Mercurial > jhg

comparison src/org/tmatesoft/hg/internal/ConfigFileParser.java @ 497:02140be396d5

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Issue 38. Towards gentle handling of config files - parse them and keep every possible user change
author Artem Tikhomirov <tikhomirov.artem@gmail.com>
date Thu, 25 Oct 2012 19:59:08 +0200
parents
children 0205a5c4566b
comparison
equal deleted inserted replaced
496:c1c8f6859d3f 497:02140be396d5
1 /*
2 * Copyright (c) 2012 TMate Software Ltd
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; version 2 of the License.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11 * GNU General Public License for more details.
12 *
13 * For information on how to redistribute this software under
14 * the terms of a license other than GNU General Public License
15 * contact TMate Software at support@hg4j.com
16 */
17 package org.tmatesoft.hg.internal;
18
19 import java.io.ByteArrayInputStream;
20 import java.io.ByteArrayOutputStream;
21 import java.io.IOException;
22 import java.io.InputStream;
23 import java.io.OutputStream;
24 import java.util.ArrayList;
25 import java.util.Collections;
26 import java.util.HashSet;
27 import java.util.Iterator;
28 import java.util.LinkedHashMap;
29 import java.util.LinkedHashSet;
30 import java.util.List;
31
32 /**
33 * Simplistic parser to allow altering configuration files without touching user modifications/formatting/comments
34 *
35 * @author Artem Tikhomirov
36 * @author TMate Software Ltd.
37 */
38 public class ConfigFileParser {
39 private enum ParseState {Initial, Section, Entry};
40 private ParseState state = ParseState.Initial;
41 private int lastNonEmptyLineEndOffset = -1;
42 private String sectionName;
43 private int sectionStart = -1;
44 private String entryKey;
45 private int entryStart = -1;
46 private int valueStart = -1, valueEnd = -1;
47 private ArrayList<Entry> entries;
48 private ArrayList<Section> sections = new ArrayList<Section>();
49 private byte[] contents;
50
51 private List<String> deletions = new ArrayList<String>(5);
52 private List<String> additions = new ArrayList<String>(5), changes = new ArrayList<String>(5);
53
54
55 public boolean exists(String section, String key) {
56 assert contents != null;
57 for (Section s : sections) {
58 if (s.name.equals(section)) {
59 for (Entry e : s.entries) {
60 if (e.name.equals(key)) {
61 return true;
62 }
63 }
64 return false;
65 }
66 }
67 return false;
68 }
69
70 public void add(String section, String key, String newValue) {
71 additions.add(section);
72 additions.add(key);
73 additions.add(newValue);
74 }
75
76 public void change(String section, String key, String newValue) {
77 changes.add(section);
78 changes.add(key);
79 changes.add(newValue);
80 }
81
82 public void delete(String section, String key) {
83 deletions.add(section);
84 deletions.add(key);
85 }
86
87 public void parse(InputStream is) throws IOException {
88 state = ParseState.Initial;
89 sections.clear();
90 contents = null;
91 ByteArrayOutputStream bos = new ByteArrayOutputStream(1024);
92 ByteArrayOutputStream line = new ByteArrayOutputStream(80);
93 int offset = 0;
94 int lineOffset = -1;
95 int lineNumber = 1;
96 boolean crDetected = false; // true when previous char was \r
97 int b;
98 while ( (b = is.read()) != -1) {
99 bos.write(b);
100 if (b == '\n' || b == '\r') {
101 if (line.size() > 0) {
102 processLine(lineNumber, lineOffset, line.toByteArray());
103 line.reset();
104 lineOffset = -1;
105 lastNonEmptyLineEndOffset = bos.size() - 1; // offset points to EOL char
106 }
107 // else: XXX does empty line closes entry???
