Merge git://git.kernel.org/pub/scm/gitk/gitk
[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.
5
6 But the patch submission requirements are a lot more relaxed
7 here, because the core GIT is thousand times smaller ;-). So
8 here is only the relevant bits.
9
10
11 (1) Make separate commits for logically separate changes.
12
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.
18
19 Describe the technical detail of the change(s).
20
21 If your description starts to get long, that's a sign that you
22 probably need to split up your commit to finer grained pieces.
23
24
25 (2) Generate your patch using git/cogito out of your commits.
26
27 git diff tools generate unidiff which is the preferred format.
28 You do not have to be afraid to use -M option to "git diff" or
29 "git format-patch", if your patch involves file renames. The
30 receiving end can handle them just fine.
31
32 Please make sure your patch does not include any extra files
33 which do not belong in a patch submission. Make sure to review
34 your patch after generating it, to ensure accuracy. Before
35 sending out, please make sure it cleanly applies to the "master"
36 branch head.
37
38
39 (3) Sending your patches.
40
41 People on the git mailing list needs to be able to read and
42 comment on the changes you are submitting. It is important for
43 a developer to be able to "quote" your changes, using standard
44 e-mail tools, so that they may comment on specific portions of
45 your code. For this reason, all patches should be submitting
46 e-mail "inline". WARNING: Be wary of your MUAs word-wrap
47 corrupting your patch. Do not cut-n-paste your patch.
48
49 It is common convention to prefix your subject line with
50 [PATCH]. This lets people easily distinguish patches from other
51 e-mail discussions.
52
53 "git format-patch" command follows the best current practice to
54 format the body of an e-mail message. At the beginning of the
55 patch should come your commit message, ending with the
56 Signed-off-by: lines, and a line that consists of three dashes,
57 followed by the diffstat information and the patch itself. If
58 you are forwarding a patch from somebody else, optionally, at
59 the beginning of the e-mail message just before the commit
60 message starts, you can put a "From: " line to name that person.
61
62 You often want to add additional explanation about the patch,
63 other than the commit message itself. Place such "cover letter"
64 material between the three dash lines and the diffstat.
65
66 Do not attach the patch as a MIME attachment, compressed or not.
67 Do not let your e-mail client send quoted-printable. Many
68 popular e-mail applications will not always transmit a MIME
69 attachment as plain text, making it impossible to comment on
70 your code. A MIME attachment also takes a bit more time to
71 process. This does not decrease the likelihood of your
72 MIME-attached change being accepted, but it makes it more likely
73 that it will be postponed.
74
75 Exception: If your mailer is mangling patches then someone may ask
76 you to re-send them using MIME, that is OK.
77
78 Do not PGP sign your patch, at least for now. Most likely, your
79 maintainer or other people on the list would not have your PGP
80 key and would not bother obtaining it anyway. Your patch is not
81 judged by who you are; a good patch from an unknown origin has a
82 far better chance of being accepted than a patch from a known,
83 respected origin that is done poorly or does incorrect things.
84
85 If you really really really really want to do a PGP signed
86 patch, format it as "multipart/signed", not a text/plain message
87 that starts with '-----BEGIN PGP SIGNED MESSAGE-----'. That is
88 not a text/plain, it's something else.
89
90 Note that your maintainer does not necessarily read everything
91 on the git mailing list. If your patch is for discussion first,
92 send it "To:" the mailing list, and optionally "cc:" him. If it
93 is trivially correct or after the list reached a consensus, send
94 it "To:" the maintainer and optionally "cc:" the list.
95
96
97 (6) Sign your work
98
99 To improve tracking of who did what, we've borrowed the
100 "sign-off" procedure from the Linux kernel project on patches
101 that are being emailed around. Although core GIT is a lot
102 smaller project it is a good discipline to follow it.
103
104 The sign-off is a simple line at the end of the explanation for
105 the patch, which certifies that you wrote it or otherwise have
106 the right to pass it on as a open-source patch. The rules are
107 pretty simple: if you can certify the below:
108
109 Developer's Certificate of Origin 1.1
110
111 By making a contribution to this project, I certify that:
112
113 (a) The contribution was created in whole or in part by me and I
114 have the right to submit it under the open source license
115 indicated in the file; or
116
117 (b) The contribution is based upon previous work that, to the best
118 of my knowledge, is covered under an appropriate open source
119 license and I have the right under that license to submit that
120 work with modifications, whether created in whole or in part
121 by me, under the same open source license (unless I am
122 permitted to submit under a different license), as indicated
123 in the file; or
124
125 (c) The contribution was provided directly to me by some other
126 person who certified (a), (b) or (c) and I have not modified
127 it.
128
129 (d) I understand and agree that this project and the contribution
130 are public and that a record of the contribution (including all
131 personal information I submit with it, including my sign-off) is
132 maintained indefinitely and may be redistributed consistent with
133 this project or the open source license(s) involved.
134
135 then you just add a line saying
136
137 Signed-off-by: Random J Developer <random@developer.example.org>
138
139 Some people also put extra tags at the end. They'll just be ignored for
140 now, but you can do this to mark internal company procedures or just
141 point out some special detail about the sign-off.
