_STRINGS.put("&Alpha", '\u0391');
158         ESCAPE_STRINGS.put("&Beta", '\u0392');
159         ESCAPE_STRINGS.put("&Gamma", '\u0393');
160         ESCAPE_STRINGS.put("&Delta", '\u0394');
161         ESCAPE_STRINGS.put("&Epsilon", '\u0395');
162         ESCAPE_STRINGS.put("&Zeta", '\u0396');
163         ESCAPE_STRINGS.put("&Eta", '\u0397');
164         ESCAPE_STRINGS.put("&Theta", '\u0398');
165         ESCAPE_STRINGS.put("&Iota", '\u0399');
166         ESCAPE_STRINGS.put("&Kappa", '\u039A');
167         ESCAPE_STRINGS.put("&Lambda", '\u039B');
168         ESCAPE_STRINGS.put("&Mu", '\u039C');
169         ESCAPE_STRINGS.put("&Nu", '\u039D');
170         ESCAPE_STRINGS.put("&Xi", '\u039E');
171         ESCAPE_STRINGS.put("&Omicron", '\u039F');
172         ESCAPE_STRINGS.put("&Pi", '\u03A0');
173         ESCAPE_STRINGS.put("&Rho", '\u03A1');
174         ESCAPE_STRINGS.put("&Sigma", '\u03A3');
175         ESCAPE_STRINGS.put("&Tau", '\u03A4');
176         ESCAPE_STRINGS.put("&Upsilon", '\u03A5');
177         ESCAPE_STRINGS.put("&Phi", '\u03A6');
178         ESCAPE_STRINGS.put("&Chi", '\u03A7');
179         ESCAPE_STRINGS.put("&Psi", '\u03A8');
180         ESCAPE_STRINGS.put("&Omega", '\u03A9');
181         ESCAPE_STRINGS.put("&alpha", '\u03B1');
182         ESCAPE_STRINGS.put("&beta", '\u03B2');
183         ESCAPE_STRINGS.put("&gamma", '\u03B3');
184         ESCAPE_STRINGS.put("&delta", '\u03B4');
185         ESCAPE_STRINGS.put("&epsilon", '\u03B5');
186         ESCAPE_STRINGS.put("&zeta", '\u03B6');
187         ESCAPE_STRINGS.put("&eta", '\u03B7');
188         ESCAPE_STRINGS.put("&theta", '\u03B8');
189         ESCAPE_STRINGS.put("&iota", '\u03B9');
190         ESCAPE_STRINGS.put("&kappa", '\u03BA');
191         ESCAPE_STRINGS.put("&lambda", '\u03BB');
192         ESCAPE_STRINGS.put("&mu", '\u03BC');
193         ESCAPE_STRINGS.put("&nu", '\u03BD');
194         ESCAPE_STRINGS.put("&xi", '\u03BE');
195         ESCAPE_STRINGS.put("&omicron", '\u03BF');
196         ESCAPE_STRINGS.put("&pi", '\u03C0');
197         ESCAPE_STRINGS.put("&rho", '\u03C1');
198         ESCAPE_STRINGS.put("&sigmaf", '\u03C2');
199         ESCAPE_STRINGS.put("&sigma", '\u03C3');
200         ESCAPE_STRINGS.put("&tau", '\u03C4');
201         ESCAPE_STRINGS.put("&upsilon", '\u03C5');
202         ESCAPE_STRINGS.put("&phi", '\u03C6');
203         ESCAPE_STRINGS.put("&chi", '\u03C7');
204         ESCAPE_STRINGS.put("&psi", '\u03C8');
205         ESCAPE_STRINGS.put("&omega", '\u03C9');
206         ESCAPE_STRINGS.put("&thetasym", '\u03D1');
207         ESCAPE_STRINGS.put("&upsih", '\u03D2');
208         ESCAPE_STRINGS.put("&piv", '\u03D6');
209         ESCAPE_STRINGS.put("&bull", '\u2022');
210         ESCAPE_STRINGS.put("&hellip", '\u2026');
211         ESCAPE_STRINGS.put("&prime", '\u2032');
212         ESCAPE_STRINGS.put("&Prime", '\u2033');
213         ESCAPE_STRINGS.put("&oline", '\u203E');
214         ESCAPE_STRINGS.put("&frasl", '\u2044');