108 // when \n follows \r, increment line count only once
109 if (!(b == '\n' && crDetected)) {
110 lineNumber++;
111 }
112 crDetected = b == '\r';
113 } else {
114 crDetected = false;
115 if (line.size() == 0) {
116 lineOffset = offset;
117 }
118 line.write(b);
119 }
120 offset++;
121 }
122 // handle last line in case it's not EOL-terminated
123 if (line.size() > 0) {
124 processLine(lineNumber, lineOffset, line.toByteArray());
125 // might need it for #closeSection() below
126 lastNonEmptyLineEndOffset = bos.size();
127 }
128 if (state == ParseState.Entry) {
129 closeEntry();
130 }
131 if (state == ParseState.Section) {
132 closeSection();
133 }
134 contents = bos.toByteArray();
135 }
136
137 public void update(OutputStream out) throws IOException {
138 if (contents == null) {
139 throw new IOException("Shall parse first");
140 }
141 HashSet<String> processedSections = new HashSet<String>();
142 int contentsOffset = 0;
143 for (Section section : sections) {
144 LinkedHashMap<String,String> additionsInSection = new LinkedHashMap<String,String>();
145 LinkedHashMap<String,String> changesInSection = new LinkedHashMap<String,String>();
146 LinkedHashSet<String> deletionsInSection = new LinkedHashSet<String>();
147 if (!processedSections.contains(section.name)) {
148 for (Iterator<String> it = additions.iterator(); it.hasNext();) {
149 String s = it.next(), k = it.next(), v = it.next();
150 if (section.name.equals(s)) {
151 additionsInSection.put(k, v);
152 }
153 }
154 for (Iterator<String> it = changes.iterator(); it.hasNext();) {
155 String s = it.next(), k = it.next(), v = it.next();
156 if (section.name.equals(s)) {
157 changesInSection.put(k, v);
158 }
159 }
160 for (Iterator<String> it = deletions.iterator(); it.hasNext();) {
161 String s = it.next(), k = it.next();
162 if (section.name.equals(s)) {
163 deletionsInSection.add(k);
164 }
165 }
166 }
167 for (Entry e : section.entries) {
168 if (deletionsInSection.contains(e.name)) {
169 // write up to key start only
170 out.write(contents, contentsOffset, e.start - contentsOffset);
171 contentsOffset = e.valueEnd + 1;
172 } else if (changesInSection.containsKey(e.name)) {
173 if (e.valueStart == -1) {
174 // e.valueEnd determines insertion point
175 out.write(contents, contentsOffset, e.valueEnd + 1 - contentsOffset);
176 } else {
177 // e.valueEnd points to last character of the value
178 out.write(contents, contentsOffset, e.valueStart - contentsOffset);
179 }
180 String value = changesInSection.get(e.name);
181 out.write(value == null ? new byte[0] : value.getBytes());
182 contentsOffset = e.valueEnd + 1;
183 }
184 // else: keep contentsOffset to point to first uncopied character
185 }
186 if (section.entries.length == 0) {
187 // no entries, empty or only comments, perhaps.
188 // use end of last meaningful line (whether [section] or comment string),
189 // which points to newline character
190 out.write(contents, contentsOffset, section.end - contentsOffset);
191 contentsOffset = section.end;
192 // since it's tricky to track \n or \r\n with lastNonEmptyLineEndOffset,
193 // we copy up to the line delimiter and insert new lines, if any, with \n prepended,
194 // so that original EOL will be moved to the very end of the section.
195 // Indeed, would be better to insert *after* lastNonEmptyLineEndOffset,
196 // but I don't want to complicate #parse (if line.size() > 0 part) method.
197 // Hope, this won't make too much trouble (if any, at all -
198 // if String.format translates \n to system EOL, then nobody would notice)
199 }
200 if (!additionsInSection.isEmpty()) {
201 // make sure additions are written once everything else is there
202 out.write(contents, contentsOffset, section.end - contentsOffset);
203 contentsOffset = section.end;
204 for (String k : additionsInSection.keySet()) {
205 String v = additionsInSection.get(k);
206 out.write(String.format("\n%s = %s", k, v == null ? "" : v).getBytes());
207 }
208 }
209 // if section comes more than once, update only first one.
