Merge branch 'mm/maint-config-explicit-bool-display' into maint
[git/git.git] / Documentation / SubmittingPatches
1 Checklist (and a short version for the impatient):
3 Commits:
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 - the first line of the commit message should be a short
10 description (50 characters is the soft limit, see DISCUSSION
11 in git-commit(1)), and should skip the full stop
12 - the body should provide a meaningful commit message, which:
13 . explains the problem the change tries to solve, iow, what
14 is wrong with the current code without the change.
15 . justifies the way the change solves the problem, iow, why
16 the result with the change is better.
17 . alternate solutions considered but discarded, if any.
18 - describe changes in imperative mood, e.g. "make xyzzy do frotz"
19 instead of "[This patch] makes xyzzy do frotz" or "[I] changed
20 xyzzy to do frotz", as if you are giving orders to the codebase
21 to change its behaviour.
22 - try to make sure your explanation can be understood without
23 external resources. Instead of giving a URL to a mailing list
24 archive, summarize the relevant points of the discussion.
25 - add a "Signed-off-by: Your Name <>" line to the
26 commit message (or just use the option "-s" when committing)
27 to confirm that you agree to the Developer's Certificate of Origin
28 - make sure that you have tests for the bug you are fixing
29 - make sure that the test suite passes after your commit
31 Patch:
33 - use "git format-patch -M" to create the patch
34 - do not PGP sign your patch
35 - do not attach your patch, but read in the mail
36 body, unless you cannot teach your mailer to
37 leave the formatting of the patch alone.
38 - be careful doing cut & paste into your mailer, not to
39 corrupt whitespaces.
40 - provide additional information (which is unsuitable for
41 the commit message) between the "---" and the diffstat
42 - if you change, add, or remove a command line option or
43 make some other user interface change, the associated
44 documentation should be updated as well.
45 - if your name is not writable in ASCII, make sure that
46 you send off a message in the correct encoding.
47 - send the patch to the list ( and the
48 maintainer ( if (and only if) the patch
49 is ready for inclusion. If you use git-send-email(1),
50 please test it first by sending email to yourself.
51 - see below for instructions specific to your mailer
53 Long version:
55 I started reading over the SubmittingPatches document for Linux
56 kernel, primarily because I wanted to have a document similar to
57 it for the core GIT to make sure people understand what they are
58 doing when they write "Signed-off-by" line.
60 But the patch submission requirements are a lot more relaxed
61 here on the technical/contents front, because the core GIT is
62 thousand times smaller ;-). So here is only the relevant bits.
64 (0) Decide what to base your work on.
66 In general, always base your work on the oldest branch that your
67 change is relevant to.
69 - A bugfix should be based on 'maint' in general. If the bug is not
70 present in 'maint', base it on 'master'. For a bug that's not yet
71 in 'master', find the topic that introduces the regression, and
72 base your work on the tip of the topic.
74 - A new feature should be based on 'master' in general. If the new
75 feature depends on a topic that is in 'pu', but not in 'master',
76 base your work on the tip of that topic.
78 - Corrections and enhancements to a topic not yet in 'master' should
79 be based on the tip of that topic. If the topic has not been merged
80 to 'next', it's alright to add a note to squash minor corrections
81 into the series.
83 - In the exceptional case that a new feature depends on several topics
84 not in 'master', start working on 'next' or 'pu' privately and send
85 out patches for discussion. Before the final merge, you may have to
86 wait until some of the dependent topics graduate to 'master', and
87 rebase your work.
89 To find the tip of a topic branch, run "git log --first-parent
90 master..pu" and look for the merge commit. The second parent of this
91 commit is the tip of the topic branch.
93 (1) Make separate commits for logically separate changes.
95 Unless your patch is really trivial, you should not be sending
96 out a patch that was generated between your working tree and
97 your commit head. Instead, always make a commit with complete
98 commit message and generate a series of patches from your
99 repository. It is a good discipline.
101 Give an explanation for the change(s) that is detailed enough so
102 that people can judge if it is good thing to do, without reading
103 the actual patch text to determine how well the code does what
104 the explanation promises to do.
