Clarify SubmittingPatches Checklist
[git/git.git] / Documentation / SubmittingPatches
1 Checklist (and a short version for the impatient):
2
3 Commits:
4
5 - make commits of logical units
6 - check for unnecessary whitespace with "git diff --check"
7 before committing
8 - do not check in commented out code or unneeded files
9 - provide a meaningful commit message
10 - the first line of the commit message should be a short
11 description and should skip the full stop
12 - if you want your work included in git.git, add a
13 "Signed-off-by: Your Name <your@email.com>" line to the
14 commit message (or just use the option "-s" when
15 committing) to confirm that you agree to the Developer's
16 Certificate of Origin
17
18 Patch:
19
20 - use "git format-patch -M" to create the patch
21 - send your patch to <git@vger.kernel.org>. If you use
22 git-send-email(1), please test it first by sending
23 email to yourself.
24 - do not PGP sign your patch
25 - do not attach your patch, but read in the mail
26 body, unless you cannot teach your mailer to
27 leave the formatting of the patch alone.
28 - be careful doing cut & paste into your mailer, not to
29 corrupt whitespaces.
30 - provide additional information (which is unsuitable for
31 the commit message) between the "---" and the diffstat
32 - send the patch to the list _and_ the maintainer
33
34 Long version:
35
36 I started reading over the SubmittingPatches document for Linux
37 kernel, primarily because I wanted to have a document similar to
38 it for the core GIT to make sure people understand what they are
39 doing when they write "Signed-off-by" line.
40
41 But the patch submission requirements are a lot more relaxed
42 here on the technical/contents front, because the core GIT is
43 thousand times smaller ;-). So here is only the relevant bits.
44
45
46 (1) Make separate commits for logically separate changes.
47
48 Unless your patch is really trivial, you should not be sending
49 out a patch that was generated between your working tree and
50 your commit head. Instead, always make a commit with complete
51 commit message and generate a series of patches from your
52 repository. It is a good discipline.
53
54 Describe the technical detail of the change(s).
55
56 If your description starts to get too long, that's a sign that you
57 probably need to split up your commit to finer grained pieces.
58
59 Oh, another thing. I am picky about whitespaces. Make sure your
60 changes do not trigger errors with the sample pre-commit hook shipped
61 in templates/hooks--pre-commit. To help ensure this does not happen,
62 run git diff --check on your changes before you commit.
63
64
65 (2) Generate your patch using git tools out of your commits.
66
67 git based diff tools (git, Cogito, and StGIT included) generate
68 unidiff which is the preferred format.
69
70 You do not have to be afraid to use -M option to "git diff" or
71 "git format-patch", if your patch involves file renames. The
72 receiving end can handle them just fine.
73
74 Please make sure your patch does not include any extra files
75 which do not belong in a patch submission. Make sure to review
76 your patch after generating it, to ensure accuracy. Before
77 sending out, please make sure it cleanly applies to the "master"
78 branch head. If you are preparing a work based on "next" branch,
79 that is fine, but please mark it as such.
80
81
82 (3) Sending your patches.
83
84 People on the git mailing list need to be able to read and
85 comment on the changes you are submitting. It is important for
86 a developer to be able to "quote" your changes, using standard
87 e-mail tools, so that they may comment on specific portions of
88 your code. For this reason, all patches should be submitted
89 "inline". WARNING: Be wary of your MUAs word-wrap
90 corrupting your patch. Do not cut-n-paste your patch; you can
91 lose tabs that way if you are not careful.
92
93 It is a common convention to prefix your subject line with
94 [PATCH]. This lets people easily distinguish patches from other
95 e-mail discussions.
96
97 "git format-patch" command follows the best current practice to
98 format the body of an e-mail message. At the beginning of the
99 patch should come your commit message, ending with the
100 Signed-off-by: lines, and a line that consists of three dashes,
101 followed by the diffstat information and the patch itself. If
102 you are forwarding a patch from somebody else, optionally, at
103 the beginning of the e-mail message just before the commit
104 message starts, you can put a "From: " line to name that person.
105
106 You often want to add additional explanation about the patch,
107 other than the commit message itself. Place such "cover letter"
108 material between the three dash lines and the diffstat.
109
110 Do not attach the patch as a MIME attachment, compressed or not.
111 Do not let your e-mail client send quoted-printable. Do not let
112 your e-mail client send format=flowed which would destroy
113 whitespaces in your patches. Many
114 popular e-mail applications will not always transmit a MIME
115 attachment as plain text, making it impossible to comment on
116 your code. A MIME attachment also takes a bit more time to
117 process. This does not decrease the likelihood of your
118 MIME-attached change being accepted, but it makes it more likely
119 that it will be postponed.