215         ESCAPE_STRINGS.put("&weierp", '\u2118');
216         ESCAPE_STRINGS.put("&image", '\u2111');
217         ESCAPE_STRINGS.put("&real", '\u211C');
218         ESCAPE_STRINGS.put("&trade", '\u2122');
219         ESCAPE_STRINGS.put("&alefsym", '\u2135');
220         ESCAPE_STRINGS.put("&larr", '\u2190');
221         ESCAPE_STRINGS.put("&uarr", '\u2191');
222         ESCAPE_STRINGS.put("&rarr", '\u2192');
223         ESCAPE_STRINGS.put("&darr", '\u2193');
224         ESCAPE_STRINGS.put("&harr", '\u2194');
225         ESCAPE_STRINGS.put("&crarr", '\u21B5');
226         ESCAPE_STRINGS.put("&lArr", '\u21D0');
227         ESCAPE_STRINGS.put("&uArr", '\u21D1');
228         ESCAPE_STRINGS.put("&rArr", '\u21D2');
229         ESCAPE_STRINGS.put("&dArr", '\u21D3');
230         ESCAPE_STRINGS.put("&hArr", '\u21D4');
231         ESCAPE_STRINGS.put("&forall", '\u2200');
232         ESCAPE_STRINGS.put("&part", '\u2202');
233         ESCAPE_STRINGS.put("&exist", '\u2203');
234         ESCAPE_STRINGS.put("&empty", '\u2205');
235         ESCAPE_STRINGS.put("&nabla", '\u2207');
236         ESCAPE_STRINGS.put("&isin", '\u2208');
237         ESCAPE_STRINGS.put("&notin", '\u2209');
238         ESCAPE_STRINGS.put("&ni", '\u220B');
239         ESCAPE_STRINGS.put("&prod", '\u220F');
240         ESCAPE_STRINGS.put("&sum", '\u2211');
241         ESCAPE_STRINGS.put("&minus", '\u2212');
242         ESCAPE_STRINGS.put("&lowast", '\u2217');
243         ESCAPE_STRINGS.put("&radic", '\u221A');
244         ESCAPE_STRINGS.put("&prop", '\u221D');
245         ESCAPE_STRINGS.put("&infin", '\u221E');
246         ESCAPE_STRINGS.put("&ang", '\u2220');
247         ESCAPE_STRINGS.put("&and", '\u2227');
248         ESCAPE_STRINGS.put("&or", '\u2228');
249         ESCAPE_STRINGS.put("&cap", '\u2229');
250         ESCAPE_STRINGS.put("&cup", '\u222A');
251         ESCAPE_STRINGS.put("&int", '\u222B');
252         ESCAPE_STRINGS.put("&there4", '\u2234');
253         ESCAPE_STRINGS.put("&sim", '\u223C');
254         ESCAPE_STRINGS.put("&cong", '\u2245');
255         ESCAPE_STRINGS.put("&asymp", '\u2248');
256         ESCAPE_STRINGS.put("&ne", '\u2260');
257         ESCAPE_STRINGS.put("&equiv", '\u2261');
258         ESCAPE_STRINGS.put("&le", '\u2264');
259         ESCAPE_STRINGS.put("&ge", '\u2265');
260         ESCAPE_STRINGS.put("&sub", '\u2282');
261         ESCAPE_STRINGS.put("&sup", '\u2283');
262         ESCAPE_STRINGS.put("&nsub", '\u2284');
263         ESCAPE_STRINGS.put("&sube", '\u2286');
264         ESCAPE_STRINGS.put("&supe", '\u2287');
265         ESCAPE_STRINGS.put("&oplus", '\u2295');
266         ESCAPE_STRINGS.put("&otimes", '\u2297');
267         ESCAPE_STRINGS.put("&perp", '\u22A5');
268         ESCAPE_STRINGS.put("&sdot", '\u22C5');
269         ESCAPE_STRINGS.put("&lceil", '\u2308');
270         ESCAPE_STRINGS.put("&rceil", '\u2309');
271         ESCAPE_STRINGS.put("&lfloor", '\u230A');
272         ESCAPE_STRINGS.put("&rfloor", '\u230B');
