2 * Copyright (C) 2008, Shawn O. Pearce <spearce@spearce.org>
6 * Redistribution and use in source and binary forms, with or
7 * without modification, are permitted provided that the following
10 * - Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
13 * - Redistributions in binary form must reproduce the above
14 * copyright notice, this list of conditions and the following
15 * disclaimer in the documentation and/or other materials provided
16 * with the distribution.
18 * - Neither the name of the Git Development Community nor the
19 * names of its contributors may be used to endorse or promote
20 * products derived from this software without specific prior
23 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
24 * CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
25 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
26 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
28 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
29 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
30 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
31 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
32 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
33 * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
34 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
35 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
38 package org
.spearce
.jgit
.util
;
40 import static org
.spearce
.jgit
.lib
.ObjectChecker
.author
;
41 import static org
.spearce
.jgit
.lib
.ObjectChecker
.committer
;
42 import static org
.spearce
.jgit
.lib
.ObjectChecker
.encoding
;
44 import java
.nio
.ByteBuffer
;
45 import java
.nio
.charset
.CharacterCodingException
;
46 import java
.nio
.charset
.Charset
;
47 import java
.nio
.charset
.CharsetDecoder
;
48 import java
.nio
.charset
.CodingErrorAction
;
49 import java
.util
.Arrays
;
51 import org
.spearce
.jgit
.lib
.Constants
;
52 import org
.spearce
.jgit
.lib
.PersonIdent
;
54 /** Handy utility functions to parse raw object contents. */
55 public final class RawParseUtils
{
56 private static final byte[] digits
;
59 digits
= new byte['9' + 1];
60 Arrays
.fill(digits
, (byte) -1);
61 for (char i
= '0'; i
<= '9'; i
++)
62 digits
[i
] = (byte) (i
- '0');
66 * Determine if b[ptr] matches src.
71 * first position within b, this should match src[0].
73 * the buffer to test for equality with b.
74 * @return ptr + src.length if b[ptr..src.length] == src; else -1.
76 public static final int match(final byte[] b
, int ptr
, final byte[] src
) {
77 if (ptr
+ src
.length
> b
.length
)
79 for (int i
= 0; i
< src
.length
; i
++, ptr
++)
85 private static final byte[] base10byte
= { '0', '1', '2', '3', '4', '5',
89 * Format a base 10 numeric into a temporary buffer.
91 * Formatting is performed backwards. The method starts at offset
92 * <code>o-1</code> and ends at <code>o-1-digits</code>, where
93 * <code>digits</code> is the number of positions necessary to store the
96 * The argument and return values from this method make it easy to chain
97 * writing, for example:
101 * final byte[] tmp = new byte[64];
102 * int ptr = tmp.length;
104 * ptr = RawParseUtils.formatBase10(tmp, ptr, 32);
106 * ptr = RawParseUtils.formatBase10(tmp, ptr, 18);
108 * final String str = new String(tmp, ptr, tmp.length - ptr);
112 * buffer to write into.
114 * one offset past the location where writing will begin; writing
115 * proceeds towards lower index values.
117 * the value to store.
118 * @return the new offset value <code>o</code>. This is the position of
119 * the last byte written. Additional writing should start at one
122 public static int formatBase10(final byte[] b
, int o
, int value
) {
127 final boolean isneg
= value
< 0;
129 b
[--o
] = base10byte
[value
% 10];
138 * Parse a base 10 numeric from a sequence of ASCII digits into an int.
140 * Digit sequences can begin with an optional run of spaces before the
141 * sequence, and may start with a '+' or a '-' to indicate sign position.
142 * Any other characters will cause the method to stop and return the current
143 * result to the caller.
148 * position within buffer to start parsing digits at.
150 * optional location to return the new ptr value through. If null
151 * the ptr value will be discarded.
152 * @return the value at this location; 0 if the location is not a valid
155 public static final int parseBase10(final byte[] b
, int ptr
,
156 final MutableInteger ptrResult
) {
160 final int sz
= b
.length
;
161 while (ptr
< sz
&& b
[ptr
] == ' ')
177 final byte v
= digits
[b
[ptr
]];
183 } catch (ArrayIndexOutOfBoundsException e
) {
184 // Not a valid digit.
186 if (ptrResult
!= null)
187 ptrResult
.value
= ptr
;
188 return sign
< 0 ?
-r
: r
;
192 * Parse a base 10 numeric from a sequence of ASCII digits into a long.
