1Generating patches with -p 2-------------------------- 3 4When "git-diff-index", "git-diff-tree", or "git-diff-files" are run 5with a '-p' option, "git diff" without the '--raw' option, or 6"git log" with the "-p" option, they 7do not produce the output described above; instead they produce a 8patch file. You can customize the creation of such patches via the 9GIT_EXTERNAL_DIFF and the GIT_DIFF_OPTS environment variables. 10 11What the -p option produces is slightly different from the traditional 12diff format. 13 141. It is preceded with a "git diff" header, that looks like 15 this: 16 17 diff --git a/file1 b/file2 18+ 19The `a/` and `b/` filenames are the same unless rename/copy is 20involved. Especially, even for a creation or a deletion, 21`/dev/null` is _not_ used in place of `a/` or `b/` filenames. 22+ 23When rename/copy is involved, `file1` and `file2` show the 24name of the source file of the rename/copy and the name of 25the file that rename/copy produces, respectively. 26 272. It is followed by one or more extended header lines: 28 29 old mode <mode> 30 new mode <mode> 31 deleted file mode <mode> 32 new file mode <mode> 33 copy from <path> 34 copy to <path> 35 rename from <path> 36 rename to <path> 37 similarity index <number> 38 dissimilarity index <number> 39 index <hash>..<hash> <mode> 40 413. TAB, LF, double quote and backslash characters in pathnames 42 are represented as `\t`, `\n`, `\"` and `\\`, respectively. 43 If there is need for such substitution then the whole 44 pathname is put in double quotes. 45 46The similarity index is the percentage of unchanged lines, and 47the dissimilarity index is the percentage of changed lines. It 48is a rounded down integer, followed by a percent sign. The 49similarity index value of 100% is thus reserved for two equal 50files, while 100% dissimilarity means that no line from the old 51file made it into the new one. 52 53 54combined diff format 55-------------------- 56 57"git-diff-tree", "git-diff-files" and "git-diff" can take '-c' or 58'--cc' option to produce 'combined diff'. For showing a merge commit 59with "git log -p", this is the default format. 60A 'combined diff' format looks like this: 61 62------------ 63diff --combined describe.c 64index fabadb8,cc95eb0..4866510 65--- a/describe.c 66+++ b/describe.c 67@@@ -98,20 -98,12 +98,20 @@@ 68 return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1; 69 } 70 71- static void describe(char *arg) 72 -static void describe(struct commit *cmit, int last_one) 73++static void describe(char *arg, int last_one) 74 { 75 + unsigned char sha1[20]; 76 + struct commit *cmit; 77 struct commit_list *list; 78 static int initialized = 0; 79 struct commit_name *n; 80 81 + if (get_sha1(arg, sha1) < 0) 82 + usage(describe_usage); 83 + cmit = lookup_commit_reference(sha1); 84 + if (!cmit) 85 + usage(describe_usage); 86 + 87 if (!initialized) { 88 initialized = 1; 89 for_each_ref(get_name); 90------------ 91 921. It is preceded with a "git diff" header, that looks like 93 this (when '-c' option is used): 94 95 diff --combined file 96+ 97or like this (when '--cc' option is used): 98 99 diff --c file 100 1012. It is followed by one or more extended header lines 102 (this example shows a merge with two parents): 103 104 index <hash>,<hash>..<hash> 105 mode <mode>,<mode>..<mode> 106 new file mode <mode> 107 deleted file mode <mode>,<mode> 108+ 109The `mode <mode>,<mode>..<mode>` line appears only if at least one of 110the <mode> is different from the rest. Extended headers with 111information about detected contents movement (renames and 112copying detection) are designed to work with diff of two 113<tree-ish> and are not used by combined diff format. 114 1153. It is followed by two-line from-file/to-file header 116 117 --- a/file 118 +++ b/file 119+ 120Similar to two-line header for traditional 'unified' diff 121format, `/dev/null` is used to signal created or deleted 122files. 123 1244. Chunk header format is modified to prevent people from 125 accidentally feeding it to `patch -p1`. Combined diff format 126 was created for review of merge commit changes, and was not 127 meant for apply. The change is similar to the change in the 128 extended 'index' header: 129 130 @@@ <from-file-range> <from-file-range> <to-file-range> @@@ 131+ 132There are (number of parents + 1) `@` characters in the chunk 133header for combined diff format. 134 135Unlike the traditional 'unified' diff format, which shows two 136files A and B with a single column that has `-` (minus -- 137appears in A but removed in B), `+` (plus -- missing in A but 138added to B), or `" "` (space -- unchanged) prefix, this format 139compares two or more files file1, file2,... with one file X, and 140shows how X differs from each of fileN. One column for each of 141fileN is prepended to the output line to note how X's line is 142different from it. 143 144A `-` character in the column N means that the line appears in 145fileN but it does not appear in the result. A `+` character 146in the column N means that the line appears in the last file, 147and fileN does not have that line (in other words, the line was 148added, from the point of view of that parent). 149 150In the above example output, the function signature was changed 151from both files (hence two `-` removals from both file1 and 152file2, plus `++` to mean one line that was added does not appear 153in either file1 nor file2). Also two other lines are the same 154from file1 but do not appear in file2 (hence prefixed with ` +`). 155 156When shown by `git diff-tree -c`, it compares the parents of a 157merge commit with the merge result (i.e. file1..fileN are the 158parents). When shown by `git diff-files -c`, it compares the 159two unresolved merge parents with the working tree file 160(i.e. file1 is stage 2 aka "our version", file2 is stage 3 aka 161"their version").