]>
Commit | Line | Data |
---|---|---|
c401b33c JS |
1 | git-filter-branch(1) |
2 | ==================== | |
3 | ||
4 | NAME | |
5 | ---- | |
6 | git-filter-branch - Rewrite branches | |
7 | ||
8 | SYNOPSIS | |
9 | -------- | |
10 | [verse] | |
11 | 'git-filter-branch' [--env-filter <command>] [--tree-filter <command>] | |
12 | [--index-filter <command>] [--parent-filter <command>] | |
13 | [--msg-filter <command>] [--commit-filter <command>] | |
14 | [--tag-name-filter <command>] [--subdirectory-filter <directory>] | |
5433235d GB |
15 | [--original <namespace>] [-d <directory>] [-f | --force] |
16 | [<rev-list options>...] | |
c401b33c JS |
17 | |
18 | DESCRIPTION | |
19 | ----------- | |
08203668 JS |
20 | Lets you rewrite git revision history by rewriting the branches mentioned |
21 | in the <rev-list options>, applying custom filters on each revision. | |
c401b33c JS |
22 | Those filters can modify each tree (e.g. removing a file or running |
23 | a perl rewrite on all files) or information about each commit. | |
24 | Otherwise, all information (including original commit times or merge | |
25 | information) will be preserved. | |
26 | ||
08203668 JS |
27 | The command will only rewrite the _positive_ refs mentioned in the |
28 | command line (i.e. if you pass 'a..b', only 'b' will be rewritten). | |
29 | If you specify no filters, the commits will be recommitted without any | |
30 | changes, which would normally have no effect. Nevertheless, this may be | |
31 | useful in the future for compensating for some git bugs or such, | |
32 | therefore such a usage is permitted. | |
c401b33c | 33 | |
73616fd3 | 34 | *WARNING*! The rewritten history will have different object names for all |
c401b33c JS |
35 | the objects and will not converge with the original branch. You will not |
36 | be able to easily push and distribute the rewritten branch on top of the | |
37 | original branch. Please do not use this command if you do not know the | |
38 | full implications, and avoid using it anyway, if a simple single commit | |
39 | would suffice to fix your problem. | |
40 | ||
dfd05e38 JS |
41 | Always verify that the rewritten version is correct: The original refs, |
42 | if different from the rewritten ones, will be stored in the namespace | |
43 | 'refs/original/'. | |
c401b33c JS |
44 | |
45 | Note that since this operation is extensively I/O expensive, it might | |
08203668 JS |
46 | be a good idea to redirect the temporary directory off-disk with the |
47 | '-d' option, e.g. on tmpfs. Reportedly the speedup is very noticeable. | |
c401b33c JS |
48 | |
49 | ||
50 | Filters | |
51 | ~~~~~~~ | |
52 | ||
53 | The filters are applied in the order as listed below. The <command> | |
6cb93bf4 JS |
54 | argument is always evaluated in shell using the 'eval' command (with the |
55 | notable exception of the commit filter, for technical reasons). | |
c401b33c JS |
56 | Prior to that, the $GIT_COMMIT environment variable will be set to contain |
57 | the id of the commit being rewritten. Also, GIT_AUTHOR_NAME, | |
58 | GIT_AUTHOR_EMAIL, GIT_AUTHOR_DATE, GIT_COMMITTER_NAME, GIT_COMMITTER_EMAIL, | |
73616fd3 | 59 | and GIT_COMMITTER_DATE are set according to the current commit. |
c401b33c JS |
60 | |
61 | A 'map' function is available that takes an "original sha1 id" argument | |
62 | and outputs a "rewritten sha1 id" if the commit has been already | |
32c37c12 JS |
63 | rewritten, and "original sha1 id" otherwise; the 'map' function can |
64 | return several ids on separate lines if your commit filter emitted | |
65 | multiple commits. | |
c401b33c JS |
66 | |
67 | ||
68 | OPTIONS | |
69 | ------- | |
70 | ||
71 | --env-filter <command>:: | |
72 | This is the filter for modifying the environment in which | |
73 | the commit will be performed. Specifically, you might want | |
74 | to rewrite the author/committer name/email/time environment | |
75 | variables (see gitlink:git-commit[1] for details). Do not forget | |
76 | to re-export the variables. | |
77 | ||
78 | --tree-filter <command>:: | |
79 | This is the filter for rewriting the tree and its contents. | |
80 | The argument is evaluated in shell with the working | |
81 | directory set to the root of the checked out tree. The new tree | |
82 | is then used as-is (new files are auto-added, disappeared files | |
83 | are auto-removed - neither .gitignore files nor any other ignore | |
73616fd3 | 84 | rules *HAVE ANY EFFECT*!). |
c401b33c JS |
85 | |
86 | --index-filter <command>:: | |
87 | This is the filter for rewriting the index. It is similar to the | |
88 | tree filter but does not check out the tree, which makes it much | |
