Merge branch 'sv/submitting-final-patch'
[git/git.git] / Documentation / SubmittingPatches
CommitLineData
adcc42e6
JH
1Here are some guidelines for people who want to contribute their code
2to this software.
31408251 3
d0c26f0f
RR
4(0) Decide what to base your work on.
5
6In general, always base your work on the oldest branch that your
7change is relevant to.
8
9 - A bugfix should be based on 'maint' in general. If the bug is not
10 present in 'maint', base it on 'master'. For a bug that's not yet
11 in 'master', find the topic that introduces the regression, and
12 base your work on the tip of the topic.
13
14 - A new feature should be based on 'master' in general. If the new
15 feature depends on a topic that is in 'pu', but not in 'master',
16 base your work on the tip of that topic.
17
18 - Corrections and enhancements to a topic not yet in 'master' should
19 be based on the tip of that topic. If the topic has not been merged
20 to 'next', it's alright to add a note to squash minor corrections
21 into the series.
22
23 - In the exceptional case that a new feature depends on several topics
24 not in 'master', start working on 'next' or 'pu' privately and send
25 out patches for discussion. Before the final merge, you may have to
26 wait until some of the dependent topics graduate to 'master', and
27 rebase your work.
28
e6da8ee8
JH
29 - Some parts of the system have dedicated maintainers with their own
30 repositories (see the section "Subsystems" below). Changes to
31 these parts should be based on their trees.
32
d0c26f0f
RR
33To find the tip of a topic branch, run "git log --first-parent
34master..pu" and look for the merge commit. The second parent of this
35commit is the tip of the topic branch.
31408251
JH
36
37(1) Make separate commits for logically separate changes.
38
39Unless your patch is really trivial, you should not be sending
40out a patch that was generated between your working tree and
41your commit head. Instead, always make a commit with complete
42commit message and generate a series of patches from your
43repository. It is a good discipline.
44
d0f7dcbf
JH
45Give an explanation for the change(s) that is detailed enough so
46that people can judge if it is good thing to do, without reading
47the actual patch text to determine how well the code does what
48the explanation promises to do.
31408251 49
45d2b286 50If your description starts to get too long, that's a sign that you
31408251 51probably need to split up your commit to finer grained pieces.
47afed5d
SV
52That being said, patches which plainly describe the things that
53help reviewers check the patch, and future maintainers understand
54the code, are the most beautiful patches. Descriptions that summarise
55the point in the subject well, and describe the motivation for the
56change, the approach taken by the change, and if relevant how this
d0f7dcbf
JH
57differs substantially from the prior version, are all good things
58to have.
31408251 59
7d5bf87b
JH
60Make sure that you have tests for the bug you are fixing.
61
62When adding a new feature, make sure that you have new tests to show
63the feature triggers the new behaviour when it should, and to show the
64feature does not trigger when it shouldn't. Also make sure that the
65test suite passes after your commit. Do not forget to update the
66documentation to describe the updated behaviour.
67
42e0fae9
MB
68Speaking of the documentation, it is currently a liberal mixture of US
69and UK English norms for spelling and grammar, which is somewhat
70unfortunate. A huge patch that touches the files all over the place
71only to correct the inconsistency is not welcome, though. Potential
72clashes with other changes that can result from such a patch are not
73worth it. We prefer to gradually reconcile the inconsistencies in
74favor of US English, with small and easily digestible patches, as a
75side effect of doing some other real work in the vicinity (e.g.
76rewriting a paragraph for clarity, while turning en_UK spelling to
77en_US). Obvious typographical fixes are much more welcomed ("teh ->
78"the"), preferably submitted as independent patches separate from
79other documentation changes.
80
81Oh, another thing. We are picky about whitespaces. Make sure your
45d2b286 82changes do not trigger errors with the sample pre-commit hook shipped
16507fcf
BL
83in templates/hooks--pre-commit. To help ensure this does not happen,
84run git diff --check on your changes before you commit.
31408251 85
31408251 86
7d5bf87b
JH
87(2) Describe your changes well.
88
89The first line of the commit message should be a short description (50
90characters is the soft limit, see DISCUSSION in git-commit(1)), and
91should skip the full stop. It is also conventional in most cases to
92prefix the first line with "area: " where the area is a filename or
93identifier for the general area of the code being modified, e.g.
94
95 . archive: ustar header checksum is computed unsigned
96 . git-cherry-pick.txt: clarify the use of revision range notation
97
98If in doubt which identifier to use, run "git log --no-merges" on the
99files you are modifying to see the current conventions.
100
101The body should provide a meaningful commit message, which:
102
103 . explains the problem the change tries to solve, iow, what is wrong
104 with the current code without the change.