194 * Digit sequences can begin with an optional run of spaces before the
195 * sequence, and may start with a '+' or a '-' to indicate sign position.
196 * Any other characters will cause the method to stop and return the current
197 * result to the caller.
202 * position within buffer to start parsing digits at.
204 * optional location to return the new ptr value through. If null
205 * the ptr value will be discarded.
206 * @return the value at this location; 0 if the location is not a valid
209 public static final long parseLongBase10(final byte[] b
, int ptr
,
210 final MutableInteger ptrResult
) {
214 final int sz
= b
.length
;
215 while (ptr
< sz
&& b
[ptr
] == ' ')
231 final byte v
= digits
[b
[ptr
]];
237 } catch (ArrayIndexOutOfBoundsException e
) {
238 // Not a valid digit.
240 if (ptrResult
!= null)
241 ptrResult
.value
= ptr
;
242 return sign
< 0 ?
-r
: r
;
246 * Parse a Git style timezone string.
248 * The sequence "-0315" will be parsed as the numeric value -195, as the
249 * lower two positions count minutes, not 100ths of an hour.
254 * position within buffer to start parsing digits at.
255 * @return the timezone at this location, expressed in minutes.
257 public static final int parseTimeZoneOffset(final byte[] b
, int ptr
) {
258 final int v
= parseBase10(b
, ptr
, null);
259 final int tzMins
= v
% 100;
260 final int tzHours
= v
/ 100;
261 return tzHours
* 60 + tzMins
;
265 * Locate the first position after a given character.
270 * position within buffer to start looking for chrA at.
273 * @return new position just after chrA.
275 public static final int next(final byte[] b
, int ptr
, final char chrA
) {
276 final int sz
= b
.length
;
278 if (b
[ptr
++] == chrA
)
285 * Locate the first position after the next LF.
287 * This method stops on the first '\n' it finds.
292 * position within buffer to start looking for LF at.
293 * @return new position just after the first LF found.
295 public static final int nextLF(final byte[] b
, int ptr
) {
296 return next(b
, ptr
, '\n');
300 * Locate the first position after either the given character or LF.
302 * This method stops on the first match it finds from either chrA or '\n'.
307 * position within buffer to start looking for chrA or LF at.
310 * @return new position just after the first chrA or LF to be found.
312 public static final int nextLF(final byte[] b
, int ptr
, final char chrA
) {
313 final int sz
= b
.length
;
315 final byte c
= b
[ptr
++];
316 if (c
== chrA
|| c
== '\n')
323 * Index the region between <code>[ptr, end)</code> to find line starts.
325 * The returned list is 1 indexed. Index 0 contains
326 * {@link Integer#MIN_VALUE} to pad the list out.
328 * Using a 1 indexed list means that line numbers can be directly accessed
329 * from the list, so <code>list.get(1)</code> (aka get line 1) returns
335 * position within the buffer corresponding to the first byte of
338 * 1 past the end of the content within <code>buf</code>.
339 * @return a line map indexing the start position of each line.
341 public static final IntList
lineMap(final byte[] buf
, int ptr
, int end
) {
342 // Experimentally derived from multiple source repositories
343 // the average number of bytes/line is 36. Its a rough guess
344 // to initially size our map close to the target.
346 final IntList map
= new IntList((end
- ptr
) / 36);
347 map
.fillTo(1, Integer
.MIN_VALUE
);
348 for (; ptr
< end
; ptr
= nextLF(buf
, ptr
))
354 * Locate the "author " header line data.
359 * position in buffer to start the scan at. Most callers should
360 * pass 0 to ensure the scan starts from the beginning of the
361 * commit buffer and does not accidentally look at message body.
362 * @return position just after the space in "author ", so the first
363 * character of the author's name. If no author header can be
364 * located -1 is returned.
366 public static final int author(final byte[] b
, int ptr
) {
367 final int sz
= b
.length
;
369 ptr
+= 46; // skip the "tree ..." line.
370 while (ptr
< sz
&& b
[ptr
] == 'p')
371 ptr
+= 48; // skip this parent.
372 return match(b
, ptr
, author
);
376 * Locate the "committer " header line data.
381 * position in buffer to start the scan at. Most callers should
382 * pass 0 to ensure the scan starts from the beginning of the
383 * commit buffer and does not accidentally look at message body.
384 * @return position just after the space in "committer ", so the first
385 * character of the committer's name. If no committer header can be
386 * located -1 is returned.
388 public static final int committer(final byte[] b
, int ptr
) {
389 final int sz
= b
.length
;
391 ptr
+= 46; // skip the "tree ..." line.