273         ESCAPE_STRINGS.put("&lang", '\u2329');
274         ESCAPE_STRINGS.put("&rang", '\u232A');
275         ESCAPE_STRINGS.put("&loz", '\u25CA');
276         ESCAPE_STRINGS.put("&spades", '\u2660');
277         ESCAPE_STRINGS.put("&clubs", '\u2663');
278         ESCAPE_STRINGS.put("&hearts", '\u2665');
279         ESCAPE_STRINGS.put("&diams", '\u2666');
280         ESCAPE_STRINGS.put("&quot", '\u0022');
281         ESCAPE_STRINGS.put("&amp", '\u0026');
282         ESCAPE_STRINGS.put("&lt", '\u003C');
283         ESCAPE_STRINGS.put("&gt", '\u003E');
284         ESCAPE_STRINGS.put("&OElig", '\u0152');
285         ESCAPE_STRINGS.put("&oelig", '\u0153');
286         ESCAPE_STRINGS.put("&Scaron", '\u0160');
287         ESCAPE_STRINGS.put("&scaron", '\u0161');
288         ESCAPE_STRINGS.put("&Yuml", '\u0178');
289         ESCAPE_STRINGS.put("&circ", '\u02C6');
290         ESCAPE_STRINGS.put("&tilde", '\u02DC');
291         ESCAPE_STRINGS.put("&ensp", '\u2002');
292         ESCAPE_STRINGS.put("&emsp", '\u2003');
293         ESCAPE_STRINGS.put("&thinsp", '\u2009');
294         ESCAPE_STRINGS.put("&zwnj", '\u200C');
295         ESCAPE_STRINGS.put("&zwj", '\u200D');
296         ESCAPE_STRINGS.put("&lrm", '\u200E');
297         ESCAPE_STRINGS.put("&rlm", '\u200F');
298         ESCAPE_STRINGS.put("&ndash", '\u2013');
299         ESCAPE_STRINGS.put("&mdash", '\u2014');
300         ESCAPE_STRINGS.put("&lsquo", '\u2018');
301         ESCAPE_STRINGS.put("&rsquo", '\u2019');
302         ESCAPE_STRINGS.put("&sbquo", '\u201A');
303         ESCAPE_STRINGS.put("&ldquo", '\u201C');
304         ESCAPE_STRINGS.put("&rdquo", '\u201D');
305         ESCAPE_STRINGS.put("&bdquo", '\u201E');
306         ESCAPE_STRINGS.put("&dagger", '\u2020');
307         ESCAPE_STRINGS.put("&Dagger", '\u2021');
308         ESCAPE_STRINGS.put("&permil", '\u2030');
309         ESCAPE_STRINGS.put("&lsaquo", '\u2039');
310         ESCAPE_STRINGS.put("&rsaquo", '\u203A');
311         ESCAPE_STRINGS.put("&euro", '\u20AC');
312     }
313 
314     /**
315      * Code to generate a short 'snippet' from either plain text or html text
316      *
317      * If the sync protocol can get plain text, that's great, but we'll still strip out extraneous
318      * whitespace.  If it's HTML, we'll 1) strip out tags, 2) turn entities into the appropriate
319      * characters, and 3) strip out extraneous whitespace, all in one pass
320      *
321      * Why not use an existing class?  The best answer is performance; yet another answer is
322      * correctness (e.g. Html.textFromHtml simply doesn't generate well-stripped text).  But
323      * performance is key; we frequently sync text that is 10K or (much) longer, yet we really only
324      * care about a small amount of text for the snippet.  So it's critically important that we just
325      * stop when we've gotten enough; existing methods that exist will go through the entire
326      * incoming string, at great (and useless, in this case) expense.
