Merge branch 'maint-1.6.0' into maint-1.6.1
[git/git.git] / Documentation / git-blame.txt
1 git-blame(1)
2 ============
3
4 NAME
5 ----
6 git-blame - Show what revision and author last modified each line of a file
7
8 SYNOPSIS
9 --------
10 [verse]
11 'git blame' [-c] [-b] [-l] [--root] [-t] [-f] [-n] [-s] [-p] [-w] [--incremental] [-L n,m]
12 [-S <revs-file>] [-M] [-C] [-C] [--since=<date>]
13 [<rev> | --contents <file> | --reverse <rev>] [--] <file>
14
15 DESCRIPTION
16 -----------
17
18 Annotates each line in the given file with information from the revision which
19 last modified the line. Optionally, start annotating from the given revision.
20
21 Also it can limit the range of lines annotated.
22
23 This report doesn't tell you anything about lines which have been deleted or
24 replaced; you need to use a tool such as 'git-diff' or the "pickaxe"
25 interface briefly mentioned in the following paragraph.
26
27 Apart from supporting file annotation, git also supports searching the
28 development history for when a code snippet occurred in a change. This makes it
29 possible to track when a code snippet was added to a file, moved or copied
30 between files, and eventually deleted or replaced. It works by searching for
31 a text string in the diff. A small example:
32
33 -----------------------------------------------------------------------------
34 $ git log --pretty=oneline -S'blame_usage'
35 5040f17eba15504bad66b14a645bddd9b015ebb7 blame -S <ancestry-file>
36 ea4c7f9bf69e781dd0cd88d2bccb2bf5cc15c9a7 git-blame: Make the output
37 -----------------------------------------------------------------------------
38
39 OPTIONS
40 -------
41 include::blame-options.txt[]
42
43 -c::
44 Use the same output mode as linkgit:git-annotate[1] (Default: off).
45
46 --score-debug::
47 Include debugging information related to the movement of
48 lines between files (see `-C`) and lines moved within a
49 file (see `-M`). The first number listed is the score.
50 This is the number of alphanumeric characters detected
51 to be moved between or within files. This must be above
52 a certain threshold for 'git-blame' to consider those lines
53 of code to have been moved.
54
55 -f::
56 --show-name::
57 Show filename in the original commit. By default
58 filename is shown if there is any line that came from a
59 file with different name, due to rename detection.
60
61 -n::
62 --show-number::
63 Show line number in the original commit (Default: off).
64
65 -s::
66 Suppress author name and timestamp from the output.
67
68 -w::
69 Ignore whitespace when comparing parent's version and
70 child's to find where the lines came from.
71
72
73 THE PORCELAIN FORMAT
74 --------------------
75
76 In this format, each line is output after a header; the
77 header at the minimum has the first line which has:
78
79 - 40-byte SHA-1 of the commit the line is attributed to;
80 - the line number of the line in the original file;
81 - the line number of the line in the final file;
82 - on a line that starts a group of line from a different
83 commit than the previous one, the number of lines in this
84 group. On subsequent lines this field is absent.
85
86 This header line is followed by the following information
87 at least once for each commit:
88
89 - author name ("author"), email ("author-mail"), time
90 ("author-time"), and timezone ("author-tz"); similarly
91 for committer.
92 - filename in the commit the line is attributed to.
93 - the first line of the commit log message ("summary").
94
95 The contents of the actual line is output after the above
96 header, prefixed by a TAB. This is to allow adding more
97 header elements later.
98
99
100 SPECIFYING RANGES
101 -----------------
102
103 Unlike 'git-blame' and 'git-annotate' in older git, the extent
104 of annotation can be limited to both line ranges and revision
105 ranges. When you are interested in finding the origin for
106 ll. 40-60 for file `foo`, you can use `-L` option like these
107 (they mean the same thing -- both ask for 21 lines starting at
108 line 40):
109
110 git blame -L 40,60 foo
111 git blame -L 40,+21 foo
112
113 Also you can use regular expression to specify the line range.
114
115 git blame -L '/^sub hello {/,/^}$/' foo
116
117 would limit the annotation to the body of `hello` subroutine.
118
119 When you are not interested in changes older than the version
120 v2.6.18, or changes older than 3 weeks, you can use revision
121 range specifiers similar to 'git-rev-list':
122
123 git blame v2.6.18.. -- foo
124 git blame --since=3.weeks -- foo
125
126 When revision range specifiers are used to limit the annotation,
127 lines that have not changed since the range boundary (either the
128 commit v2.6.18 or the most recent commit that is more than 3
129 weeks old in the above example) are blamed for that range
130 boundary commit.
131
132 A particularly useful way is to see if an added file have lines
133 created by copy-and-paste from existing files. Sometimes this
134 indicates that the developer was being sloppy and did not
135 refactor the code properly. You can first find the commit that
136 introduced the file with:
137
138 git log --diff-filter=A --pretty=short -- foo
139
140 and then annotate the change between the commit and its
141 parents, using `commit{caret}!` notation:
142
143 git blame -C -C -f $commit^! -- foo
144
145
146 INCREMENTAL OUTPUT
147 ------------------
148
149 When called with `--incremental` option, the command outputs the
150 result as it is built. The output generally will talk about
151 lines touched by more recent commits first (i.e. the lines will
152 be annotated out of order) and is meant to be used by
153 interactive viewers.
154
155 The output format is similar to the Porcelain format, but it
156 does not contain the actual lines from the file that is being
157 annotated.
158
159 . Each blame entry always starts with a line of:
160
161 <40-byte hex sha1> <sourceline> <resultline> <num_lines>
162 +
163 Line numbers count from 1.
164
165 . The first time that commit shows up in the stream, it has various
166 other information about it printed out with a one-word tag at the
167 beginning of each line about that "extended commit info" (author,
168 email, committer, dates, summary etc).
169
170 . Unlike Porcelain format, the filename information is always
171 given and terminates the entry:
172
173 "filename" <whitespace-quoted-filename-goes-here>
174 +
175 and thus it's really quite easy to parse for some line- and word-oriented
176 parser (which should be quite natural for most scripting languages).
177 +
178 [NOTE]
179 For people who do parsing: to make it more robust, just ignore any
180 lines in between the first and last one ("<sha1>" and "filename" lines)
181 where you don't recognize the tag-words (or care about that particular
182 one) at the beginning of the "extended information" lines. That way, if
183 there is ever added information (like the commit encoding or extended
184 commit commentary), a blame viewer won't ever care.
185
186
187 SEE ALSO
188 --------
189 linkgit:git-annotate[1]
190
191 AUTHOR
192 ------
193 Written by Junio C Hamano <gitster@pobox.com>
194
195 GIT
196 ---
197 Part of the linkgit:git[1] suite