392 while (ptr
< sz
&& b
[ptr
] == 'p')
393 ptr
+= 48; // skip this parent.
394 if (ptr
< sz
&& b
[ptr
] == 'a')
395 ptr
= nextLF(b
, ptr
);
396 return match(b
, ptr
, committer
);
400 * Locate the "encoding " header line.
405 * position in buffer to start the scan at. Most callers should
406 * pass 0 to ensure the scan starts from the beginning of the
407 * buffer and does not accidentally look at the message body.
408 * @return position just after the space in "encoding ", so the first
409 * character of the encoding's name. If no encoding header can be
410 * located -1 is returned (and UTF-8 should be assumed).
412 public static final int encoding(final byte[] b
, int ptr
) {
413 final int sz
= b
.length
;
419 ptr
= nextLF(b
, ptr
);
421 return match(b
, ptr
, encoding
);
425 * Parse the "encoding " header into a character set reference.
427 * Locates the "encoding " header (if present) by first calling
428 * {@link #encoding(byte[], int)} and then returns the proper character set
429 * to apply to this buffer to evaluate its contents as character data.
431 * If no encoding header is present, {@link Constants#CHARSET} is assumed.
435 * @return the Java character set representation. Never null.
437 public static Charset
parseEncoding(final byte[] b
) {
438 final int enc
= encoding(b
, 0);
440 return Constants
.CHARSET
;
441 final int lf
= nextLF(b
, enc
);
442 return Charset
.forName(decode(Constants
.CHARSET
, b
, enc
, lf
- 1));
446 * Parse a name line (e.g. author, committer, tagger) into a PersonIdent.
448 * When passing in a value for <code>nameB</code> callers should use the
449 * return value of {@link #author(byte[], int)} or
450 * {@link #committer(byte[], int)}, as these methods provide the proper
451 * position within the buffer.
454 * the buffer to parse character data from.
456 * first position of the identity information. This should be the
457 * first position after the space which delimits the header field
458 * name (e.g. "author" or "committer") from the rest of the
460 * @return the parsed identity. Never null.
462 public static PersonIdent
parsePersonIdent(final byte[] raw
, final int nameB
) {
463 final Charset cs
= parseEncoding(raw
);
464 final int emailB
= nextLF(raw
, nameB
, '<');
465 final int emailE
= nextLF(raw
, emailB
, '>');
467 final String name
= decode(cs
, raw
, nameB
, emailB
- 2);
468 final String email
= decode(cs
, raw
, emailB
, emailE
- 1);
470 final MutableInteger ptrout
= new MutableInteger();
471 final long when
= parseLongBase10(raw
, emailE
+ 1, ptrout
);
472 final int tz
= parseTimeZoneOffset(raw
, ptrout
.value
);
474 return new PersonIdent(name
, email
, when
* 1000L, tz
);
478 * Decode a buffer under UTF-8, if possible.
480 * If the byte stream cannot be decoded that way, the platform default is tried
481 * and if that too fails, the fail-safe ISO-8859-1 encoding is tried.
484 * buffer to pull raw bytes from.
485 * @return a string representation of the range <code>[start,end)</code>,
486 * after decoding the region through the specified character set.
488 public static String
decode(final byte[] buffer
) {
489 return decode(Constants
.CHARSET
, buffer
, 0, buffer
.length
);
493 * Decode a buffer under the specified character set if possible.
495 * If the byte stream cannot be decoded that way, the platform default is tried
496 * and if that too fails, the fail-safe ISO-8859-1 encoding is tried.
499 * character set to use when decoding the buffer.
501 * buffer to pull raw bytes from.
502 * @return a string representation of the range <code>[start,end)</code>,
503 * after decoding the region through the specified character set.
505 public static String
decode(final Charset cs
, final byte[] buffer
) {
506 return decode(cs
, buffer
, 0, buffer
.length
);
510 * Decode a region of the buffer under the specified character set if possible.
512 * If the byte stream cannot be decoded that way, the platform default is tried
513 * and if that too fails, the fail-safe ISO-8859-1 encoding is tried.
516 * character set to use when decoding the buffer.
518 * buffer to pull raw bytes from.
520 * first position within the buffer to take data from.
522 * one position past the last location within the buffer to take
524 * @return a string representation of the range <code>[start,end)</code>,
525 * after decoding the region through the specified character set.
