001/* 002 * Licensed to the Apache Software Foundation (ASF) under one 003 * or more contributor license agreements. See the NOTICE file 004 * distributed with this work for additional information 005 * regarding copyright ownership. The ASF licenses this file 006 * to you under the Apache License, Version 2.0 (the 007 * "License"); you may not use this file except in compliance 008 * with the License. You may obtain a copy of the License at 009 * 010 * http://www.apache.org/licenses/LICENSE-2.0 011 * 012 * Unless required by applicable law or agreed to in writing, 013 * software distributed under the License is distributed on an 014 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY 015 * KIND, either express or implied. See the License for the 016 * specific language governing permissions and limitations 017 * under the License. 018 */ 019package org.apache.commons.compress.archivers.ar; 020 021import java.io.EOFException; 022import java.io.IOException; 023import java.io.InputStream; 024 025import org.apache.commons.compress.archivers.ArchiveEntry; 026import org.apache.commons.compress.archivers.ArchiveInputStream; 027import org.apache.commons.compress.utils.ArchiveUtils; 028import org.apache.commons.compress.utils.IOUtils; 029 030/** 031 * Implements the "ar" archive format as an input stream. 032 * 033 * @NotThreadSafe 034 * 035 */ 036public class ArArchiveInputStream extends ArchiveInputStream { 037 038 private final InputStream input; 039 private long offset = 0; 040 private boolean closed; 041 042 /* 043 * If getNextEnxtry has been called, the entry metadata is stored in 044 * currentEntry. 045 */ 046 private ArArchiveEntry currentEntry = null; 047 048 // Storage area for extra long names (GNU ar) 049 private byte[] namebuffer = null; 050 051 /* 052 * The offset where the current entry started. -1 if no entry has been 053 * called 054 */ 055 private long entryOffset = -1; 056 057 // cached buffers - must only be used locally in the class (COMPRESS-172 - reduce garbage collection) 058 private final byte[] nameBuf = new byte[16]; 059 private final byte[] lastModifiedBuf = new byte[12]; 060 private final byte[] idBuf = new byte[6]; 061 private final byte[] fileModeBuf = new byte[8]; 062 private final byte[] lengthBuf = new byte[10]; 063 064 /** 065 * Constructs an Ar input stream with the referenced stream 066 * 067 * @param pInput 068 * the ar input stream 069 */ 070 public ArArchiveInputStream(final InputStream pInput) { 071 input = pInput; 072 closed = false; 073 } 074 075 /** 076 * Returns the next AR entry in this stream. 077 * 078 * @return the next AR entry. 079 * @throws IOException 080 * if the entry could not be read 081 */ 082 public ArArchiveEntry getNextArEntry() throws IOException { 083 if (currentEntry != null) { 084 final long entryEnd = entryOffset + currentEntry.getLength(); 085 IOUtils.skip(this, entryEnd - offset); 086 currentEntry = null; 087 } 088 089 if (offset == 0) { 090 final byte[] expected = ArchiveUtils.toAsciiBytes(ArArchiveEntry.HEADER); 091 final byte[] realized = new byte[expected.length]; 092 final int read = IOUtils.readFully(this, realized); 093 if (read != expected.length) { 094 throw new IOException("failed to read header. Occured at byte: " + getBytesRead()); 095 } 096 for (int i = 0; i < expected.length; i++) { 097 if (expected[i] != realized[i]) { 098 throw new IOException("invalid header " + ArchiveUtils.toAsciiString(realized)); 099 } 100 } 101 } 102 103 if (offset % 2 != 0 && read() < 0) { 104 // hit eof 105 return null; 106 } 107 108 if (input.available() == 0) { 109 return null; 110 } 111 112 IOUtils.readFully(this, nameBuf); 113 IOUtils.readFully(this, lastModifiedBuf); 114 IOUtils.readFully(this, idBuf); 115 final int userId = asInt(idBuf, true); 116 IOUtils.readFully(this, idBuf); 117 IOUtils.readFully(this, fileModeBuf); 118 IOUtils.readFully(this, lengthBuf); 119 120 { 121 final byte[] expected = ArchiveUtils.toAsciiBytes(ArArchiveEntry.TRAILER); 122 final byte[] realized = new byte[expected.length]; 123 final int read = IOUtils.readFully(this, realized); 124 if (read != expected.length) { 125 throw new IOException("failed to read entry trailer. Occured at byte: " + getBytesRead()); 126 } 127 for (int i = 0; i < expected.length; i++) { 128 if (expected[i] != realized[i]) { 129 throw new IOException("invalid entry trailer. not read the content? Occured at byte: " + getBytesRead()); 130 } 131 } 132 } 133 134 entryOffset = offset; 135 136// GNU ar uses a '/' to mark the end of the filename; this allows for the use of spaces without the use of an extended filename. 