106 If your description starts to get too long, that's a sign that you
107 probably need to split up your commit to finer grained pieces.
108 That being said, patches which plainly describe the things that
109 help reviewers check the patch, and future maintainers understand
110 the code, are the most beautiful patches. Descriptions that summarise
111 the point in the subject well, and describe the motivation for the
112 change, the approach taken by the change, and if relevant how this
113 differs substantially from the prior version, are all good things
114 to have.
116 Oh, another thing. I am picky about whitespaces. Make sure your
117 changes do not trigger errors with the sample pre-commit hook shipped
118 in templates/hooks--pre-commit. To help ensure this does not happen,
119 run git diff --check on your changes before you commit.
122 (1a) Try to be nice to older C compilers
124 We try to support a wide range of C compilers to compile
125 git with. That means that you should not use C99 initializers, even
126 if a lot of compilers grok it.
128 Also, variables have to be declared at the beginning of the block
129 (you can check this with gcc, using the -Wdeclaration-after-statement
130 option).
132 Another thing: NULL pointers shall be written as NULL, not as 0.
135 (2) Generate your patch using git tools out of your commits.
137 git based diff tools generate unidiff which is the preferred format.
139 You do not have to be afraid to use -M option to "git diff" or
140 "git format-patch", if your patch involves file renames. The
141 receiving end can handle them just fine.
143 Please make sure your patch does not include any extra files
144 which do not belong in a patch submission. Make sure to review
145 your patch after generating it, to ensure accuracy. Before
146 sending out, please make sure it cleanly applies to the "master"
147 branch head. If you are preparing a work based on "next" branch,
148 that is fine, but please mark it as such.
151 (3) Sending your patches.
153 People on the git mailing list need to be able to read and
154 comment on the changes you are submitting. It is important for
155 a developer to be able to "quote" your changes, using standard
156 e-mail tools, so that they may comment on specific portions of
157 your code. For this reason, all patches should be submitted
158 "inline". WARNING: Be wary of your MUAs word-wrap
159 corrupting your patch. Do not cut-n-paste your patch; you can
160 lose tabs that way if you are not careful.
162 It is a common convention to prefix your subject line with
163 [PATCH]. This lets people easily distinguish patches from other
164 e-mail discussions. Use of additional markers after PATCH and
165 the closing bracket to mark the nature of the patch is also
166 encouraged. E.g. [PATCH/RFC] is often used when the patch is
167 not ready to be applied but it is for discussion, [PATCH v2],
168 [PATCH v3] etc. are often seen when you are sending an update to
169 what you have previously sent.
171 "git format-patch" command follows the best current practice to
172 format the body of an e-mail message. At the beginning of the
173 patch should come your commit message, ending with the
174 Signed-off-by: lines, and a line that consists of three dashes,
175 followed by the diffstat information and the patch itself. If
176 you are forwarding a patch from somebody else, optionally, at
177 the beginning of the e-mail message just before the commit
178 message starts, you can put a "From: " line to name that person.
180 You often want to add additional explanation about the patch,
181 other than the commit message itself. Place such "cover letter"
182 material between the three dash lines and the diffstat.
184 Do not attach the patch as a MIME attachment, compressed or not.
185 Do not let your e-mail client send quoted-printable. Do not let
186 your e-mail client send format=flowed which would destroy
187 whitespaces in your patches. Many
188 popular e-mail applications will not always transmit a MIME
189 attachment as plain text, making it impossible to comment on
190 your code. A MIME attachment also takes a bit more time to
191 process. This does not decrease the likelihood of your
192 MIME-attached change being accepted, but it makes it more likely
193 that it will be postponed.
195 Exception: If your mailer is mangling patches then someone may ask
196 you to re-send them using MIME, that is OK.
198 Do not PGP sign your patch, at least for now. Most likely, your
199 maintainer or other people on the list would not have your PGP
200 key and would not bother obtaining it anyway. Your patch is not
201 judged by who you are; a good patch from an unknown origin has a
202 far better chance of being accepted than a patch from a known,
203 respected origin that is done poorly or does incorrect things.