327      */
328 
makeSnippetFromHtmlText(String text)329     public static String makeSnippetFromHtmlText(String text) {
330         return makeSnippetFromText(text, true);
331     }
332 
makeSnippetFromPlainText(String text)333     public static String makeSnippetFromPlainText(String text) {
334         return makeSnippetFromText(text, false);
335     }
336 
337     /**
338      * Find the end of this tag; there are two alternatives: <tag .../> or <tag ...> ... </tag>
339      * @param htmlText some HTML text
340      * @param tag the HTML tag
341      * @param startPos the start position in the HTML text where the tag starts
342      * @return the position just before the end of the tag or -1 if not found
343      */
findTagEnd(String htmlText, String tag, int startPos)344     /*package*/ static int findTagEnd(String htmlText, String tag, int startPos) {
345         if (tag.endsWith(" ")) {
346             tag = tag.substring(0, tag.length() - 1);
347         }
348         int length = htmlText.length();
349         char prevChar = 0;
350         for (int i = startPos; i < length; i++) {
351             char c = htmlText.charAt(i);
352             if (c == '>') {
353                if (prevChar == '/') {
354                    return i - 1;
355                }
356                break;
357             }
358             prevChar = c;
359         }
360         // We didn't find /> at the end of the tag so find </tag>
361         return htmlText.indexOf("/" + tag, startPos);
362     }
363 
makeSnippetFromText(String text, boolean stripHtml)364     public static String makeSnippetFromText(String text, boolean stripHtml) {
365         // Handle null and empty string
366         if (TextUtils.isEmpty(text)) return "";
367 
368         final int length = text.length();
369         // Use char[] instead of StringBuilder purely for performance; fewer method calls, etc.
370         char[] buffer = new char[MAX_SNIPPET_LENGTH];
371         // skipCount is an array of a single int; that int is set inside stripHtmlEntity and is
372         // used to determine how many characters can be "skipped" due to the transformation of the
373         // entity to a single character.  When Java allows multiple return values, we can make this
374         // much cleaner :-)
375         int[] skipCount = new int[1];
376         int bufferCount = 0;
377         // Start with space as last character to avoid leading whitespace
378         char last = ' ';
379         // Indicates whether we're in the middle of an HTML tag
380         boolean inTag = false;
381 
382         // Walk through the text until we're done with the input OR we've got a large enough snippet
383         for (int i = 0; i < length && bufferCount < MAX_SNIPPET_LENGTH; i++) {
384             char c = text.charAt(i);
385             if (stripHtml && !inTag && (c == '<')) {
386                 // Find tags to strip; they will begin with <! or !- or </ or <letter
387                 if (i < (length - 1)) {
388                     char peek = text.charAt(i + 1);
389                     if (peek == '!' || peek == '-' || peek == '/' || Character.isLetter(peek)) {
390                         inTag = true;
391                         // Strip content of title, script, style and applet tags
392                         if (i < (length - (MAX_STRIP_TAG_LENGTH + 2))) {
393                             String tag = text.substring(i + 1, i + MAX_STRIP_TAG_LENGTH + 1);
394                             String tagLowerCase = tag.toLowerCase();
395                             boolean stripContent = false;
396                             for (String stripTag: STRIP_TAGS) {
397                                 if (tagLowerCase.startsWith(stripTag)) {
398                                     stripContent = true;
399                                     tag = tag.substring(0, stripTag.length());
400                                     break;
401                                 }
402                             }
403                             if (stripContent) {
404                                 // Look for the end of this tag
405                                 int endTagPosition = findTagEnd(text, tag, i);
406                                 if (endTagPosition < 0) {
407                                     break;
408                                 } else {
409                                     i = endTagPosition;
410                                 }
411                             }
412                         }
413                     }
414                 }
415             } else if (stripHtml && inTag && (c == '>')) {
416                 // Terminate stripping here
417                 inTag = false;
418                 continue;
419             }
420 
421             if (inTag) {
422                 // We just skip by everything while we're in a tag
423                 continue;
424             } else if (stripHtml && (c == '&')) {
425                 // Handle a possible HTML entity here
426                 // We always get back a character to use; we also get back a "skip count",
427                 // indicating how many characters were eaten from the entity
428                 c = stripHtmlEntity(text, i, skipCount);
429                 i += skipCount[0];
430             }
431 
432             if (Character.isWhitespace(c) || (c == NON_BREAKING_SPACE_CHARACTER)) {