89 | faster. For hairy cases, see gitlink:git-update-index[1]. | |
90 | ||
91 | --parent-filter <command>:: | |
92 | This is the filter for rewriting the commit's parent list. | |
93 | It will receive the parent string on stdin and shall output | |
94 | the new parent string on stdout. The parent string is in | |
95 | a format accepted by gitlink:git-commit-tree[1]: empty for | |
96 | the initial commit, "-p parent" for a normal commit and | |
97 | "-p parent1 -p parent2 -p parent3 ..." for a merge commit. | |
98 | ||
99 | --msg-filter <command>:: | |
100 | This is the filter for rewriting the commit messages. | |
101 | The argument is evaluated in the shell with the original | |
102 | commit message on standard input; its standard output is | |
103 | used as the new commit message. | |
104 | ||
105 | --commit-filter <command>:: | |
106 | This is the filter for performing the commit. | |
107 | If this filter is specified, it will be called instead of the | |
108 | gitlink:git-commit-tree[1] command, with arguments of the form | |
109 | "<TREE_ID> [-p <PARENT_COMMIT_ID>]..." and the log message on | |
110 | stdin. The commit id is expected on stdout. | |
111 | + | |
112 | As a special extension, the commit filter may emit multiple | |
113 | commit ids; in that case, ancestors of the original commit will | |
114 | have all of them as parents. | |
115 | ||
116 | --tag-name-filter <command>:: | |
117 | This is the filter for rewriting tag names. When passed, | |
118 | it will be called for every tag ref that points to a rewritten | |
119 | object (or to a tag object which points to a rewritten object). | |
120 | The original tag name is passed via standard input, and the new | |
121 | tag name is expected on standard output. | |
122 | + | |
123 | The original tags are not deleted, but can be overwritten; | |
5876b8ee | 124 | use "--tag-name-filter cat" to simply update the tags. In this |
c401b33c JS |
125 | case, be very careful and make sure you have the old tags |
126 | backed up in case the conversion has run afoul. | |
127 | + | |
128 | Note that there is currently no support for proper rewriting of | |
129 | tag objects; in layman terms, if the tag has a message or signature | |
130 | attached, the rewritten tag won't have it. Sorry. (It is by | |
131 | definition impossible to preserve signatures at any rate.) | |
132 | ||
133 | --subdirectory-filter <directory>:: | |
73616fd3 JS |
134 | Only look at the history which touches the given subdirectory. |
135 | The result will contain that directory (and only that) as its | |
136 | project root. | |
c401b33c | 137 | |
5433235d GB |
138 | --original <namespace>:: |
139 | Use this option to set the namespace where the original commits | |
140 | will be stored. The default value is 'refs/original'. | |
141 | ||
c401b33c JS |
142 | -d <directory>:: |
143 | Use this option to set the path to the temporary directory used for | |
144 | rewriting. When applying a tree filter, the command needs to | |
145 | temporary checkout the tree to some directory, which may consume | |
146 | considerable space in case of large projects. By default it | |
147 | does this in the '.git-rewrite/' directory but you can override | |
148 | that choice by this parameter. | |
149 | ||
dfd05e38 JS |
150 | -f\|--force:: |
151 | `git filter-branch` refuses to start with an existing temporary | |
152 | directory or when there are already refs starting with | |
153 | 'refs/original/', unless forced. | |
154 | ||
c401b33c JS |
155 | <rev-list-options>:: |
156 | When options are given after the new branch name, they will | |
157 | be passed to gitlink:git-rev-list[1]. Only commits in the resulting | |
158 | output will be filtered, although the filtered commits can still | |
159 | reference parents which are outside of that set. | |
160 | ||
161 | ||
162 | Examples | |
163 | -------- | |
164 | ||
165 | Suppose you want to remove a file (containing confidential information | |
166 | or copyright violation) from all commits: | |
167 | ||
168 | ------------------------------------------------------- | |
dfd05e38 | 169 | git filter-branch --tree-filter 'rm filename' HEAD |
c401b33c JS |
170 | ------------------------------------------------------- |
171 | ||
172 | A significantly faster version: | |
173 | ||
dfd05e38 JS |
174 | -------------------------------------------------------------------------- |
175 | git filter-branch --index-filter 'git update-index --remove filename' HEAD | |
176 | -------------------------------------------------------------------------- | |
c401b33c JS |
177 | |
178 | Now, you will get the rewritten history saved in the branch 'newbranch' | |