205 If you really really really really want to do a PGP signed
206 patch, format it as "multipart/signed", not a text/plain message
207 that starts with '-----BEGIN PGP SIGNED MESSAGE-----'. That is
208 not a text/plain, it's something else.
210 Unless your patch is a very trivial and an obviously correct one,
211 first send it with "To:" set to the mailing list, with "cc:" listing
212 people who are involved in the area you are touching (the output from
213 "git blame $path" and "git shortlog --no-merges $path" would help to
214 identify them), to solicit comments and reviews. After the list
215 reached a consensus that it is a good idea to apply the patch, re-send
216 it with "To:" set to the maintainer and optionally "cc:" the list for
217 inclusion. Do not forget to add trailers such as "Acked-by:",
218 "Reviewed-by:" and "Tested-by:" after your "Signed-off-by:" line as
219 necessary.
222 (4) Sign your work
224 To improve tracking of who did what, we've borrowed the
225 "sign-off" procedure from the Linux kernel project on patches
226 that are being emailed around. Although core GIT is a lot
227 smaller project it is a good discipline to follow it.
229 The sign-off is a simple line at the end of the explanation for
230 the patch, which certifies that you wrote it or otherwise have
231 the right to pass it on as a open-source patch. The rules are
232 pretty simple: if you can certify the below:
234 Developer's Certificate of Origin 1.1
236 By making a contribution to this project, I certify that:
238 (a) The contribution was created in whole or in part by me and I
239 have the right to submit it under the open source license
240 indicated in the file; or
242 (b) The contribution is based upon previous work that, to the best
243 of my knowledge, is covered under an appropriate open source
244 license and I have the right under that license to submit that
245 work with modifications, whether created in whole or in part
246 by me, under the same open source license (unless I am
247 permitted to submit under a different license), as indicated
248 in the file; or
250 (c) The contribution was provided directly to me by some other
251 person who certified (a), (b) or (c) and I have not modified
252 it.
254 (d) I understand and agree that this project and the contribution
255 are public and that a record of the contribution (including all
256 personal information I submit with it, including my sign-off) is
257 maintained indefinitely and may be redistributed consistent with
258 this project or the open source license(s) involved.
260 then you just add a line saying
262 Signed-off-by: Random J Developer <>
264 This line can be automatically added by git if you run the git-commit
265 command with the -s option.
267 Notice that you can place your own Signed-off-by: line when
268 forwarding somebody else's patch with the above rules for
269 D-C-O. Indeed you are encouraged to do so. Do not forget to
270 place an in-body "From: " line at the beginning to properly attribute
271 the change to its true author (see (2) above).
273 Also notice that a real name is used in the Signed-off-by: line. Please
274 don't hide your real name.
276 If you like, you can put extra tags at the end:
278 1. "Reported-by:" is used to credit someone who found the bug that
279 the patch attempts to fix.
280 2. "Acked-by:" says that the person who is more familiar with the area
281 the patch attempts to modify liked the patch.
282 3. "Reviewed-by:", unlike the other tags, can only be offered by the
283 reviewer and means that she is completely satisfied that the patch
284 is ready for application. It is usually offered only after a
285 detailed review.
286 4. "Tested-by:" is used to indicate that the person applied the patch
287 and found it to have the desired effect.
289 You can also create your own tag or use one that's in common usage
290 such as "Thanks-to:", "Based-on-patch-by:", or "Mentored-by:".
292 ------------------------------------------------
293 An ideal patch flow
295 Here is an ideal patch flow for this project the current maintainer
296 suggests to the contributors:
298 (0) You come up with an itch. You code it up.
300 (1) Send it to the list and cc people who may need to know about
301 the change.
303 The people who may need to know are the ones whose code you
304 are butchering. These people happen to be the ones who are
305 most likely to be knowledgeable enough to help you, but
306 they have no obligation to help you (i.e. you ask for help,
307 don't demand). "git log -p -- $area_you_are_modifying" would
308 help you find out who they are.
310 (2) You get comments and suggestions for improvements. You may
311 even get them in a "on top of your change" patch form.
