]>
Commit | Line | Data |
---|---|---|
4a1332d0 JH |
1 | Tweaking diff output |
2 | ==================== | |
3 | June 2005 | |
4 | ||
5 | ||
6 | Introduction | |
7 | ------------ | |
8 | ||
4cc41a16 JH |
9 | The diff commands git-diff-index, git-diff-files, and git-diff-tree |
10 | can be told to manipulate differences they find in | |
59df2a11 CS |
11 | unconventional ways before showing diff(1) output. The manipulation |
12 | is collectively called "diffcore transformation". This short note | |
13 | describes what they are and how to use them to produce diff outputs | |
14 | that are easier to understand than the conventional kind. | |
4a1332d0 JH |
15 | |
16 | ||
17 | The chain of operation | |
18 | ---------------------- | |
19 | ||
20 | The git-diff-* family works by first comparing two sets of | |
21 | files: | |
22 | ||
215a7ad1 | 23 | - git-diff-index compares contents of a "tree" object and the |
e1ccf53a YS |
24 | working directory (when '\--cached' flag is not used) or a |
25 | "tree" object and the index file (when '\--cached' flag is | |
4a1332d0 JH |
26 | used); |
27 | ||
28 | - git-diff-files compares contents of the index file and the | |
29 | working directory; | |
30 | ||
59df2a11 CS |
31 | - git-diff-tree compares contents of two "tree" objects; |
32 | ||
4a1332d0 JH |
33 | In all of these cases, the commands themselves compare |
34 | corresponding paths in the two sets of files. The result of | |
35 | comparison is passed from these commands to what is internally | |
36 | called "diffcore", in a format similar to what is output when | |
37 | the -p option is not used. E.g. | |
38 | ||
8db9307c JH |
39 | ------------------------------------------------ |
40 | in-place edit :100644 100644 bcd1234... 0123456... M file0 | |
41 | create :000000 100644 0000000... 1234567... A file4 | |
42 | delete :100644 000000 1234567... 0000000... D file5 | |
43 | unmerged :000000 000000 0000000... 0000000... U file6 | |
44 | ------------------------------------------------ | |
4a1332d0 JH |
45 | |
46 | The diffcore mechanism is fed a list of such comparison results | |
47 | (each of which is called "filepair", although at this point each | |
48 | of them talks about a single file), and transforms such a list | |
28f8faff | 49 | into another list. There are currently 6 such transformations: |
4a1332d0 | 50 | |
8db9307c JH |
51 | - diffcore-pathspec |
52 | - diffcore-break | |
53 | - diffcore-rename | |
54 | - diffcore-merge-broken | |
55 | - diffcore-pickaxe | |
56 | - diffcore-order | |
4a1332d0 | 57 | |
8db9307c | 58 | These are applied in sequence. The set of filepairs git-diff-\* |
4a1332d0 JH |
59 | commands find are used as the input to diffcore-pathspec, and |
60 | the output from diffcore-pathspec is used as the input to the | |
61 | next transformation. The final result is then passed to the | |
62 | output routine and generates either diff-raw format (see Output | |
8db9307c | 63 | format sections of the manual for git-diff-\* commands) or |
4a1332d0 JH |
64 | diff-patch format. |
65 | ||
66 | ||
59df2a11 | 67 | diffcore-pathspec: For Ignoring Files Outside Our Consideration |
a67c1d08 | 68 | --------------------------------------------------------------- |
4a1332d0 JH |
69 | |
70 | The first transformation in the chain is diffcore-pathspec, and | |
71 | is controlled by giving the pathname parameters to the | |
72 | git-diff-* commands on the command line. The pathspec is used | |
73 | to limit the world diff operates in. It removes the filepairs | |
59df2a11 CS |
74 | outside the specified set of pathnames. E.g. If the input set |
75 | of filepairs included: | |
76 | ||
77 | ------------------------------------------------ | |
78 | :100644 100644 bcd1234... 0123456... M junkfile | |
79 | ------------------------------------------------ | |
80 | ||
81 | but the command invocation was "git-diff-files myfile", then the | |
82 | junkfile entry would be removed from the list because only "myfile" | |
83 | is under consideration. | |
4a1332d0 JH |
84 | |
85 | Implementation note. For performance reasons, git-diff-tree | |
86 | uses the pathname parameters on the command line to cull set of | |
87 | filepairs it feeds the diffcore mechanism itself, and does not | |
88 | use diffcore-pathspec, but the end result is the same. | |
89 | ||
90 | ||
59df2a11 | 91 | diffcore-break: For Splitting Up "Complete Rewrites" |
a67c1d08 | 92 | ---------------------------------------------------- |
4a1332d0 JH |
93 | |
94 | The second transformation in the chain is diffcore-break, and is | |
95 | controlled by the -B option to the git-diff-* commands. This is | |
96 | used to detect a filepair that represents "complete rewrite" and | |
97 | break such filepair into two filepairs that represent delete and | |
98 | create. E.g. If the input contained this filepair: | |