433                 // The idea is to find the content in the message, not the whitespace, so we'll
434                 // turn any combination of contiguous whitespace into a single space
435                 if (last == ' ') {
436                     continue;
437                 } else {
438                     // Make every whitespace character a simple space
439                     c = ' ';
440                 }
441             } else if ((c == '-' || c == '=') && (last == c)) {
442                 // Lots of messages (especially digests) have whole lines of --- or ===
443                 // We'll get rid of those duplicates here
444                 continue;
445             }
446 
447             // After all that, maybe we've got a character for our snippet
448             buffer[bufferCount++] = c;
449             last = c;
450         }
451 
452         // Lose trailing space and return our snippet
453         if ((bufferCount > 0) && (last == ' ')) {
454             bufferCount--;
455         }
456         return new String(buffer, 0, bufferCount);
457     }
458 
stripHtmlEntity(String text, int pos, int[] skipCount)459     static /*package*/ char stripHtmlEntity(String text, int pos, int[] skipCount) {
460         int length = text.length();
461         // Ugly, but we store our skip count in this array; we can't use a static here, because
462         // multiple threads might be calling in
463         skipCount[0] = 0;
464         // All entities are <= 8 characters long, so that's how far we'll look for one (+ & and ;)
465         int end = pos + 10;
466         String entity = null;
467         // Isolate the entity
468         for (int i = pos; (i < length) && (i < end); i++) {
469             if (text.charAt(i) == ';') {
470                 entity = text.substring(pos, i);
471                 break;
472             }
473         }
474         if (entity == null) {
475             // This wasn't really an HTML entity
476             return '&';
477         } else {
478             // Skip count is the length of the entity
479             Character mapping = ESCAPE_STRINGS.get(entity);
480             int entityLength = entity.length();
481             if (mapping != null) {
482                 skipCount[0] = entityLength;
483                 return mapping;
484             } else if ((entityLength > 2) && (entity.charAt(1) == '#')) {
485                 // &#nn; means ascii nn (decimal) and &#xnn means ascii nn (hex)
486                 char c = '?';
487                 try {
488                     int i;
489                     if ((entity.charAt(2) == 'x') && (entityLength > 3)) {
490                         i = Integer.parseInt(entity.substring(3), 16);
491                     } else {
492                         i = Integer.parseInt(entity.substring(2));
493                     }
494                     c = (char)i;
495                 } catch (NumberFormatException e) {
496                     // We'll just return the ? in this case
497                 }
498                 skipCount[0] = entityLength;
499                 return c;
500             }
501         }
502         // Worst case, we return the original start character, ampersand
503         return '&';
504     }
505 
506     /**
507      * Given a string of HTML text and a query containing any number of search terms, returns
508      * an HTML string in which those search terms are highlighted (intended for use in a WebView)
509      *
510      * @param text the HTML text to process
511      * @param query the search terms
512      * @return HTML text with the search terms highlighted
513      */
514     @VisibleForTesting
highlightTermsInHtml(String text, String query)515     public static String highlightTermsInHtml(String text, String query) {
516         try {
517             return highlightTerms(text, query, true).toString();
518         } catch (IOException e) {
519             // Can't happen, but we must catch this
520             return text;
521         }
522     }
523 
524     /**
525      * Given a string of plain text and a query containing any number of search terms, returns
526      * a CharSequence in which those search terms are highlighted (intended for use in a TextView)
527      *
528      * @param text the text to process
529      * @param query the search terms
530      * @return a CharSequence with the search terms highlighted
531      */
highlightTermsInText(String text, String query)532     public static CharSequence highlightTermsInText(String text, String query) {
533         try {
534             return highlightTerms(text, query, false);
535         } catch (IOException e) {
536             // Can't happen, but we must catch this
537             return text;
538         }
539     }
540 
541     static class SearchTerm {
542         final String mTerm;
543         final String mTermLowerCase;
544         final int mLength;
545         int mMatchLength = 0;
546         int mMatchStart = -1;
547 
SearchTerm(String term)548         SearchTerm(String term) {
549             mTerm = term;
550             mTermLowerCase = term.toLowerCase();
551             mLength = term.length();
552         }
553     }
554 
555     /**
556      * Generate a version of the incoming text in which all search terms in a query are highlighted.