313 (3) Polish, refine, and re-send to the list and the people who
314 spend their time to improve your patch. Go back to step (2).
316 (4) The list forms consensus that the last round of your patch is
317 good. Send it to the list and cc the maintainer.
319 (5) A topic branch is created with the patch and is merged to 'next',
320 and cooked further and eventually graduates to 'master'.
322 In any time between the (2)-(3) cycle, the maintainer may pick it up
323 from the list and queue it to 'pu', in order to make it easier for
324 people play with it without having to pick up and apply the patch to
325 their trees themselves.
327 ------------------------------------------------
328 Know the status of your patch after submission
330 * You can use Git itself to find out when your patch is merged in
331 master. 'git pull --rebase' will automatically skip already-applied
332 patches, and will let you know. This works only if you rebase on top
333 of the branch in which your patch has been merged (i.e. it will not
334 tell you if your patch is merged in pu if you rebase on top of
335 master).
337 * Read the git mailing list, the maintainer regularly posts messages
338 entitled "What's cooking in git.git" and "What's in git.git" giving
339 the status of various proposed changes.
341 ------------------------------------------------
342 MUA specific hints
344 Some of patches I receive or pick up from the list share common
345 patterns of breakage. Please make sure your MUA is set up
346 properly not to corrupt whitespaces.
348 See the DISCUSSION section of git-format-patch(1) for hints on
349 checking your patch by mailing it to yourself and applying with
350 git-am(1).
352 While you are at it, check the resulting commit log message from
353 a trial run of applying the patch. If what is in the resulting
354 commit is not exactly what you would want to see, it is very
355 likely that your maintainer would end up hand editing the log
356 message when he applies your patch. Things like "Hi, this is my
357 first patch.\n", if you really want to put in the patch e-mail,
358 should come after the three-dash line that signals the end of the
359 commit message.
362 Pine
363 ----
365 (Johannes Schindelin)
367 I don't know how many people still use pine, but for those poor
368 souls it may be good to mention that the quell-flowed-text is
369 needed for recent versions.
371 ... the "no-strip-whitespace-before-send" option, too. AFAIK it
372 was introduced in 4.60.
374 (Linus Torvalds)
376 And 4.58 needs at least this.
378 ---
379 diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1)
380 Author: Linus Torvalds <>
381 Date: Mon Aug 15 17:23:51 2005 -0700
383 Fix pine whitespace-corruption bug
385 There's no excuse for unconditionally removing whitespace from
386 the pico buffers on close.
388 diff --git a/pico/pico.c b/pico/pico.c
389 --- a/pico/pico.c
390 +++ b/pico/pico.c
391 @@ -219,7 +219,9 @@ PICO *pm;
392 switch(pico_all_done){ /* prepare for/handle final events */
393 case COMP_EXIT : /* already confirmed */
394 packheader();
395 +#if 0
396 stripwhitespace();
397 +#endif
398 c |= COMP_EXIT;
399 break;
402 (Daniel Barkalow)
404 > A patch to SubmittingPatches, MUA specific help section for
405 > users of Pine 4.63 would be very much appreciated.
407 Ah, it looks like a recent version changed the default behavior to do the
408 right thing, and inverted the sense of the configuration option. (Either
409 that or Gentoo did it.) So you need to set the
410 "no-strip-whitespace-before-send" option, unless the option you have is
411 "strip-whitespace-before-send", in which case you should avoid checking
412 it.
415 Thunderbird, KMail, GMail
416 -------------------------
418 See the MUA-SPECIFIC HINTS section of git-format-patch(1).
420 Gnus
421 ----
423 '|' in the *Summary* buffer can be used to pipe the current
424 message to an external program, and this is a handy way to drive
425 "git am". However, if the message is MIME encoded, what is
426 piped into the program is the representation you see in your
427 *Article* buffer after unwrapping MIME. This is often not what
428 you would want for two reasons. It tends to screw up non ASCII
429 characters (most notably in people's names), and also
430 whitespaces (fatal in patches). Running 'C-u g' to display the
431 message in raw form before using '|' to run the pipe can work
432 this problem around.