1 /* 2 * Copyright (C) 2015, Matthias Sohn <matthias.sohn@sap.com> and others 3 * 4 * This program and the accompanying materials are made available under the 5 * terms of the Eclipse Distribution License v. 1.0 which is available at 6 * https://www.eclipse.org/org/documents/edl-v10.php. 7 * 8 * SPDX-License-Identifier: BSD-3-Clause 9 */ 10 11 package org.eclipse.jgit.lfs.lib; 12 13 import java.io.Serializable; 14 import java.text.MessageFormat; 15 16 import org.eclipse.jgit.lfs.errors.InvalidLongObjectIdException; 17 import org.eclipse.jgit.lfs.internal.LfsText; 18 import org.eclipse.jgit.util.NB; 19 import org.eclipse.jgit.util.RawParseUtils; 20 21 /** 22 * A prefix abbreviation of an {@link org.eclipse.jgit.lfs.lib.LongObjectId}. 23 * <p> 24 * Enable abbreviating SHA-256 strings used by Git LFS, using sufficient leading 25 * digits from the LongObjectId name to still be unique within the repository 26 * the string was generated from. These ids are likely to be unique for a useful 27 * period of time, especially if they contain at least 6-10 hex digits. 28 * <p> 29 * This class converts the hex string into a binary form, to make it more 30 * efficient for matching against an object. 31 * 32 * Ported to SHA-256 from {@link org.eclipse.jgit.lib.AbbreviatedObjectId} 33 * 34 * @since 4.3 35 */ 36 public final class AbbreviatedLongObjectId implements Serializable { 37 private static final long serialVersionUID = 1L; 38 39 /** 40 * Test a string of characters to verify it is a hex format. 41 * <p> 42 * If true the string can be parsed with {@link #fromString(String)}. 43 * 44 * @param id 45 * the string to test. 46 * @return true if the string can converted into an AbbreviatedObjectId. 47 */ isId(String id)48 public static final boolean isId(String id) { 49 if (id.length() < 2 50 || Constants.LONG_OBJECT_ID_STRING_LENGTH < id.length()) 51 return false; 52 try { 53 for (int i = 0; i < id.length(); i++) 54 RawParseUtils.parseHexInt4((byte) id.charAt(i)); 55 return true; 56 } catch (ArrayIndexOutOfBoundsException e) { 57 return false; 58 } 59 } 60 61 /** 62 * Convert an AbbreviatedObjectId from hex characters (US-ASCII). 63 * 64 * @param buf 65 * the US-ASCII buffer to read from. 66 * @param offset 67 * position to read the first character from. 68 * @param end 69 * one past the last position to read (<code>end-offset</code> is 70 * the length of the string). 71 * @return the converted object id. 72 */ fromString(final byte[] buf, final int offset, final int end)73 public static final AbbreviatedLongObjectId fromString(final byte[] buf, 74 final int offset, final int end) { 75 if (end - offset > Constants.LONG_OBJECT_ID_STRING_LENGTH) 76 throw new IllegalArgumentException(MessageFormat.format( 77 LfsText.get().invalidLongIdLength, 78 Integer.valueOf(end - offset), 79 Integer.valueOf(Constants.LONG_OBJECT_ID_STRING_LENGTH))); 80 return fromHexString(buf, offset, end); 81 } 82 83 /** 84 * Convert an AbbreviatedObjectId from an 85 * {@link org.eclipse.jgit.lib.AnyObjectId}. 86 * <p> 87 * This method copies over all bits of the Id, and is therefore complete 88 * (see {@link #isComplete()}). 89 * 90 * @param id 91 * the {@link org.eclipse.jgit.lib.ObjectId} to convert from. 92 * @return the converted object id. 93 */ fromLongObjectId( AnyLongObjectId id)94 public static final AbbreviatedLongObjectId fromLongObjectId( 95 AnyLongObjectId id) { 96 return new AbbreviatedLongObjectId( 97 Constants.LONG_OBJECT_ID_STRING_LENGTH, id.w1, id.w2, id.w3, 98 id.w4); 99 } 100 101 /** 102 * Convert an AbbreviatedLongObjectId from hex characters. 103 * 104 * @param str 105 * the string to read from. Must be <= 64 characters. 