120
121 Exception: If your mailer is mangling patches then someone may ask
122 you to re-send them using MIME, that is OK.
123
124 Do not PGP sign your patch, at least for now. Most likely, your
125 maintainer or other people on the list would not have your PGP
126 key and would not bother obtaining it anyway. Your patch is not
127 judged by who you are; a good patch from an unknown origin has a
128 far better chance of being accepted than a patch from a known,
129 respected origin that is done poorly or does incorrect things.
130
131 If you really really really really want to do a PGP signed
132 patch, format it as "multipart/signed", not a text/plain message
133 that starts with '-----BEGIN PGP SIGNED MESSAGE-----'. That is
134 not a text/plain, it's something else.
135
136 Note that your maintainer does not necessarily read everything
137 on the git mailing list. If your patch is for discussion first,
138 send it "To:" the mailing list, and optionally "cc:" him. If it
139 is trivially correct or after the list reached a consensus, send
140 it "To:" the maintainer and optionally "cc:" the list.
141
142 Also note that your maintainer does not actively involve himself in
143 maintaining what are in contrib/ hierarchy. When you send fixes and
144 enhancements to them, do not forget to "cc: " the person who primarily
145 worked on that hierarchy in contrib/.
146
147
148 (4) Sign your work
149
150 To improve tracking of who did what, we've borrowed the
151 "sign-off" procedure from the Linux kernel project on patches
152 that are being emailed around. Although core GIT is a lot
153 smaller project it is a good discipline to follow it.
154
155 The sign-off is a simple line at the end of the explanation for
156 the patch, which certifies that you wrote it or otherwise have
157 the right to pass it on as a open-source patch. The rules are
158 pretty simple: if you can certify the below:
159
160 Developer's Certificate of Origin 1.1
161
162 By making a contribution to this project, I certify that:
163
164 (a) The contribution was created in whole or in part by me and I
165 have the right to submit it under the open source license
166 indicated in the file; or
167
168 (b) The contribution is based upon previous work that, to the best
169 of my knowledge, is covered under an appropriate open source
170 license and I have the right under that license to submit that
171 work with modifications, whether created in whole or in part
172 by me, under the same open source license (unless I am
173 permitted to submit under a different license), as indicated
174 in the file; or
175
176 (c) The contribution was provided directly to me by some other
177 person who certified (a), (b) or (c) and I have not modified
178 it.
179
180 (d) I understand and agree that this project and the contribution
181 are public and that a record of the contribution (including all
182 personal information I submit with it, including my sign-off) is
183 maintained indefinitely and may be redistributed consistent with
184 this project or the open source license(s) involved.
185
186 then you just add a line saying
187
188 Signed-off-by: Random J Developer <random@developer.example.org>
189
190 This line can be automatically added by git if you run the git-commit
191 command with the -s option.
192
193 Some people also put extra tags at the end. They'll just be ignored for
194 now, but you can do this to mark internal company procedures or just
195 point out some special detail about the sign-off.
196
197
198 ------------------------------------------------
199 MUA specific hints
200
201 Some of patches I receive or pick up from the list share common
202 patterns of breakage. Please make sure your MUA is set up
203 properly not to corrupt whitespaces. Here are two common ones
204 I have seen:
205
206 * Empty context lines that do not have _any_ whitespace.
207
208 * Non empty context lines that have one extra whitespace at the
209 beginning.
210
211 One test you could do yourself if your MUA is set up correctly is:
212
213 * Send the patch to yourself, exactly the way you would, except
214 To: and Cc: lines, which would not contain the list and
215 maintainer address.
216
217 * Save that patch to a file in UNIX mailbox format. Call it say
218 a.patch.
219
220 * Try to apply to the tip of the "master" branch from the
221 git.git public repository:
222
223 $ git fetch http://kernel.org/pub/scm/git/git.git master:test-apply
224 $ git checkout test-apply
225 $ git reset --hard
226 $ git applymbox a.patch
227
228 If it does not apply correctly, there can be various reasons.
229
230 * Your patch itself does not apply cleanly. That is _bad_ but
231 does not have much to do with your MUA. Please rebase the
232 patch appropriately.
233
234 * Your MUA corrupted your patch; applymbox would complain that
235 the patch does not apply. Look at .dotest/ subdirectory and
236 see what 'patch' file contains and check for the common
237 corruption patterns mentioned above.
238
239 * While you are at it, check what are in 'info' and
240 'final-commit' files as well. If what is in 'final-commit' is
241 not exactly what you would want to see in the commit log
242 message, it is very likely that your maintainer would end up
243 hand editing the log message when he applies your patch.