105
106 . justifies the way the change solves the problem, iow, why the
107 result with the change is better.
108
109 . alternate solutions considered but discarded, if any.
110
111Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
112instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
113to do frotz", as if you are giving orders to the codebase to change
114its behaviour. Try to make sure your explanation can be understood
115without external resources. Instead of giving a URL to a mailing list
116archive, summarize the relevant points of the discussion.
117
118
2de9b711 119(3) Generate your patch using Git tools out of your commits.
45d2b286 120
2de9b711 121Git based diff tools generate unidiff which is the preferred format.
45d2b286 122
31408251
JH
123You do not have to be afraid to use -M option to "git diff" or
124"git format-patch", if your patch involves file renames. The
125receiving end can handle them just fine.
126
7d5bf87b
JH
127Please make sure your patch does not add commented out debugging code,
128or include any extra files which do not relate to what your patch
129is trying to achieve. Make sure to review
31408251
JH
130your patch after generating it, to ensure accuracy. Before
131sending out, please make sure it cleanly applies to the "master"
45d2b286
JH
132branch head. If you are preparing a work based on "next" branch,
133that is fine, but please mark it as such.
31408251
JH
134
135
7d5bf87b 136(4) Sending your patches.
31408251 137
2de9b711 138People on the Git mailing list need to be able to read and
31408251
JH
139comment on the changes you are submitting. It is important for
140a developer to be able to "quote" your changes, using standard
141e-mail tools, so that they may comment on specific portions of
eaa6c987
RS
142your code. For this reason, each patch should be submitted
143"inline" in a separate message.
144
145Multiple related patches should be grouped into their own e-mail
146thread to help readers find all parts of the series. To that end,
147send them as replies to either an additional "cover letter" message
148(see below), the first patch, or the respective preceding patch.
149
150If your log message (including your name on the
7d5bf87b
JH
151Signed-off-by line) is not writable in ASCII, make sure that
152you send off a message in the correct encoding.
153
154WARNING: Be wary of your MUAs word-wrap
45d2b286
JH
155corrupting your patch. Do not cut-n-paste your patch; you can
156lose tabs that way if you are not careful.
31408251 157
45d2b286 158It is a common convention to prefix your subject line with
31408251 159[PATCH]. This lets people easily distinguish patches from other
4e891acf
JH
160e-mail discussions. Use of additional markers after PATCH and
161the closing bracket to mark the nature of the patch is also
162encouraged. E.g. [PATCH/RFC] is often used when the patch is
163not ready to be applied but it is for discussion, [PATCH v2],
164[PATCH v3] etc. are often seen when you are sending an update to
165what you have previously sent.
31408251
JH
166
167"git format-patch" command follows the best current practice to
168format the body of an e-mail message. At the beginning of the
169patch should come your commit message, ending with the
170Signed-off-by: lines, and a line that consists of three dashes,
171followed by the diffstat information and the patch itself. If
172you are forwarding a patch from somebody else, optionally, at
173the beginning of the e-mail message just before the commit
174message starts, you can put a "From: " line to name that person.
175
176You often want to add additional explanation about the patch,
177other than the commit message itself. Place such "cover letter"
76323c67
PO
178material between the three dash lines and the diffstat. Git-notes
179can also be inserted using the `--notes` option.
31408251
JH
180
181Do not attach the patch as a MIME attachment, compressed or not.
e30b217b
JH
182Do not let your e-mail client send quoted-printable. Do not let
183your e-mail client send format=flowed which would destroy
184whitespaces in your patches. Many
31408251
JH
185popular e-mail applications will not always transmit a MIME
186attachment as plain text, making it impossible to comment on
187your code. A MIME attachment also takes a bit more time to
188process. This does not decrease the likelihood of your
189MIME-attached change being accepted, but it makes it more likely
190that it will be postponed.
191
192Exception: If your mailer is mangling patches then someone may ask
9847f7e0 193you to re-send them using MIME, that is OK.
31408251 194
9847f7e0
JH
195Do not PGP sign your patch, at least for now. Most likely, your
196maintainer or other people on the list would not have your PGP
197key and would not bother obtaining it anyway. Your patch is not
198judged by who you are; a good patch from an unknown origin has a
199far better chance of being accepted than a patch from a known,
200respected origin that is done poorly or does incorrect things.
201
202If you really really really really want to do a PGP signed
203patch, format it as "multipart/signed", not a text/plain message
204that starts with '-----BEGIN PGP SIGNED MESSAGE-----'. That is
205not a text/plain, it's something else.
206
7d5bf87b 207Send your patch with "To:" set to the mailing list, with "cc:" listing
d0c26f0f
RR
208people who are involved in the area you are touching (the output from
209"git blame $path" and "git shortlog --no-merges $path" would help to
7d5bf87b 210identify them), to solicit comments and reviews.