137 138 // entry name is stored as ASCII string 139 String temp = ArchiveUtils.toAsciiString(nameBuf).trim(); 140 if (isGNUStringTable(temp)) { // GNU extended filenames entry 141 currentEntry = readGNUStringTable(lengthBuf); 142 return getNextArEntry(); 143 } 144 145 long len = asLong(lengthBuf); 146 if (temp.endsWith("/")) { // GNU terminator 147 temp = temp.substring(0, temp.length() - 1); 148 } else if (isGNULongName(temp)) { 149 final int off = Integer.parseInt(temp.substring(1));// get the offset 150 temp = getExtendedName(off); // convert to the long name 151 } else if (isBSDLongName(temp)) { 152 temp = getBSDLongName(temp); 153 // entry length contained the length of the file name in 154 // addition to the real length of the entry. 155 // assume file name was ASCII, there is no "standard" otherwise 156 final int nameLen = temp.length(); 157 len -= nameLen; 158 entryOffset += nameLen; 159 } 160 161 currentEntry = new ArArchiveEntry(temp, len, userId, 162 asInt(idBuf, true), 163 asInt(fileModeBuf, 8), 164 asLong(lastModifiedBuf)); 165 return currentEntry; 166 } 167 168 /** 169 * Get an extended name from the GNU extended name buffer. 170 * 171 * @param offset pointer to entry within the buffer 172 * @return the extended file name; without trailing "/" if present. 173 * @throws IOException if name not found or buffer not set up 174 */ 175 private String getExtendedName(final int offset) throws IOException { 176 if (namebuffer == null) { 177 throw new IOException("Cannot process GNU long filename as no // record was found"); 178 } 179 for (int i = offset; i < namebuffer.length; i++) { 180 if (namebuffer[i] == '\012' || namebuffer[i] == 0) { 181 if (namebuffer[i - 1] == '/') { 182 i--; // drop trailing / 183 } 184 return ArchiveUtils.toAsciiString(namebuffer, offset, i - offset); 185 } 186 } 187 throw new IOException("Failed to read entry: " + offset); 188 } 189 190 private long asLong(final byte[] byteArray) { 191 return Long.parseLong(ArchiveUtils.toAsciiString(byteArray).trim()); 192 } 193 194 private int asInt(final byte[] byteArray) { 195 return asInt(byteArray, 10, false); 196 } 197 198 private int asInt(final byte[] byteArray, final boolean treatBlankAsZero) { 199 return asInt(byteArray, 10, treatBlankAsZero); 200 } 201 202 private int asInt(final byte[] byteArray, final int base) { 203 return asInt(byteArray, base, false); 204 } 205 206 private int asInt(final byte[] byteArray, final int base, final boolean treatBlankAsZero) { 207 final String string = ArchiveUtils.toAsciiString(byteArray).trim(); 208 if (string.length() == 0 && treatBlankAsZero) { 209 return 0; 210 } 211 return Integer.parseInt(string, base); 212 } 213 214 /* 215 * (non-Javadoc) 216 * 217 * @see 218 * org.apache.commons.compress.archivers.ArchiveInputStream#getNextEntry() 219 */ 220 @Override 221 public ArchiveEntry getNextEntry() throws IOException { 222 return getNextArEntry(); 223 } 224 225 /* 226 * (non-Javadoc) 227 * 228 * @see java.io.InputStream#close() 229 */ 230 @Override 231 public void close() throws IOException { 232 if (!closed) { 233 closed = true; 234 input.close(); 235 } 236 currentEntry = null; 237 } 238 239 /* 240 * (non-Javadoc) 241 * 242 * @see java.io.InputStream#read(byte[], int, int) 243 */ 244 @Override 245 public int read(final byte[] b, final int off, final int len) throws IOException { 246 int toRead = len; 247 if (currentEntry != null) { 248 final long entryEnd = entryOffset + currentEntry.getLength(); 249 if (len > 0 && entryEnd > offset) { 250 toRead = (int) Math.min(len, entryEnd - offset); 251 } else { 252 return -1; 253 } 254 } 255 final int ret = this.input.read(b, off, toRead); 256 count(ret); 257 offset += ret > 0 ? ret : 0; 258 return ret; 259 } 260 261 /** 262 * Checks if the signature matches ASCII "!