557      * If the input is HTML, we return a StringBuilder with additional markup as required
558      * If the input is text, we return a SpannableStringBuilder with additional spans as required
559      *
560      * @param text the text to be processed
561      * @param query the query, which can contain multiple terms separated by whitespace
562      * @param html whether or not the text to be processed is HTML
563      * @return highlighted text
564      *
565      * @throws IOException as Appendable requires this
566      */
highlightTerms(String text, String query, boolean html)567     public static CharSequence highlightTerms(String text, String query, boolean html)
568             throws IOException {
569         // Handle null and empty string
570         if (TextUtils.isEmpty(text)) return "";
571         final int length = text.length();
572 
573         // Break up the query into search terms
574         ArrayList<SearchTerm> terms = new ArrayList<SearchTerm>();
575         if (query != null) {
576             StringTokenizer st = new StringTokenizer(query);
577             while (st.hasMoreTokens()) {
578                 terms.add(new SearchTerm(st.nextToken()));
579             }
580         }
581 
582         // Our appendable depends on whether we're building HTML text (for webview) or spannable
583         // text (for UI)
584         final Appendable sb = html ? new StringBuilder() : new SpannableStringBuilder();
585         // Indicates whether we're in the middle of an HTML tag
586         boolean inTag = false;
587         // The position of the last input character copied to output
588         int lastOut = -1;
589 
590         // Walk through the text until we're done with the input
591         // Just copy any HTML tags directly into the output; search for terms in the remaining text
592         for (int i = 0; i < length; i++) {
593             char chr = text.charAt(i);
594             if (html) {
595                 if (!inTag && (chr == '<')) {
596                     // Find tags; they will begin with <! or !- or </ or <letter
597                     if (i < (length - 1)) {
598                         char peek = text.charAt(i + 1);
599                         if (peek == '!' || peek == '-' || peek == '/' || Character.isLetter(peek)) {
600                             inTag = true;
601                             // Skip content of title, script, style and applet tags
602                             if (i < (length - (MAX_STRIP_TAG_LENGTH + 2))) {
603                                 String tag = text.substring(i + 1, i + MAX_STRIP_TAG_LENGTH + 1);
604                                 String tagLowerCase = tag.toLowerCase();
605                                 boolean stripContent = false;
606                                 for (String stripTag: STRIP_TAGS) {
607                                     if (tagLowerCase.startsWith(stripTag)) {
608                                         stripContent = true;
609                                         tag = tag.substring(0, stripTag.length());
610                                         break;
611                                     }
612                                 }
613                                 if (stripContent) {
614                                     // Look for the end of this tag
615                                     int endTagPosition = findTagEnd(text, tag, i);
616                                     if (endTagPosition < 0) {
617                                         sb.append(text.substring(i));
618                                         break;
619                                     } else {
620                                         sb.append(text.substring(i, endTagPosition - 1));
621                                         i = endTagPosition - 1;
622                                         chr = text.charAt(i);
623                                     }
624                                 }
625                             }
626                         }
627                     }
628                 } else if (inTag && (chr == '>')) {
629                     inTag = false;
630                 }
631 
632                 if (inTag) {
633                     sb.append(chr);
634                     continue;
635                 }
636             }
637 
638             // After all that, we've got some "body" text
639             char chrLowerCase = Character.toLowerCase(chr);
640             // Whether or not the current character should be appended to the output; we inhibit
641             // this while any search terms match