04d24455 211
7d5bf87b 212After the list reached a consensus that it is a good idea to apply the
92a865e7
JH
213patch, re-send it with "To:" set to the maintainer [*1*] and "cc:" the
214list [*2*] for inclusion.
31408251 215
7d5bf87b
JH
216Do not forget to add trailers such as "Acked-by:", "Reviewed-by:" and
217"Tested-by:" lines as necessary to credit people who helped your
218patch.
04d24455 219
92a865e7
JH
220 [Addresses]
221 *1* The current maintainer: gitster@pobox.com
222 *2* The mailing list: git@vger.kernel.org
223
31408251 224
7d5bf87b 225(5) Sign your work
31408251
JH
226
227To improve tracking of who did what, we've borrowed the
228"sign-off" procedure from the Linux kernel project on patches
48a8c26c 229that are being emailed around. Although core Git is a lot
31408251
JH
230smaller project it is a good discipline to follow it.
231
232The sign-off is a simple line at the end of the explanation for
233the patch, which certifies that you wrote it or otherwise have
234the right to pass it on as a open-source patch. The rules are
235pretty simple: if you can certify the below:
236
237 Developer's Certificate of Origin 1.1
238
239 By making a contribution to this project, I certify that:
240
241 (a) The contribution was created in whole or in part by me and I
242 have the right to submit it under the open source license
243 indicated in the file; or
244
245 (b) The contribution is based upon previous work that, to the best
246 of my knowledge, is covered under an appropriate open source
247 license and I have the right under that license to submit that
248 work with modifications, whether created in whole or in part
249 by me, under the same open source license (unless I am
250 permitted to submit under a different license), as indicated
251 in the file; or
252
253 (c) The contribution was provided directly to me by some other
254 person who certified (a), (b) or (c) and I have not modified
255 it.
256
257 (d) I understand and agree that this project and the contribution
258 are public and that a record of the contribution (including all
259 personal information I submit with it, including my sign-off) is
260 maintained indefinitely and may be redistributed consistent with
261 this project or the open source license(s) involved.
262
263then you just add a line saying
264
265 Signed-off-by: Random J Developer <random@developer.example.org>
266
2de9b711 267This line can be automatically added by Git if you run the git-commit
69945602
PC
268command with the -s option.
269
c11c3b56
JH
270Notice that you can place your own Signed-off-by: line when
271forwarding somebody else's patch with the above rules for
272D-C-O. Indeed you are encouraged to do so. Do not forget to
273place an in-body "From: " line at the beginning to properly attribute
274the change to its true author (see (2) above).
275
67275247
MV
276Also notice that a real name is used in the Signed-off-by: line. Please
277don't hide your real name.
278
95b7a41a
RR
279If you like, you can put extra tags at the end:
280
0353a0c4 2811. "Reported-by:" is used to credit someone who found the bug that
95b7a41a
RR
282 the patch attempts to fix.
2832. "Acked-by:" says that the person who is more familiar with the area
284 the patch attempts to modify liked the patch.
2853. "Reviewed-by:", unlike the other tags, can only be offered by the
286 reviewer and means that she is completely satisfied that the patch
287 is ready for application. It is usually offered only after a
288 detailed review.
2894. "Tested-by:" is used to indicate that the person applied the patch
290 and found it to have the desired effect.
291
292You can also create your own tag or use one that's in common usage
293such as "Thanks-to:", "Based-on-patch-by:", or "Mentored-by:".
9740d289 294
e6da8ee8
JH
295------------------------------------------------
296Subsystems with dedicated maintainers
297
298Some parts of the system have dedicated maintainers with their own
299repositories.
300
301 - git-gui/ comes from git-gui project, maintained by Pat Thoyts:
302
303 git://repo.or.cz/git-gui.git
304
305 - gitk-git/ comes from Paul Mackerras's gitk project:
306
307 git://ozlabs.org/~paulus/gitk
308
309 - po/ comes from the localization coordinator, Jiang Xin:
310
311 https://github.com/git-l10n/git-po/
312
313Patches to these parts should be based on their trees.
314
a941fb4a
JH
315------------------------------------------------
316An ideal patch flow
317
318Here is an ideal patch flow for this project the current maintainer
319suggests to the contributors:
320
321 (0) You come up with an itch. You code it up.
322
323 (1) Send it to the list and cc people who may need to know about
324 the change.
325
326 The people who may need to know are the ones whose code you
327 are butchering. These people happen to be the ones who are
328 most likely to be knowledgeable enough to help you, but
329 they have no obligation to help you (i.e. you ask for help,
330 don't demand). "git log -p -- $area_you_are_modifying" would
331 help you find out who they are.