<arch>" followed by a single LF 263 * control character 264 * 265 * @param signature 266 * the bytes to check 267 * @param length 268 * the number of bytes to check 269 * @return true, if this stream is an Ar archive stream, false otherwise 270 */ 271 public static boolean matches(final byte[] signature, final int length) { 272 // 3c21 7261 6863 0a3e 273 274 return length >= 8 && signature[0] == 0x21 && 275 signature[1] == 0x3c && signature[2] == 0x61 && 276 signature[3] == 0x72 && signature[4] == 0x63 && 277 signature[5] == 0x68 && signature[6] == 0x3e && 278 signature[7] == 0x0a; 279 } 280 281 static final String BSD_LONGNAME_PREFIX = "#1/"; 282 private static final int BSD_LONGNAME_PREFIX_LEN = 283 BSD_LONGNAME_PREFIX.length(); 284 private static final String BSD_LONGNAME_PATTERN = 285 "^" + BSD_LONGNAME_PREFIX + "\\d+"; 286 287 /** 288 * Does the name look like it is a long name (or a name containing 289 * spaces) as encoded by BSD ar? 290 * 291 * <p>From the FreeBSD ar(5) man page:</p> 292 * <pre> 293 * BSD In the BSD variant, names that are shorter than 16 294 * characters and without embedded spaces are stored 295 * directly in this field. If a name has an embedded 296 * space, or if it is longer than 16 characters, then 297 * the string "#1/" followed by the decimal represen- 298 * tation of the length of the file name is placed in 299 * this field. The actual file name is stored immedi- 300 * ately after the archive header. The content of the 301 * archive member follows the file name. The ar_size 302 * field of the header (see below) will then hold the 303 * sum of the size of the file name and the size of 304 * the member. 305 * </pre> 306 * 307 * @since 1.3 308 */ 309 private static boolean isBSDLongName(final String name) { 310 return name != null && name.matches(BSD_LONGNAME_PATTERN); 311 } 312 313 /** 314 * Reads the real name from the current stream assuming the very 315 * first bytes to be read are the real file name. 316 * 317 * @see #isBSDLongName 318 * 319 * @since 1.3 320 */ 321 private String getBSDLongName(final String bsdLongName) throws IOException { 322 final int nameLen = 323 Integer.parseInt(bsdLongName.substring(BSD_LONGNAME_PREFIX_LEN)); 324 final byte[] name = new byte[nameLen]; 325 final int read = IOUtils.readFully(this, name); 326 if (read != nameLen) { 327 throw new EOFException(); 328 } 329 return ArchiveUtils.toAsciiString(name); 330 } 331 332 private static final String GNU_STRING_TABLE_NAME = "//"; 333 334 /** 335 * Is this the name of the "Archive String Table" as used by 336 * SVR4/GNU to store long file names? 337 * 338 * <p>GNU ar stores multiple extended filenames in the data section 339 * of a file with the name "//", this record is referred to by 340 * future headers.</p> 341 * 342 * <p>A header references an extended filename by storing a "/" 343 * followed by a decimal offset to the start of the filename in 344 * the extended filename data section.</p> 345 * 346 * <p>The format of the "//" file itself is simply a list of the 347 * long filenames, each separated by one or more LF 348 * characters. Note that the decimal offsets are number of 349 * characters, not line or string number within the "//" file.</p> 350 */ 351 private static boolean isGNUStringTable(final String name) { 352 return GNU_STRING_TABLE_NAME.equals(name); 353 } 354 355 /** 356 * Reads the GNU archive String Table. 357 * 358 * @see #isGNUStringTable 359 */ 360 private ArArchiveEntry readGNUStringTable(final byte[] length) throws IOException { 361 final int bufflen = asInt(length); // Assume length will fit in an int 362 namebuffer = new byte[bufflen]; 363 final int read = IOUtils.readFully(this, namebuffer, 0, bufflen); 364 if (read != bufflen){ 365 throw new IOException("Failed to read complete // record: expected=" 366 + bufflen + " read=" + read); 367 } 368 return new ArArchiveEntry(GNU_STRING_TABLE_NAME, bufflen); 369 } 370 371 private static final String GNU_LONGNAME_PATTERN = "^/\\d+"; 372 373 /** 374 * Does the name look like it is a long name (or a name containing 375 * spaces) as encoded by SVR4/GNU ar? 376 * 377 * @see #isGNUStringTable 378 */ 379 private boolean isGNULongName(final String name) { 380 return name != null && name.matches(GNU_LONGNAME_PATTERN); 381 } 382}