Documentation/diff-generate-patch.txt

   1 Generating patches with -p
   2 --------------------------
   3
   4 When "git-diff-index", "git-diff-tree", or "git-diff-files" are run
   5 with a `-p` option, "git diff" without the `--raw` option, or
   6 "git log" with the "-p" option, they
   7 do not produce the output described above; instead they produce a
   8 patch file.  You can customize the creation of such patches via the
   9 `GIT_EXTERNAL_DIFF` and the `GIT_DIFF_OPTS` environment variables.
  10
  11 What the -p option produces is slightly different from the traditional
  12 diff format:
  13
  14 1.   It is preceded with a "git diff" header that looks like this:
  15
  16        diff --git a/file1 b/file2
  17 +
  18 The `a/` and `b/` filenames are the same unless rename/copy is
  19 involved.  Especially, even for a creation or a deletion,
  20 `/dev/null` is _not_ used in place of the `a/` or `b/` filenames.
  21 +
  22 When rename/copy is involved, `file1` and `file2` show the
  23 name of the source file of the rename/copy and the name of
  24 the file that rename/copy produces, respectively.
  25
  26 2.   It is followed by one or more extended header lines:
  27
  28        old mode <mode>
  29        new mode <mode>
  30        deleted file mode <mode>
  31        new file mode <mode>
  32        copy from <path>
  33        copy to <path>
  34        rename from <path>
  35        rename to <path>
  36        similarity index <number>
  37        dissimilarity index <number>
  38        index <hash>..<hash> <mode>
  39 +
  40 File modes are printed as 6-digit octal numbers including the file type
  41 and file permission bits.
  42 +
  43 Path names in extended headers do not include the `a/` and `b/` prefixes.
  44 +
  45 The similarity index is the percentage of unchanged lines, and
  46 the dissimilarity index is the percentage of changed lines.  It
  47 is a rounded down integer, followed by a percent sign.  The
  48 similarity index value of 100% is thus reserved for two equal
  49 files, while 100% dissimilarity means that no line from the old
  50 file made it into the new one.
  51 +
  52 The index line includes the SHA-1 checksum before and after the change.
  53 The <mode> is included if the file mode does not change; otherwise,
  54 separate lines indicate the old and the new mode.
  55
  56 3.  Pathnames with "unusual" characters are quoted as explained for
  57     the configuration variable `core.quotePath` (see
  58     linkgit:git-config[1]).
  59
  60 4.  All the `file1` files in the output refer to files before the
  61     commit, and all the `file2` files refer to files after the commit.
  62     It is incorrect to apply each change to each file sequentially.  For
  63     example, this patch will swap a and b:
  64
  65       diff --git a/a b/b
  66       rename from a
  67       rename to b
  68       diff --git a/b b/a
  69       rename from b
  70       rename to a
  71
  72
  73 combined diff format
  74 --------------------
  75
  76 Any diff-generating command can take the `-c` or `--cc` option to
  77 produce a 'combined diff' when showing a merge. This is the default
  78 format when showing merges with linkgit:git-diff[1] or
  79 linkgit:git-show[1]. Note also that you can give the `-m` option to any
  80 of these commands to force generation of diffs with individual parents
  81 of a merge.
  82
  83 A 'combined diff' format looks like this:
  84
  85 ------------
  86 diff --combined describe.c
  87 index fabadb8,cc95eb0..4866510
  88 --- a/describe.c
  89 +++ b/describe.c
  90 @@@ -98,20 -98,12 +98,20 @@@
  91         return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
  92   }
  93
  94 - static void describe(char *arg)
  95  -static void describe(struct commit *cmit, int last_one)
  96 ++static void describe(char *arg, int last_one)
  97   {
  98  +      unsigned char sha1[20];
  99  +      struct commit *cmit;
 100         struct commit_list *list;
 101         static int initialized = 0;
 102         struct commit_name *n;
 103
 104  +      if (get_sha1(arg, sha1) < 0)
 105  +              usage(describe_usage);
 106  +      cmit = lookup_commit_reference(sha1);
 107  +      if (!cmit)
 108  +              usage(describe_usage);
 109  +
 110         if (!initialized) {
 111                 initialized = 1;
 112                 for_each_ref(get_name);
 113 ------------
 114
 115 1.   It is preceded with a "git diff" header, that looks like
 116      this (when `-c` option is used):
 117
 118        diff --combined file
 119 +
 120 or like this (when `--cc` option is used):
 121
 122        diff --cc file
 123
 124 2.   It is followed by one or more extended header lines
 125      (this example shows a merge with two parents):
 126
 127        index <hash>,<hash>..<hash>
 128        mode <mode>,<mode>..<mode>
 129        new file mode <mode>
 130        deleted file mode <mode>,<mode>
 131 +
 132 The `mode <mode>,<mode>..<mode>` line appears only if at least one of
 133 the <mode> is different from the rest. Extended headers with
 134 information about detected contents movement (renames and
 135 copying detection) are designed to work with diff of two
 136 <tree-ish> and are not used by combined diff format.
 137
 138 3.   It is followed by two-line from-file/to-file header
 139
 140        --- a/file
 141        +++ b/file
 142 +
 143 Similar to two-line header for traditional 'unified' diff
 144 format, `/dev/null` is used to signal created or deleted
 145 files.
 146 +
 147 However, if the --combined-all-paths option is provided, instead of a
 148 two-line from-file/to-file you get a N+1 line from-file/to-file header,
 149 where N is the number of parents in the merge commit
 150
 151        --- a/file
 152        --- a/file
 153        --- a/file
 154        +++ b/file
 155 +
 156 This extended format can be useful if rename or copy detection is
 157 active, to allow you to see the original name of the file in different
 158 parents.
 159
 160 4.   Chunk header format is modified to prevent people from
 161      accidentally feeding it to `patch -p1`. Combined diff format
 162      was created for review of merge commit changes, and was not
 163      meant for apply. The change is similar to the change in the
 164      extended 'index' header:
 165
 166        @@@ <from-file-range> <from-file-range> <to-file-range> @@@
 167 +
 168 There are (number of parents + 1) `@` characters in the chunk
 169 header for combined diff format.
 170
 171 Unlike the traditional 'unified' diff format, which shows two
 172 files A and B with a single column that has `-` (minus --
 173 appears in A but removed in B), `+` (plus -- missing in A but
 174 added to B), or `" "` (space -- unchanged) prefix, this format
 175 compares two or more files file1, file2,... with one file X, and
 176 shows how X differs from each of fileN.  One column for each of
 177 fileN is prepended to the output line to note how X's line is
 178 different from it.
 179
 180 A `-` character in the column N means that the line appears in
 181 fileN but it does not appear in the result.  A `+` character
 182 in the column N means that the line appears in the result,
 183 and fileN does not have that line (in other words, the line was
 184 added, from the point of view of that parent).
 185
 186 In the above example output, the function signature was changed
 187 from both files (hence two `-` removals from both file1 and
 188 file2, plus `++` to mean one line that was added does not appear
 189 in either file1 or file2).  Also eight other lines are the same
 190 from file1 but do not appear in file2 (hence prefixed with `+`).
 191
 192 When shown by `git diff-tree -c`, it compares the parents of a
 193 merge commit with the merge result (i.e. file1..fileN are the
 194 parents).  When shown by `git diff-files -c`, it compares the
 195 two unresolved merge parents with the working tree file
 196 (i.e. file1 is stage 2 aka "our version", file2 is stage 3 aka
 197 "their version").