106 * @return the converted object id. 107 */ fromString(String str)108 public static final AbbreviatedLongObjectId fromString(String str) { 109 if (str.length() > Constants.LONG_OBJECT_ID_STRING_LENGTH) 110 throw new IllegalArgumentException( 111 MessageFormat.format(LfsText.get().invalidLongId, str)); 112 final byte[] b = org.eclipse.jgit.lib.Constants.encodeASCII(str); 113 return fromHexString(b, 0, b.length); 114 } 115 fromHexString(final byte[] bs, int ptr, final int end)116 private static final AbbreviatedLongObjectId fromHexString(final byte[] bs, 117 int ptr, final int end) { 118 try { 119 final long a = hexUInt64(bs, ptr, end); 120 final long b = hexUInt64(bs, ptr + 16, end); 121 final long c = hexUInt64(bs, ptr + 32, end); 122 final long d = hexUInt64(bs, ptr + 48, end); 123 return new AbbreviatedLongObjectId(end - ptr, a, b, c, d); 124 } catch (ArrayIndexOutOfBoundsException e) { 125 InvalidLongObjectIdException e1 = new InvalidLongObjectIdException( 126 bs, ptr, end - ptr); 127 e1.initCause(e); 128 throw e1; 129 } 130 } 131 hexUInt64(final byte[] bs, int p, final int end)132 private static final long hexUInt64(final byte[] bs, int p, final int end) { 133 if (16 <= end - p) 134 return RawParseUtils.parseHexInt64(bs, p); 135 136 long r = 0; 137 int n = 0; 138 while (n < 16 && p < end) { 139 r <<= 4; 140 r |= RawParseUtils.parseHexInt4(bs[p++]); 141 n++; 142 } 143 return r << ((16 - n) * 4); 144 } 145 mask(int nibbles, long word, long v)146 static long mask(int nibbles, long word, long v) { 147 final long b = (word - 1) * 16; 148 if (b + 16 <= nibbles) { 149 // We have all of the bits required for this word. 150 // 151 return v; 152 } 153 154 if (nibbles <= b) { 155 // We have none of the bits required for this word. 156 // 157 return 0; 158 } 159 160 final long s = 64 - (nibbles - b) * 4; 161 return (v >>> s) << s; 162 } 163 164 /** Number of half-bytes used by this id. */ 165 final int nibbles; 166 167 final long w1; 168 169 final long w2; 170 171 final long w3; 172 173 final long w4; 174 AbbreviatedLongObjectId(final int n, final long new_1, final long new_2, final long new_3, final long new_4)175 AbbreviatedLongObjectId(final int n, final long new_1, final long new_2, 176 final long new_3, final long new_4) { 177 nibbles = n; 178 w1 = new_1; 179 w2 = new_2; 180 w3 = new_3; 181 w4 = new_4; 182 } 183 184 /** 185 * Get length 186 * 187 * @return number of hex digits appearing in this id. 188 */ length()189 public int length() { 190 return nibbles; 191 } 192 193 /** 194 * Check if this id is complete 195 * 196 * @return true if this ObjectId is actually a complete id. 197 */ isComplete()198 public boolean isComplete() { 199 return length() == Constants.LONG_OBJECT_ID_STRING_LENGTH; 200 } 201 202 /** 203 * Convert to LongObjectId 204 * 205 * @return a complete ObjectId; null if {@link #isComplete()} is false. 206 */ toLongObjectId()207 public LongObjectId toLongObjectId() { 208 return isComplete() ? new LongObjectId(w1, w2, w3, w4) : null; 209 } 210 211 /** 212 * Compares this abbreviation to a full object id. 213 * 214 * @param other 215 * the other object id. 216 * @return <0 if this abbreviation names an object that is less than 217 * <code>other</code>; 0 if this abbreviation exactly matches the 218 * first {@link #length()} digits of <code>other.name()</code>; 219 * >0 if this abbreviation names an object that is after 220 * <code>other</code>. 221 */ prefixCompare(AnyLongObjectId other)222 public final int prefixCompare(AnyLongObjectId other) { 223 int cmp; 224 225 cmp = NB.compareUInt64(w1, mask(1, other.w1)); 226 if (cmp != 0) 227 return cmp; 228 229 cmp = NB.compareUInt64(w2, mask(2, other.w2)); 230 if (cmp != 0) 231 return cmp; 232 233 cmp = NB.compareUInt64(w3, mask(3, other.w3)); 234 if (cmp != 0) 235 return cmp; 236 237 return NB.compareUInt64(w4, mask(4, other.w4)); 238 } 239 240 /** 241 * Compare this abbreviation to a network-byte-order LongObjectId. 