git log -p --merge [[--] paths...]
[git/git.git] / Documentation / SubmittingPatches
1 I started reading over the SubmittingPatches document for Linux
2 kernel, primarily because I wanted to have a document similar to
3 it for the core GIT to make sure people understand what they are
4 doing when they write "Signed-off-by" line.
6 But the patch submission requirements are a lot more relaxed
7 here on the technical/contents front, because the core GIT is
8 thousand times smaller ;-). So here is only the relevant bits.
11 (1) Make separate commits for logically separate changes.
13 Unless your patch is really trivial, you should not be sending
14 out a patch that was generated between your working tree and
15 your commit head. Instead, always make a commit with complete
16 commit message and generate a series of patches from your
17 repository. It is a good discipline.
19 Describe the technical detail of the change(s).
21 If your description starts to get too long, that's a sign that you
22 probably need to split up your commit to finer grained pieces.
24 Oh, another thing. I am picky about whitespaces. Make sure your
25 changes do not trigger errors with the sample pre-commit hook shipped
26 in templates/hooks--pre-commit.
29 (2) Generate your patch using git tools out of your commits.
31 git based diff tools (git, Cogito, and StGIT included) generate
32 unidiff which is the preferred format.
34 You do not have to be afraid to use -M option to "git diff" or
35 "git format-patch", if your patch involves file renames. The
36 receiving end can handle them just fine.
38 Please make sure your patch does not include any extra files
39 which do not belong in a patch submission. Make sure to review
40 your patch after generating it, to ensure accuracy. Before
41 sending out, please make sure it cleanly applies to the "master"
42 branch head. If you are preparing a work based on "next" branch,
43 that is fine, but please mark it as such.
46 (3) Sending your patches.
48 People on the git mailing list need to be able to read and
49 comment on the changes you are submitting. It is important for
50 a developer to be able to "quote" your changes, using standard
51 e-mail tools, so that they may comment on specific portions of
52 your code. For this reason, all patches should be submited
53 "inline". WARNING: Be wary of your MUAs word-wrap
54 corrupting your patch. Do not cut-n-paste your patch; you can
55 lose tabs that way if you are not careful.
57 It is a common convention to prefix your subject line with
58 [PATCH]. This lets people easily distinguish patches from other
59 e-mail discussions.
61 "git format-patch" command follows the best current practice to
62 format the body of an e-mail message. At the beginning of the
63 patch should come your commit message, ending with the
64 Signed-off-by: lines, and a line that consists of three dashes,
65 followed by the diffstat information and the patch itself. If
66 you are forwarding a patch from somebody else, optionally, at
67 the beginning of the e-mail message just before the commit
68 message starts, you can put a "From: " line to name that person.
70 You often want to add additional explanation about the patch,
71 other than the commit message itself. Place such "cover letter"
72 material between the three dash lines and the diffstat.
74 Do not attach the patch as a MIME attachment, compressed or not.
75 Do not let your e-mail client send quoted-printable. Many
76 popular e-mail applications will not always transmit a MIME
77 attachment as plain text, making it impossible to comment on
78 your code. A MIME attachment also takes a bit more time to
79 process. This does not decrease the likelihood of your
80 MIME-attached change being accepted, but it makes it more likely
81 that it will be postponed.
83 Exception: If your mailer is mangling patches then someone may ask
84 you to re-send them using MIME, that is OK.
86 Do not PGP sign your patch, at least for now. Most likely, your
87 maintainer or other people on the list would not have your PGP
88 key and would not bother obtaining it anyway. Your patch is not
89 judged by who you are; a good patch from an unknown origin has a
90 far better chance of being accepted than a patch from a known,
91 respected origin that is done poorly or does incorrect things.
93 If you really really really really want to do a PGP signed
94 patch, format it as "multipart/signed", not a text/plain message
95 that starts with '-----BEGIN PGP SIGNED MESSAGE-----'. That is
96 not a text/plain, it's something else.
98 Note that your maintainer does not necessarily read everything
99 on the git mailing list. If your patch is for discussion first,
100 send it "To:" the mailing list, and optionally "cc:" him. If it
101 is trivially correct or after the list reached a consensus, send
102 it "To:" the maintainer and optionally "cc:" the list.
105 (6) Sign your work
107 To improve tracking of who did what, we've borrowed the
108 "sign-off" procedure from the Linux kernel project on patches
109 that are being emailed around. Although core GIT is a lot
110 smaller project it is a good discipline to follow it.
112 The sign-off is a simple line at the end of the explanation for
113 the patch, which certifies that you wrote it or otherwise have
114 the right to pass it on as a open-source patch. The rules are
115 pretty simple: if you can certify the below:
117 Developer's Certificate of Origin 1.1
119 By making a contribution to this project, I certify that:
121 (a) The contribution was created in whole or in part by me and I
122 have the right to submit it under the open source license
123 indicated in the file; or
125 (b) The contribution is based upon previous work that, to the best
126 of my knowledge, is covered under an appropriate open source
127 license and I have the right under that license to submit that
128 work with modifications, whether created in whole or in part
129 by me, under the same open source license (unless I am
130 permitted to submit under a different license), as indicated
131 in the file; or
133 (c) The contribution was provided directly to me by some other
134 person who certified (a), (b) or (c) and I have not modified
135 it.