244 Things like "Hi, this is my first patch.\n", if you really
245 want to put in the patch e-mail, should come after the
246 three-dash line that signals the end of the commit message.
247
248
249 Pine
250 ----
251
252 (Johannes Schindelin)
253
254 I don't know how many people still use pine, but for those poor
255 souls it may be good to mention that the quell-flowed-text is
256 needed for recent versions.
257
258 ... the "no-strip-whitespace-before-send" option, too. AFAIK it
259 was introduced in 4.60.
260
261 (Linus Torvalds)
262
263 And 4.58 needs at least this.
264
265 ---
266 diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1)
267 Author: Linus Torvalds <torvalds@g5.osdl.org>
268 Date: Mon Aug 15 17:23:51 2005 -0700
269
270 Fix pine whitespace-corruption bug
271
272 There's no excuse for unconditionally removing whitespace from
273 the pico buffers on close.
274
275 diff --git a/pico/pico.c b/pico/pico.c
276 --- a/pico/pico.c
277 +++ b/pico/pico.c
278 @@ -219,7 +219,9 @@ PICO *pm;
279 switch(pico_all_done){ /* prepare for/handle final events */
280 case COMP_EXIT : /* already confirmed */
281 packheader();
282 +#if 0
283 stripwhitespace();
284 +#endif
285 c |= COMP_EXIT;
286 break;
287
288
289 (Daniel Barkalow)
290
291 > A patch to SubmittingPatches, MUA specific help section for
292 > users of Pine 4.63 would be very much appreciated.
293
294 Ah, it looks like a recent version changed the default behavior to do the
295 right thing, and inverted the sense of the configuration option. (Either
296 that or Gentoo did it.) So you need to set the
297 "no-strip-whitespace-before-send" option, unless the option you have is
298 "strip-whitespace-before-send", in which case you should avoid checking
299 it.
300
301
302 Thunderbird
303 -----------
304
305 (A Large Angry SCM)
306
307 Here are some hints on how to successfully submit patches inline using
308 Thunderbird.
309
310 This recipe appears to work with the current [*1*] Thunderbird from Suse.
311
312 The following Thunderbird extensions are needed:
313 AboutConfig 0.5
314 http://aboutconfig.mozdev.org/
315 External Editor 0.7.2
316 http://globs.org/articles.php?lng=en&pg=8
317
318 1) Prepare the patch as a text file using your method of choice.
319
320 2) Before opening a compose window, use Edit->Account Settings to
321 uncheck the "Compose messages in HTML format" setting in the
322 "Composition & Addressing" panel of the account to be used to send the
323 patch. [*2*]
324
325 3) In the main Thunderbird window, _before_ you open the compose window
326 for the patch, use Tools->about:config to set the following to the
327 indicated values:
328 mailnews.send_plaintext_flowed => false
329 mailnews.wraplength => 0
330
331 4) Open a compose window and click the external editor icon.
332
333 5) In the external editor window, read in the patch file and exit the
334 editor normally.
335
336 6) Back in the compose window: Add whatever other text you wish to the
337 message, complete the addressing and subject fields, and press send.
338
339 7) Optionally, undo the about:config/account settings changes made in
340 steps 2 & 3.
341
342
343 [Footnotes]
344 *1* Version 1.0 (20041207) from the MozillaThunderbird-1.0-5 rpm of Suse
345 9.3 professional updates.
346
347 *2* It may be possible to do this with about:config and the following
348 settings but I haven't tried, yet.
349 mail.html_compose => false
350 mail.identity.default.compose_html => false
351 mail.identity.id?.compose_html => false
352
353
354 Gnus
355 ----
356
357 '|' in the *Summary* buffer can be used to pipe the current
358 message to an external program, and this is a handy way to drive
359 "git am". However, if the message is MIME encoded, what is
360 piped into the program is the representation you see in your
361 *Article* buffer after unwrapping MIME. This is often not what
362 you would want for two reasons. It tends to screw up non ASCII
363 characters (most notably in people's names), and also
364 whitespaces (fatal in patches). Running 'C-u g' to display the
365 message in raw form before using '|' to run the pipe can work
366 this problem around.
367
368
369 KMail
370 -----
371
372 This should help you to submit patches inline using KMail.
373
374 1) Prepare the patch as a text file.
375
376 2) Click on New Mail.
377
378 3) Go under "Options" in the Composer window and be sure that
379 "Word wrap" is not set.
380
381 4) Use Message -> Insert file... and insert the patch.
382
383 5) Back in the compose window: add whatever other text you wish to the
384 message, complete the addressing and subject fields, and press send.