242 * 243 * @param bs 244 * array containing the other LongObjectId in network byte order. 245 * @param p 246 * position within {@code bs} to start the compare at. At least 247 * 32 bytes, starting at this position are required. 248 * @return <0 if this abbreviation names an object that is less than 249 * <code>other</code>; 0 if this abbreviation exactly matches the 250 * first {@link #length()} digits of <code>other.name()</code>; 251 * >0 if this abbreviation names an object that is after 252 * <code>other</code>. 253 */ prefixCompare(byte[] bs, int p)254 public final int prefixCompare(byte[] bs, int p) { 255 int cmp; 256 257 cmp = NB.compareUInt64(w1, mask(1, NB.decodeInt64(bs, p))); 258 if (cmp != 0) 259 return cmp; 260 261 cmp = NB.compareUInt64(w2, mask(2, NB.decodeInt64(bs, p + 8))); 262 if (cmp != 0) 263 return cmp; 264 265 cmp = NB.compareUInt64(w3, mask(3, NB.decodeInt64(bs, p + 16))); 266 if (cmp != 0) 267 return cmp; 268 269 return NB.compareUInt64(w4, mask(4, NB.decodeInt64(bs, p + 24))); 270 } 271 272 /** 273 * Compare this abbreviation to a network-byte-order LongObjectId. 274 * 275 * @param bs 276 * array containing the other LongObjectId in network byte order. 277 * @param p 278 * position within {@code bs} to start the compare at. At least 4 279 * longs, starting at this position are required. 280 * @return <0 if this abbreviation names an object that is less than 281 * <code>other</code>; 0 if this abbreviation exactly matches the 282 * first {@link #length()} digits of <code>other.name()</code>; 283 * >0 if this abbreviation names an object that is after 284 * <code>other</code>. 285 */ prefixCompare(long[] bs, int p)286 public final int prefixCompare(long[] bs, int p) { 287 int cmp; 288 289 cmp = NB.compareUInt64(w1, mask(1, bs[p])); 290 if (cmp != 0) 291 return cmp; 292 293 cmp = NB.compareUInt64(w2, mask(2, bs[p + 1])); 294 if (cmp != 0) 295 return cmp; 296 297 cmp = NB.compareUInt64(w3, mask(3, bs[p + 2])); 298 if (cmp != 0) 299 return cmp; 300 301 return NB.compareUInt64(w4, mask(4, bs[p + 3])); 302 } 303 304 /** 305 * Get the first byte of this id 306 * 307 * @return value for a fan-out style map, only valid of length >= 2. 308 */ getFirstByte()309 public final int getFirstByte() { 310 return (int) (w1 >>> 56); 311 } 312 mask(long word, long v)313 private long mask(long word, long v) { 314 return mask(nibbles, word, v); 315 } 316 317 /** {@inheritDoc} */ 318 @Override hashCode()319 public int hashCode() { 320 return (int) (w1 >> 32); 321 } 322 323 /** {@inheritDoc} */ 324 @Override equals(Object o)325 public boolean equals(Object o) { 326 if (o instanceof AbbreviatedLongObjectId) { 327 final AbbreviatedLongObjectId b = (AbbreviatedLongObjectId) o; 328 return nibbles == b.nibbles && w1 == b.w1 && w2 == b.w2 329 && w3 == b.w3 && w4 == b.w4; 330 } 331 return false; 332 } 333 334 /** 335 * <p>name.</p> 336 * 337 * @return string form of the abbreviation, in lower case hexadecimal. 338 */ name()339 public final String name() { 340 final char[] b = new char[Constants.LONG_OBJECT_ID_STRING_LENGTH]; 341 342 AnyLongObjectId.formatHexChar(b, 0, w1); 343 if (nibbles <= 16) 344 return new String(b, 0, nibbles); 345 346 AnyLongObjectId.formatHexChar(b, 16, w2); 347 if (nibbles <= 32) 348 return new String(b, 0, nibbles); 349 350 AnyLongObjectId.formatHexChar(b, 32, w3); 351 if (nibbles <= 48) 352 return new String(b, 0, nibbles); 353 354 AnyLongObjectId.formatHexChar(b, 48, w4); 355 return new String(b, 0, nibbles); 356 } 357 358 /** {@inheritDoc} */ 359 @SuppressWarnings("nls") 360 @Override toString()361 public String toString() { 362 return "AbbreviatedLongObjectId[" + name() + "]"; //$NON-NLS-1$ 363 } 364 } 365