137 (d) I understand and agree that this project and the contribution
138 are public and that a record of the contribution (including all
139 personal information I submit with it, including my sign-off) is
140 maintained indefinitely and may be redistributed consistent with
141 this project or the open source license(s) involved.
143 then you just add a line saying
145 Signed-off-by: Random J Developer <>
147 Some people also put extra tags at the end. They'll just be ignored for
148 now, but you can do this to mark internal company procedures or just
149 point out some special detail about the sign-off.
152 ------------------------------------------------
153 MUA specific hints
155 Some of patches I receive or pick up from the list share common
156 patterns of breakage. Please make sure your MUA is set up
157 properly not to corrupt whitespaces. Here are two common ones
158 I have seen:
160 * Empty context lines that do not have _any_ whitespace.
162 * Non empty context lines that have one extra whitespace at the
163 beginning.
165 One test you could do yourself if your MUA is set up correctly is:
167 * Send the patch to yourself, exactly the way you would, except
168 To: and Cc: lines, which would not contain the list and
169 maintainer address.
171 * Save that patch to a file in UNIX mailbox format. Call it say
172 a.patch.
174 * Try to apply to the tip of the "master" branch from the
175 git.git public repository:
177 $ git fetch master:test-apply
178 $ git checkout test-apply
179 $ git reset --hard
180 $ git applymbox a.patch
182 If it does not apply correctly, there can be various reasons.
184 * Your patch itself does not apply cleanly. That is _bad_ but
185 does not have much to do with your MUA. Please rebase the
186 patch appropriately.
188 * Your MUA corrupted your patch; applymbox would complain that
189 the patch does not apply. Look at .dotest/ subdirectory and
190 see what 'patch' file contains and check for the common
191 corruption patterns mentioned above.
193 * While you are at it, check what are in 'info' and
194 'final-commit' files as well. If what is in 'final-commit' is
195 not exactly what you would want to see in the commit log
196 message, it is very likely that your maintainer would end up
197 hand editing the log message when he applies your patch.
198 Things like "Hi, this is my first patch.\n", if you really
199 want to put in the patch e-mail, should come after the
200 three-dash line that signals the end of the commit message.
203 Pine
204 ----
206 (Johannes Schindelin)
208 I don't know how many people still use pine, but for those poor
209 souls it may be good to mention that the quell-flowed-text is
210 needed for recent versions.
212 ... the "no-strip-whitespace-before-send" option, too. AFAIK it
213 was introduced in 4.60.
215 (Linus Torvalds)
217 And 4.58 needs at least this.
219 ---
220 diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1)
221 Author: Linus Torvalds <>
222 Date: Mon Aug 15 17:23:51 2005 -0700
224 Fix pine whitespace-corruption bug
226 There's no excuse for unconditionally removing whitespace from
227 the pico buffers on close.
229 diff --git a/pico/pico.c b/pico/pico.c
230 --- a/pico/pico.c
231 +++ b/pico/pico.c
232 @@ -219,7 +219,9 @@ PICO *pm;
233 switch(pico_all_done){ /* prepare for/handle final events */
234 case COMP_EXIT : /* already confirmed */
235 packheader();
236 +#if 0
237 stripwhitespace();
238 +#endif
239 c |= COMP_EXIT;
240 break;
243 (Daniel Barkalow)
245 > A patch to SubmittingPatches, MUA specific help section for
246 > users of Pine 4.63 would be very much appreciated.
248 Ah, it looks like a recent version changed the default behavior to do the
249 right thing, and inverted the sense of the configuration option. (Either
250 that or Gentoo did it.) So you need to set the
251 "no-strip-whitespace-before-send" option, unless the option you have is
252 "strip-whitespace-before-send", in which case you should avoid checking
253 it.
256 Thunderbird
257 -----------
259 (A Large Angry SCM)
261 Here are some hints on how to successfully submit patches inline using
262 Thunderbird.
264 This recipe appears to work with the current [*1*] Thunderbird from Suse.
266 The following Thunderbird extensions are needed:
267 AboutConfig 0.5
269 External Editor 0.7.2
272 1) Prepare the patch as a text file using your method of choice.
274 2) Before opening a compose window, use Edit->Account Settings to
275 uncheck the "Compose messages in HTML format" setting in the
276 "Composition & Addressing" panel of the account to be used to send the
277 patch. [*2*]
279 3) In the main Thunderbird window, _before_ you open the compose window
280 for the patch, use Tools->about:config to set the following to the
281 indicated values:
282 mailnews.send_plaintext_flowed => false
283 mailnews.wraplength => 0
285 4) Open a compose window and click the external editor icon.
287 5) In the external editor window, read in the patch file and exit the
288 editor normally.
290 6) Back in the compose window: Add whatever other text you wish to the
291 message, complete the addressing and subject fields, and press send.
293 7) Optionally, undo the about:config/account settings changes made in
294 steps 2 & 3.
297 [Footnotes]
298 *1* Version 1.0 (20041207) from the MozillaThunderbird-1.0-5 rpm of Suse
299 9.3 professional updates.
301 *2* It may be possible to do this with about:config and the following
302 settings but I haven't tried, yet.
303 mail.html_compose => false
304 mail.identity.default.compose_html => false
305 => false