179 | (your current branch is left untouched). | |
180 | ||
32c37c12 JS |
181 | To set a commit (which typically is at the tip of another |
182 | history) to be the parent of the current initial commit, in | |
183 | order to paste the other history behind the current history: | |
c401b33c | 184 | |
dfd05e38 JS |
185 | ------------------------------------------------------------------- |
186 | git filter-branch --parent-filter 'sed "s/^\$/-p <graft-id>/"' HEAD | |
187 | ------------------------------------------------------------------- | |
c401b33c | 188 | |
08203668 JS |
189 | (if the parent string is empty - which happens when we are dealing with |
190 | the initial commit - add graftcommit as a parent). Note that this assumes | |
c401b33c JS |
191 | history with a single root (that is, no merge without common ancestors |
192 | happened). If this is not the case, use: | |
193 | ||
dfd05e38 | 194 | -------------------------------------------------------------------------- |
c401b33c | 195 | git filter-branch --parent-filter \ |
dfd05e38 JS |
196 | 'cat; test $GIT_COMMIT = <commit-id> && echo "-p <graft-id>"' HEAD |
197 | -------------------------------------------------------------------------- | |
c401b33c | 198 | |
32c37c12 JS |
199 | or even simpler: |
200 | ||
201 | ----------------------------------------------- | |
202 | echo "$commit-id $graft-id" >> .git/info/grafts | |
dfd05e38 | 203 | git filter-branch $graft-id..HEAD |
32c37c12 JS |
204 | ----------------------------------------------- |
205 | ||
c401b33c JS |
206 | To remove commits authored by "Darl McBribe" from the history: |
207 | ||
208 | ------------------------------------------------------------------------------ | |
209 | git filter-branch --commit-filter ' | |
210 | if [ "$GIT_AUTHOR_NAME" = "Darl McBribe" ]; | |
211 | then | |
212 | shift; | |
213 | while [ -n "$1" ]; | |
214 | do | |
215 | shift; | |
216 | echo "$1"; | |
217 | shift; | |
218 | done; | |
219 | else | |
220 | git commit-tree "$@"; | |
dfd05e38 | 221 | fi' HEAD |
c401b33c JS |
222 | ------------------------------------------------------------------------------ |
223 | ||
224 | The shift magic first throws away the tree id and then the -p | |
225 | parameters. Note that this handles merges properly! In case Darl | |
226 | committed a merge between P1 and P2, it will be propagated properly | |
227 | and all children of the merge will become merge commits with P1,P2 | |
228 | as their parents instead of the merge commit. | |
229 | ||
230 | To restrict rewriting to only part of the history, specify a revision | |
231 | range in addition to the new branch name. The new branch name will | |
232 | point to the top-most revision that a 'git rev-list' of this range | |
233 | will print. | |
234 | ||
08203668 JS |
235 | *NOTE* the changes introduced by the commits, and which are not reverted |
236 | by subsequent commits, will still be in the rewritten branch. If you want | |
c401b33c JS |
237 | to throw out _changes_ together with the commits, you should use the |
238 | interactive mode of gitlink:git-rebase[1]. | |
239 | ||
08203668 | 240 | |
c401b33c JS |
241 | Consider this history: |
242 | ||
243 | ------------------ | |
244 | D--E--F--G--H | |
245 | / / | |
246 | A--B-----C | |
247 | ------------------ | |
248 | ||
249 | To rewrite only commits D,E,F,G,H, but leave A, B and C alone, use: | |
250 | ||
251 | -------------------------------- | |
dfd05e38 | 252 | git filter-branch ... C..H |
c401b33c JS |
253 | -------------------------------- |
254 | ||
255 | To rewrite commits E,F,G,H, use one of these: | |
256 | ||
257 | ---------------------------------------- | |
dfd05e38 JS |
258 | git filter-branch ... C..H --not D |
259 | git filter-branch ... D..H --not C | |
c401b33c JS |
260 | ---------------------------------------- |
261 | ||
262 | To move the whole tree into a subdirectory, or remove it from there: | |
263 | ||
264 | --------------------------------------------------------------- | |
265 | git filter-branch --index-filter \ | |
266 | 'git ls-files -s | sed "s-\t-&newsubdir/-" | | |
267 | GIT_INDEX_FILE=$GIT_INDEX_FILE.new \ | |
268 | git update-index --index-info && | |
dfd05e38 | 269 | mv $GIT_INDEX_FILE.new $GIT_INDEX_FILE' HEAD |
c401b33c JS |
270 | --------------------------------------------------------------- |
271 | ||
272 | ||
273 | Author | |
274 | ------ | |
275 | Written by Petr "Pasky" Baudis <pasky@suse.cz>, | |
276 | and the git list <git@vger.kernel.org> | |
277 | ||
278 | Documentation | |
279 | -------------- | |
280 | Documentation by Petr Baudis and the git list. | |
281 | ||
282 | GIT | |
283 | --- | |
284 | Part of the gitlink:git[7] suite |