142
143
144 ------------------------------------------------
145 MUA specific hints
146
147 Some of patches I receive or pick up from the list share common
148 patterns of breakage. Please make sure your MUA is set up
149 properly not to corrupt whitespaces. Here are two common ones
150 I have seen:
151
152 * Empty context lines that do not have _any_ whitespace.
153
154 * Non empty context lines that have one extra whitespace at the
155 beginning.
156
157 One test you could do yourself if your MUA is set up correctly is:
158
159 * Send the patch to yourself, exactly the way you would, except
160 To: and Cc: lines, which would not contain the list and
161 maintainer address.
162
163 * Save that patch to a file in UNIX mailbox format. Call it say
164 a.patch.
165
166 * Try to apply to the tip of the "master" branch from the
167 git.git public repository:
168
169 $ git fetch http://kernel.org/pub/scm/git/git.git master:test-apply
170 $ git checkout test-apply
171 $ git reset --hard
172 $ git applymbox a.patch
173
174 If it does not apply correctly, there can be various reasons.
175
176 * Your patch itself does not apply cleanly. That is _bad_ but
177 does not have much to do with your MUA. Please rebase the
178 patch appropriately.
179
180 * Your MUA corrupted your patch; applymbox would complain that
181 the patch does not apply. Look at .dotest/ subdirectory and
182 see what 'patch' file contains and check for the common
183 corruption patterns mentioned above.
184
185 * While you are at it, check what are in 'info' and
186 'final-commit' files as well. If what is in 'final-commit' is
187 not exactly what you would want to see in the commit log
188 message, it is very likely that your maintainer would end up
189 hand editing the log message when he applies your patch.
190 Things like "Hi, this is my first patch.\n", if you really
191 want to put in the patch e-mail, should come after the
192 three-dash line that signals the end of the commit message.
193
194
195 Pine
196 ----
197
198 (Johannes Schindelin)
199
200 I don't know how many people still use pine, but for those poor
201 souls it may be good to mention that the quell-flowed-text is
202 needed for recent versions.
203
204 ... the "no-strip-whitespace-before-send" option, too. AFAIK it
205 was introduced in 4.60.
206
207 (Linus Torvalds)
208
209 And 4.58 needs at least this.
210
211 ---
212 diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1)
213 Author: Linus Torvalds <torvalds@g5.osdl.org>
214 Date: Mon Aug 15 17:23:51 2005 -0700
215
216 Fix pine whitespace-corruption bug
217
218 There's no excuse for unconditionally removing whitespace from
219 the pico buffers on close.
220
221 diff --git a/pico/pico.c b/pico/pico.c
222 --- a/pico/pico.c
223 +++ b/pico/pico.c
224 @@ -219,7 +219,9 @@ PICO *pm;
225 switch(pico_all_done){ /* prepare for/handle final events */
226 case COMP_EXIT : /* already confirmed */
227 packheader();
228 +#if 0
229 stripwhitespace();
230 +#endif
231 c |= COMP_EXIT;
232 break;
233
234
235 (Daniel Barkalow)
236
237 > A patch to SubmittingPatches, MUA specific help section for
238 > users of Pine 4.63 would be very much appreciated.
239
240 Ah, it looks like a recent version changed the default behavior to do the
241 right thing, and inverted the sense of the configuration option. (Either
242 that or Gentoo did it.) So you need to set the
243 "no-strip-whitespace-before-send" option, unless the option you have is
244 "strip-whitespace-before-send", in which case you should avoid checking
245 it.
246
247
248 Thunderbird
249 -----------
250
251 (A Large Angry SCM)
252
253 Here are some hints on how to successfully submit patches inline using
254 Thunderbird.
255
256 This recipe appears to work with the current [*1*] Thunderbird from Suse.
257
258 The following Thunderbird extensions are needed:
259 AboutConfig 0.5
260 http://aboutconfig.mozdev.org/
261 External Editor 0.5.4
262 http://extensionroom.mozdev.org/more-info/exteditor
263
264 1) Prepare the patch as a text file using your method of choice.
265
266 2) Before opening a compose window, use Edit->Account Settings to
267 uncheck the "Compose messages in HTML format" setting in the
268 "Composition & Addressing" panel of the account to be used to send the
269 patch. [*2*]
270
271 3) In the main Thunderbird window, _before_ you open the compose window
272 for the patch, use Tools->about:config to set the following to the
273 indicated values:
274 mailnews.send_plaintext_flowed => false
275 mailnews.wraplength => 0
276
277 4) Open a compose window and click the external editor icon.
278
279 5) In the external editor window, read in the patch file and exit the
280 editor normally.
281
282 6) Back in the compose window: Add whatever other text you wish to the
283 message, complete the addressing and subject fields, and press send.
284
285 7) Optionally, undo the about:config/account settings changes made in
286 steps 2 & 3.
287
288
289 [Footnotes]
290 *1* Version 1.0 (20041207) from the MozillaThunderbird-1.0-5 rpm of Suse
291 9.3 professional updates.
292
293 *2* It may be possible to do this with about:config and the following
294 settings but I haven't tried, yet.
295 mail.html_compose => false
296 mail.identity.default.compose_html => false
297 mail.identity.id?.compose_html => false
298