332
333 (2) You get comments and suggestions for improvements. You may
334 even get them in a "on top of your change" patch form.
335
336 (3) Polish, refine, and re-send to the list and the people who
337 spend their time to improve your patch. Go back to step (2).
338
339 (4) The list forms consensus that the last round of your patch is
faa8fac1 340 good. Send it to the maintainer and cc the list.
a941fb4a
JH
341
342 (5) A topic branch is created with the patch and is merged to 'next',
343 and cooked further and eventually graduates to 'master'.
344
345In any time between the (2)-(3) cycle, the maintainer may pick it up
346from the list and queue it to 'pu', in order to make it easier for
347people play with it without having to pick up and apply the patch to
348their trees themselves.
349
63cb8215
MM
350------------------------------------------------
351Know the status of your patch after submission
352
353* You can use Git itself to find out when your patch is merged in
354 master. 'git pull --rebase' will automatically skip already-applied
355 patches, and will let you know. This works only if you rebase on top
356 of the branch in which your patch has been merged (i.e. it will not
357 tell you if your patch is merged in pu if you rebase on top of
358 master).
359
2de9b711 360* Read the Git mailing list, the maintainer regularly posts messages
63cb8215
MM
361 entitled "What's cooking in git.git" and "What's in git.git" giving
362 the status of various proposed changes.
363
9740d289
JH
364------------------------------------------------
365MUA specific hints
366
367Some of patches I receive or pick up from the list share common
368patterns of breakage. Please make sure your MUA is set up
57756161
JN
369properly not to corrupt whitespaces.
370
371See the DISCUSSION section of git-format-patch(1) for hints on
372checking your patch by mailing it to yourself and applying with
373git-am(1).
374
375While you are at it, check the resulting commit log message from
376a trial run of applying the patch. If what is in the resulting
377commit is not exactly what you would want to see, it is very
378likely that your maintainer would end up hand editing the log
379message when he applies your patch. Things like "Hi, this is my
380first patch.\n", if you really want to put in the patch e-mail,
381should come after the three-dash line that signals the end of the
382commit message.
9847f7e0 383
9740d289
JH
384
385Pine
386----
387
388(Johannes Schindelin)
389
390I don't know how many people still use pine, but for those poor
391souls it may be good to mention that the quell-flowed-text is
392needed for recent versions.
393
394... the "no-strip-whitespace-before-send" option, too. AFAIK it
395was introduced in 4.60.
396
397(Linus Torvalds)
398
399And 4.58 needs at least this.
400
401---
402diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1)
403Author: Linus Torvalds <torvalds@g5.osdl.org>
404Date: Mon Aug 15 17:23:51 2005 -0700
405
406 Fix pine whitespace-corruption bug
407
408 There's no excuse for unconditionally removing whitespace from
409 the pico buffers on close.
410
411diff --git a/pico/pico.c b/pico/pico.c
412--- a/pico/pico.c
413+++ b/pico/pico.c
414@@ -219,7 +219,9 @@ PICO *pm;
a6080a0a
JH
415 switch(pico_all_done){ /* prepare for/handle final events */
416 case COMP_EXIT : /* already confirmed */
417 packheader();
9740d289 418+#if 0
a6080a0a 419 stripwhitespace();
9740d289 420+#endif
a6080a0a
JH
421 c |= COMP_EXIT;
422 break;
423
9740d289 424
1eb446fa
JH
425(Daniel Barkalow)
426
427> A patch to SubmittingPatches, MUA specific help section for
428> users of Pine 4.63 would be very much appreciated.
429
430Ah, it looks like a recent version changed the default behavior to do the
431right thing, and inverted the sense of the configuration option. (Either
432that or Gentoo did it.) So you need to set the
433"no-strip-whitespace-before-send" option, unless the option you have is
434"strip-whitespace-before-send", in which case you should avoid checking
435it.
436
9740d289 437
36c10e6d
JN
438Thunderbird, KMail, GMail
439-------------------------
9740d289 440
dc53151f 441See the MUA-SPECIFIC HINTS section of git-format-patch(1).
e30b217b 442
e30b217b
JH
443Gnus
444----
445
446'|' in the *Summary* buffer can be used to pipe the current
447message to an external program, and this is a handy way to drive
448"git am". However, if the message is MIME encoded, what is
449piped into the program is the representation you see in your
450*Article* buffer after unwrapping MIME. This is often not what
451you would want for two reasons. It tends to screw up non ASCII
452characters (most notably in people's names), and also
453whitespaces (fatal in patches). Running 'C-u g' to display the
454message in raw form before using '|' to run the pipe can work
455this problem around.