99 | ||
8db9307c JH |
100 | ------------------------------------------------ |
101 | :100644 100644 bcd1234... 0123456... M file0 | |
102 | ------------------------------------------------ | |
4a1332d0 JH |
103 | |
104 | and if it detects that the file "file0" is completely rewritten, | |
105 | it changes it to: | |
106 | ||
8db9307c JH |
107 | ------------------------------------------------ |
108 | :100644 000000 bcd1234... 0000000... D file0 | |
109 | :000000 100644 0000000... 0123456... A file0 | |
110 | ------------------------------------------------ | |
4a1332d0 JH |
111 | |
112 | For the purpose of breaking a filepair, diffcore-break examines | |
113 | the extent of changes between the contents of the files before | |
114 | and after modification (i.e. the contents that have "bcd1234..." | |
115 | and "0123456..." as their SHA1 content ID, in the above | |
116 | example). The amount of deletion of original contents and | |
117 | insertion of new material are added together, and if it exceeds | |
118 | the "break score", the filepair is broken into two. The break | |
119 | score defaults to 50% of the size of the smaller of the original | |
120 | and the result (i.e. if the edit shrinks the file, the size of | |
121 | the result is used; if the edit lengthens the file, the size of | |
122 | the original is used), and can be customized by giving a number | |
123 | after "-B" option (e.g. "-B75" to tell it to use 75%). | |
124 | ||
125 | ||
59df2a11 | 126 | diffcore-rename: For Detection Renames and Copies |
a67c1d08 | 127 | ------------------------------------------------- |
4a1332d0 JH |
128 | |
129 | This transformation is used to detect renames and copies, and is | |
130 | controlled by the -M option (to detect renames) and the -C option | |
131 | (to detect copies as well) to the git-diff-* commands. If the | |
132 | input contained these filepairs: | |
133 | ||
8db9307c JH |
134 | ------------------------------------------------ |
135 | :100644 000000 0123456... 0000000... D fileX | |
136 | :000000 100644 0000000... 0123456... A file0 | |
137 | ------------------------------------------------ | |
4a1332d0 JH |
138 | |
139 | and the contents of the deleted file fileX is similar enough to | |
140 | the contents of the created file file0, then rename detection | |
141 | merges these filepairs and creates: | |
142 | ||
8db9307c JH |
143 | ------------------------------------------------ |
144 | :100644 100644 0123456... 0123456... R100 fileX file0 | |
145 | ------------------------------------------------ | |
4a1332d0 | 146 | |
59df2a11 CS |
147 | When the "-C" option is used, the original contents of modified files, |
148 | and deleted files (and also unmodified files, if the | |
149 | "\--find-copies-harder" option is used) are considered as candidates | |
150 | of the source files in rename/copy operation. If the input were like | |
151 | these filepairs, that talk about a modified file fileY and a newly | |
4a1332d0 JH |
152 | created file file0: |
153 | ||
8db9307c JH |
154 | ------------------------------------------------ |
155 | :100644 100644 0123456... 1234567... M fileY | |
59df2a11 | 156 | :000000 100644 0000000... bcd3456... A file0 |
8db9307c | 157 | ------------------------------------------------ |
4a1332d0 JH |
158 | |
159 | the original contents of fileY and the resulting contents of | |
160 | file0 are compared, and if they are similar enough, they are | |
161 | changed to: | |
162 | ||
8db9307c JH |
163 | ------------------------------------------------ |
164 | :100644 100644 0123456... 1234567... M fileY | |
59df2a11 | 165 | :100644 100644 0123456... bcd3456... C100 fileY file0 |
8db9307c | 166 | ------------------------------------------------ |
4a1332d0 JH |
167 | |
168 | In both rename and copy detection, the same "extent of changes" | |
169 | algorithm used in diffcore-break is used to determine if two | |
170 | files are "similar enough", and can be customized to use | |
59df2a11 CS |
171 | a similarity score different from the default of 50% by giving a |
172 | number after the "-M" or "-C" option (e.g. "-M8" to tell it to use | |
4a1332d0 JH |
173 | 8/10 = 80%). |
174 | ||
e1ccf53a | 175 | Note. When the "-C" option is used with `\--find-copies-harder` |
8db9307c | 176 | option, git-diff-\* commands feed unmodified filepairs to |
232b75ab JH |
177 | diffcore mechanism as well as modified ones. This lets the copy |
178 | detector consider unmodified files as copy source candidates at | |
e1ccf53a | 179 | the expense of making it slower. Without `\--find-copies-harder`, |
8db9307c | 180 | git-diff-\* commands can detect copies only if the file that was |
232b75ab | 181 | copied happened to have been modified in the same changeset. |
4a1332d0 JH |
182 | |
183 | ||