527 public static String
decode(final Charset cs
, final byte[] buffer
,
528 final int start
, final int end
) {
530 return decodeNoFallback(cs
, buffer
, start
, end
);
531 } catch (CharacterCodingException e
) {
532 // Fall back to an ISO-8859-1 style encoding. At least all of
533 // the bytes will be present in the output.
535 return extractBinaryString(buffer
, start
, end
);
540 * Decode a region of the buffer under the specified character set if
543 * If the byte stream cannot be decoded that way, the platform default is
544 * tried and if that too fails, an exception is thrown.
547 * character set to use when decoding the buffer.
549 * buffer to pull raw bytes from.
551 * first position within the buffer to take data from.
553 * one position past the last location within the buffer to take
555 * @return a string representation of the range <code>[start,end)</code>,
556 * after decoding the region through the specified character set.
557 * @throws CharacterCodingException
558 * the input is not in any of the tested character sets.
560 public static String
decodeNoFallback(final Charset cs
,
561 final byte[] buffer
, final int start
, final int end
)
562 throws CharacterCodingException
{
563 final ByteBuffer b
= ByteBuffer
.wrap(buffer
, start
, end
- start
);
566 // Try our built-in favorite. The assumption here is that
567 // decoding will fail if the data is not actually encoded
568 // using that encoder.
571 return decode(b
, Constants
.CHARSET
);
572 } catch (CharacterCodingException e
) {
576 if (!cs
.equals(Constants
.CHARSET
)) {
577 // Try the suggested encoding, it might be right since it was
578 // provided by the caller.
581 return decode(b
, cs
);
582 } catch (CharacterCodingException e
) {
587 // Try the default character set. A small group of people
588 // might actually use the same (or very similar) locale.
590 final Charset defcs
= Charset
.defaultCharset();
591 if (!defcs
.equals(cs
) && !defcs
.equals(Constants
.CHARSET
)) {
593 return decode(b
, defcs
);
594 } catch (CharacterCodingException e
) {
599 throw new CharacterCodingException();
603 * Decode a region of the buffer under the ISO-8859-1 encoding.
605 * Each byte is treated as a single character in the 8859-1 character
606 * encoding, performing a raw binary->char conversion.
609 * buffer to pull raw bytes from.
611 * first position within the buffer to take data from.
613 * one position past the last location within the buffer to take
615 * @return a string representation of the range <code>[start,end)</code>.
617 public static String
extractBinaryString(final byte[] buffer
,
618 final int start
, final int end
) {
619 final StringBuilder r
= new StringBuilder(end
- start
);
620 for (int i
= start
; i
< end
; i
++)
621 r
.append((char) (buffer
[i
] & 0xff));
625 private static String
decode(final ByteBuffer b
, final Charset charset
)
626 throws CharacterCodingException
{
627 final CharsetDecoder d
= charset
.newDecoder();
628 d
.onMalformedInput(CodingErrorAction
.REPORT
);
629 d
.onUnmappableCharacter(CodingErrorAction
.REPORT
);
630 return d
.decode(b
).toString();
634 * Locate the position of the commit message body.
639 * position in buffer to start the scan at. Most callers should
640 * pass 0 to ensure the scan starts from the beginning of the
642 * @return position of the user's message buffer.
644 public static final int commitMessage(final byte[] b
, int ptr
) {
645 final int sz
= b
.length
;
647 ptr
+= 46; // skip the "tree ..." line.
648 while (ptr
< sz
&& b
[ptr
] == 'p')
649 ptr
+= 48; // skip this parent.
651 // skip any remaining header lines, ignoring what their actual
652 // header line type is.
654 while (ptr
< sz
&& b
[ptr
] != '\n')
655 ptr
= nextLF(b
, ptr
);
656 if (ptr
< sz
&& b
[ptr
] == '\n')
662 * Locate the end of a paragraph.
664 * A paragraph is ended by two consecutive LF bytes.
669 * position in buffer to start the scan at. Most callers will
670 * want to pass the first position of the commit message (as
671 * found by {@link #commitMessage(byte[], int)}.
672 * @return position of the LF at the end of the paragraph;
673 * <code>b.length</code> if no paragraph end could be located.
675 public static final int endOfParagraph(final byte[] b
, final int start
) {
677 final int sz
= b
.length
;
678 while (ptr
< sz
&& b
[ptr
] != '\n')
679 ptr
= nextLF(b
, ptr
);
680 while (0 < ptr
&& start
< ptr
&& b
[ptr
- 1] == '\n')
685 private RawParseUtils() {
686 // Don't create instances of a static only utility.