210 processedSections.add(section.name);
211 }
212 out.write(contents, contentsOffset, contents.length - contentsOffset);
213 }
214
215 private void processLine(int lineNumber, int offset, byte[] line) throws IOException {
216 int localOffset = 0, i = 0;
217 while (i < line.length && Character.isWhitespace(line[i])) {
218 i++;
219 }
220 if (i == line.length) {
221 return;
222 }
223 localOffset = i;
224 if (line[i] == '[') {
225 if (state == ParseState.Entry) {
226 closeEntry();
227 }
228 if (state == ParseState.Section) {
229 closeSection();
230 }
231
232 while (i < line.length && line[i] != ']') {
233 i++;
234 }
235 if (i == line.length) {
236 throw new IOException(String.format("Can't find closing ']' for section name in line %d", lineNumber));
237 }
238 sectionName = new String(line, localOffset+1, i-localOffset-1);
239 sectionStart = offset + localOffset;
240 state = ParseState.Section;
241 } else if (line[i] == '#' || line[i] == ';') {
242 // comment line, nothing to process
243 return;
244 } else {
245 // entry
246 if (state == ParseState.Initial) {
247 throw new IOException(String.format("Line %d doesn't belong to any section", lineNumber));
248 }
249 if (localOffset > 0) {
250 if (state == ParseState.Section) {
251 throw new IOException(String.format("Non-indented key is expected in line %d", lineNumber));
252 }
253 assert state == ParseState.Entry;
254 // whitespace-indented continuation of the previous entry
255 if (valueStart == -1) {
256 // value didn't start at the same line the key was found at
257 valueStart = offset + localOffset;
258 }
259 // value ends with eol (assumption is trailing comments are not allowed)
260 valueEnd = offset + line.length - 1;
261 } else {
262 if (state == ParseState.Entry) {
263 closeEntry();
264 }
265 assert state == ParseState.Section;
266 // it's a new entry
267 state = ParseState.Entry;
268 // get name of the entry
269 while (i < line.length && !Character.isWhitespace(line[i]) && line[i] != '=') {
270 i++;
271 }
272 if (i == line.length) {
273 throw new IOException(String.format("Can't process entry in line %d", lineNumber));
274 }
275 entryKey = new String(line, localOffset, i - localOffset);
276 entryStart = offset + localOffset;
277 // look for '=' after key name
278 while (i < line.length && line[i] != '=') {
279 i++;
280 }
281 if (i == line.length) {
282 throw new IOException(String.format("Can't find '=' after key %s in line %d", entryKey, lineNumber));
283 }
284 // skip whitespaces after '='
285 i++; // line[i] == '='
286 while (i < line.length && Character.isWhitespace(line[i])) {
287 i++;
288 }
289 // valueStart might be -1 in case no value is specified in the same line as key
290 // but valueEnd is always initialized just in case there's no next, value continuation line
291 if (i == line.length) {
292 valueStart = -1;
293 } else {
294 valueStart = offset + i;
295 }
296
297 // if trailing comments are allowed, shall
298 // look up comment char and set valueEnd to its position-1
299 valueEnd = offset + line.length - 1;
300 }
301 }
302 }
303
304 private void closeSection() {
305 assert state == ParseState.Section;
306 assert sectionName != null;
307 assert lastNonEmptyLineEndOffset != -1;
308 Section s = new Section(sectionName, sectionStart, lastNonEmptyLineEndOffset, entries == null ? Collections.<Entry>emptyList() : entries);
309 sections.add(s);
310 sectionName = null;
311 sectionStart = -1;
312 state = ParseState.Initial;
313 entries = null;
314 }
315
316 private void closeEntry() {
317 assert state == ParseState.Entry;
318 assert entryKey != null;
319 state = ParseState.Section;
320 Entry e = new Entry(entryKey, entryStart, valueStart, valueEnd);
321 if (entries == null) {
322 entries = new ArrayList<Entry>();
323 }
324 entries.add(e);
325 entryKey = null;
326 entryStart = valueStart = valueEnd -1;
327 }
328
329
330 private static class Block {
331 public final int start;
332 Block(int s) {
333 start = s;
334 }
335 }
336
337 private static class Entry extends Block {
338 public final int valueStart, valueEnd;
339 public final String name;
340
341 Entry(String n, int s, int vs, int ve) {
342 super(s);
343 name = n;
344 valueStart = vs;
345 valueEnd = ve;
346 }
347 }
348
349 private static class Section extends Block {
350 public final String name;
351 public final Entry[] entries;
352 public final int end;
353
354 Section(String n, int s, int endOffset, List<Entry> e) {
355 super(s);
356 name = n;
357 end = endOffset;
358 entries = new Entry[e.size()];
359 e.toArray(entries);
360 }
361 }
362
363 public static void main(String[] args) throws Exception {
364 ConfigFileParser p = new ConfigFileParser();
365 p.parse(new ByteArrayInputStream(xx.getBytes()));
366 System.out.println(">>>");
367 System.out.println(xx);
368 System.out.println("===");
369 p.add("sect1", "key5", "x");
370 ByteArrayOutputStream out = new ByteArrayOutputStream(xx.length());
371 p.update(out);
372 System.out.println(new String(out.toByteArray()));
373 /*
374 for (Section s : p.sections) {
375 System.out.printf("[%s@%d]\n", s.name, s.start);
376 for (Entry e : s.entries) {
377 System.out.printf("%s@%d = %d..%d\n", e.name, e.start, e.valueStart, e.valueEnd);
378 }
379 }
380 */
381 }
382 private static final String xx = "#comment1\n [sect1]\nkey = value #not a comment2\n#comment3\nkey2= \nkey3 = \n value1, #cc\n value2\nkey4 = v1,\n v2 \n ,v3\n\n\n[sect2]\nx = a";
383 }