642             boolean appendNow = true;
643             // Look through search terms for matches
644             for (SearchTerm t: terms) {
645                 if (chrLowerCase == t.mTermLowerCase.charAt(t.mMatchLength)) {
646                     if (t.mMatchLength++ == 0) {
647                         // New match start
648                         t.mMatchStart = i;
649                     }
650                     if (t.mMatchLength == t.mLength) {
651                         String matchText = text.substring(t.mMatchStart, t.mMatchStart + t.mLength);
652                         // Completed match; add highlight and reset term
653                         if (t.mMatchStart <= lastOut) {
654                             matchText = text.substring(lastOut + 1, i + 1);
655                         }
656                         /*else*/
657                         if (matchText.length() == 0) {} else
658                         if (html) {
659                             sb.append("<span style=\"background-color: " + HIGHLIGHT_COLOR_STRING +
660                                     "\">");
661                             sb.append(matchText);
662                             sb.append("</span>");
663                         } else {
664                             SpannableString highlightSpan = new SpannableString(matchText);
665                             highlightSpan.setSpan(new BackgroundColorSpan(HIGHLIGHT_COLOR_INT), 0,
666                                     highlightSpan.length(), Spannable.SPAN_EXCLUSIVE_EXCLUSIVE);
667                             sb.append(highlightSpan);
668                         }
669                         lastOut = t.mMatchStart + t.mLength - 1;
670                         t.mMatchLength = 0;
671                         t.mMatchStart = -1;
672                     }
673                     appendNow = false;
674                 } else {
675                     if (t.mMatchStart >= 0) {
676                         // We're no longer matching; check for other matches in progress
677                         int leastOtherStart = -1;
678                         for (SearchTerm ot: terms) {
679                             // Save away the lowest match start for other search terms
680                             if ((ot != t) && (ot.mMatchStart >= 0) && ((leastOtherStart < 0) ||
681                                     (ot.mMatchStart <= leastOtherStart))) {
682                                 leastOtherStart = ot.mMatchStart;
683                             }
684                         }
685                         int matchEnd = t.mMatchStart + t.mMatchLength;
686                         if (leastOtherStart < 0 || leastOtherStart > matchEnd) {
687                             // Append the whole thing
688                             if (t.mMatchStart > lastOut) {
689                                 sb.append(text.substring(t.mMatchStart, matchEnd));
690                                 lastOut = matchEnd;
691                             }
692                         } else if (leastOtherStart == t.mMatchStart) {
693                             // Ok to append the current char
694                         } else if (leastOtherStart < t.mMatchStart) {
695                             // We're already covered by another search term, so don't append
696                             appendNow = false;
697                         } else if (t.mMatchStart > lastOut) {
698                             // Append the piece of our term that's not already covered
699                             sb.append(text.substring(t.mMatchStart, leastOtherStart));
700                             lastOut = leastOtherStart;
701                         }
702                     }
703                     // Reset this term
704                     t.mMatchLength = 0;
705                     t.mMatchStart = -1;
706                 }
707             }
708 
709             if (appendNow) {
710                 sb.append(chr);
711                 lastOut = i;
712             }
713         }
714 
715         return (CharSequence)sb;
716    }
717 
718     /**
719      * Determine whether two Strings (either of which might be null) are the same; this is true
720      * when both are null or both are Strings that are equal.
721      */
stringOrNullEquals(String a, String b)722     public static boolean stringOrNullEquals(String a, String b) {
723         if (a == null && b == null) return true;
724         if (a != null && b != null && a.equals(b)) return true;
725         return false;
726     }
727 
728 }
729