59df2a11 | 184 | diffcore-merge-broken: For Putting "Complete Rewrites" Back Together |
a67c1d08 | 185 | -------------------------------------------------------------------- |
4a1332d0 JH |
186 | |
187 | This transformation is used to merge filepairs broken by | |
f73ae1fc | 188 | diffcore-break, and not transformed into rename/copy by |
4a1332d0 JH |
189 | diffcore-rename, back into a single modification. This always |
190 | runs when diffcore-break is used. | |
191 | ||
192 | For the purpose of merging broken filepairs back, it uses a | |
193 | different "extent of changes" computation from the ones used by | |
194 | diffcore-break and diffcore-rename. It counts only the deletion | |
195 | from the original, and does not count insertion. If you removed | |
196 | only 10 lines from a 100-line document, even if you added 910 | |
197 | new lines to make a new 1000-line document, you did not do a | |
198 | complete rewrite. diffcore-break breaks such a case in order to | |
199 | help diffcore-rename to consider such filepairs as candidate of | |
200 | rename/copy detection, but if filepairs broken that way were not | |
201 | matched with other filepairs to create rename/copy, then this | |
202 | transformation merges them back into the original | |
203 | "modification". | |
204 | ||
205 | The "extent of changes" parameter can be tweaked from the | |
206 | default 80% (that is, unless more than 80% of the original | |
207 | material is deleted, the broken pairs are merged back into a | |
208 | single modification) by giving a second number to -B option, | |
209 | like these: | |
210 | ||
8db9307c JH |
211 | * -B50/60 (give 50% "break score" to diffcore-break, use 60% |
212 | for diffcore-merge-broken). | |
213 | ||
214 | * -B/60 (the same as above, since diffcore-break defaults to 50%). | |
4a1332d0 | 215 | |
366175ef | 216 | Note that earlier implementation left a broken pair as a separate |
f73ae1fc | 217 | creation and deletion patches. This was an unnecessary hack and |
366175ef JH |
218 | the latest implementation always merges all the broken pairs |
219 | back into modifications, but the resulting patch output is | |
f73ae1fc | 220 | formatted differently for easier review in case of such |
366175ef JH |
221 | a complete rewrite by showing the entire contents of old version |
222 | prefixed with '-', followed by the entire contents of new | |
223 | version prefixed with '+'. | |
224 | ||
4a1332d0 | 225 | |
59df2a11 | 226 | diffcore-pickaxe: For Detecting Addition/Deletion of Specified String |
a67c1d08 | 227 | --------------------------------------------------------------------- |
4a1332d0 JH |
228 | |
229 | This transformation is used to find filepairs that represent | |
230 | changes that touch a specified string, and is controlled by the | |
e1ccf53a | 231 | -S option and the `\--pickaxe-all` option to the git-diff-* |
4a1332d0 JH |
232 | commands. |
233 | ||
234 | When diffcore-pickaxe is in use, it checks if there are | |
235 | filepairs whose "original" side has the specified string and | |
236 | whose "result" side does not. Such a filepair represents "the | |
237 | string appeared in this changeset". It also checks for the | |
238 | opposite case that loses the specified string. | |
239 | ||
e1ccf53a | 240 | When `\--pickaxe-all` is not in effect, diffcore-pickaxe leaves |
59df2a11 | 241 | only such filepairs that touch the specified string in its |
e1ccf53a | 242 | output. When `\--pickaxe-all` is used, diffcore-pickaxe leaves all |
4a1332d0 JH |
243 | filepairs intact if there is such a filepair, or makes the |
244 | output empty otherwise. The latter behaviour is designed to | |
245 | make reviewing of the changes in the context of the whole | |
246 | changeset easier. | |
247 | ||
248 | ||
59df2a11 | 249 | diffcore-order: For Sorting the Output Based on Filenames |
a67c1d08 | 250 | --------------------------------------------------------- |
4a1332d0 JH |
251 | |
252 | This is used to reorder the filepairs according to the user's | |
253 | (or project's) taste, and is controlled by the -O option to the | |
254 | git-diff-* commands. | |
255 | ||
59df2a11 | 256 | This takes a text file each of whose lines is a shell glob |
4a1332d0 JH |
257 | pattern. Filepairs that match a glob pattern on an earlier line |
258 | in the file are output before ones that match a later line, and | |
259 | filepairs that do not match any glob pattern are output last. | |
260 | ||
59df2a11 | 261 | As an example, a typical orderfile for the core git probably |
8db9307c | 262 | would look like this: |
4a1332d0 | 263 | |
8db9307c | 264 | ------------------------------------------------ |
df8baa42 JF |
265 | README |
266 | Makefile | |
267 | Documentation | |
268 | *.h | |
269 | *.c | |
270 | t | |
8db9307c | 271 | ------------------------------------------------ |
4a1332d0 | 272 |