git
837 строк · 35.7 Кб
1Submitting Patches
2==================
3
4== Guidelines
5
6Here are some guidelines for contributing back to this
7project. There is also a link:MyFirstContribution.html[step-by-step tutorial]
8available which covers many of these same guidelines.
9
10[[patch-flow]]
11=== A typical life cycle of a patch series
12
13To help us understand the reason behind various guidelines given later
14in the document, first let's understand how the life cycle of a
15typical patch series for this project goes.
16
17. You come up with an itch. You code it up. You do not need any
18pre-authorization from the project to do so.
19+
20Your patches will be reviewed by other contributors on the mailing
21list, and the reviews will be done to assess the merit of various
22things, like the general idea behind your patch (including "is it
23solving a problem worth solving in the first place?"), the reason
24behind the design of the solution, and the actual implementation.
25The guidelines given here are there to help your patches by making
26them easier to understand by the reviewers.
27
28. You send the patches to the list and cc people who may need to know
29about the change. Your goal is *not* necessarily to convince others
30that what you are building is good. Your goal is to get help in
31coming up with a solution for the "itch" that is better than what
32you can build alone.
33+
34The people who may need to know are the ones who worked on the code
35you are touching. These people happen to be the ones who are
36most likely to be knowledgeable enough to help you, but
37they have no obligation to help you (i.e. you ask them for help,
38you don't demand). +git log -p {litdd} _$area_you_are_modifying_+ would
39help you find out who they are.
40
41. You get comments and suggestions for improvements. You may even get
42them in an "on top of your change" patch form. You are expected to
43respond to them with "Reply-All" on the mailing list, while taking
44them into account while preparing an updated set of patches.
45
46. Polish, refine, and re-send your patches to the list and to the people
47who spent their time to improve your patch. Go back to step (2).
48
49. While the above iterations improve your patches, the maintainer may
50pick the patches up from the list and queue them to the `seen`
51branch, in order to make it easier for people to play with it
52without having to pick up and apply the patches to their trees
53themselves. Being in `seen` has no other meaning. Specifically, it
54does not mean the patch was "accepted" in any way.
55
56. When the discussion reaches a consensus that the latest iteration of
57the patches are in good enough shape, the maintainer includes the
58topic in the "What's cooking" report that are sent out a few times a
59week to the mailing list, marked as "Will merge to 'next'." This
60decision is primarily made by the maintainer with help from those
61who participated in the review discussion.
62
63. After the patches are merged to the 'next' branch, the discussion
64can still continue to further improve them by adding more patches on
65top, but by the time a topic gets merged to 'next', it is expected
66that everybody agrees that the scope and the basic direction of the
67topic are appropriate, so such an incremental updates are limited to
68small corrections and polishing. After a topic cooks for some time
69(like 7 calendar days) in 'next' without needing further tweaks on
70top, it gets merged to the 'master' branch and wait to become part
71of the next major release.
72
73In the following sections, many techniques and conventions are listed
74to help your patches get reviewed effectively in such a life cycle.
75
76
77[[choose-starting-point]]
78=== Choose a starting point.
79
80As a preliminary step, you must first choose a starting point for your
81work. Typically this means choosing a branch, although technically
82speaking it is actually a particular commit (typically the HEAD, or tip,
83of the branch).
84
85There are several important branches to be aware of. Namely, there are
86four integration branches as discussed in linkgit:gitworkflows[7]:
87
88* maint
89* master
90* next
91* seen
92
93The branches lower on the list are typically descendants of the ones
94that come before it. For example, `maint` is an "older" branch than
95`master` because `master` usually has patches (commits) on top of
96`maint`.
97
98There are also "topic" branches, which contain work from other
99contributors. Topic branches are created by the Git maintainer (in
100their fork) to organize the current set of incoming contributions on
101the mailing list, and are itemized in the regular "What's cooking in
102git.git" announcements. To find the tip of a topic branch, run `git log
103--first-parent master..seen` and look for the merge commit. The second
104parent of this commit is the tip of the topic branch.
105
106There is one guiding principle for choosing the right starting point: in
107general, always base your work on the oldest integration branch that
108your change is relevant to (see "Merge upwards" in
109linkgit:gitworkflows[7]). What this principle means is that for the
110vast majority of cases, the starting point for new work should be the
111latest HEAD commit of `maint` or `master` based on the following cases:
112
113* If you are fixing bugs in the released version, use `maint` as the
114starting point (which may mean you have to fix things without using
115new API features on the cutting edge that recently appeared in
116`master` but were not available in the released version).
117
118* Otherwise (such as if you are adding new features) use `master`.
119
120
121NOTE: In exceptional cases, a bug that was introduced in an old
122version may have to be fixed for users of releases that are much older
123than the recent releases. `git describe --contains X` may describe
124`X` as `v2.30.0-rc2-gXXXXXX` for the commit `X` that introduced the
125bug, and the bug may be so high-impact that we may need to issue a new
126maintenance release for Git 2.30.x series, when "Git 2.41.0" is the
127current release. In such a case, you may want to use the tip of the
128maintenance branch for the 2.30.x series, which may be available in the
129`maint-2.30` branch in https://github.com/gitster/git[the maintainer's
130"broken out" repo].
131
132This also means that `next` or `seen` are inappropriate starting points
133for your work, if you want your work to have a realistic chance of
134graduating to `master`. They are simply not designed to be used as a
135base for new work; they are only there to make sure that topics in
136flight work well together. This is why both `next` and `seen` are
137frequently re-integrated with incoming patches on the mailing list and
138force-pushed to replace previous versions of themselves. A topic that is
139literally built on top of `next` cannot be merged to `master` without
140dragging in all the other topics in `next`, some of which may not be
141ready.
142
143For example, if you are making tree-wide changes, while somebody else is
144also making their own tree-wide changes, your work may have severe
145overlap with the other person's work. This situation may tempt you to
146use `next` as your starting point (because it would have the other
147person's work included in it), but doing so would mean you'll not only
148depend on the other person's work, but all the other random things from
149other contributors that are already integrated into `next`. And as soon
150as `next` is updated with a new version, all of your work will need to
151be rebased anyway in order for them to be cleanly applied by the
152maintainer.
153
154Under truly exceptional circumstances where you absolutely must depend
155on a select few topic branches that are already in `next` but not in
156`master`, you may want to create your own custom base-branch by forking
157`master` and merging the required topic branches into it. You could then
158work on top of this base-branch. But keep in mind that this base-branch
159would only be known privately to you. So when you are ready to send
160your patches to the list, be sure to communicate how you created it in
161your cover letter. This critical piece of information would allow
162others to recreate your base-branch on their end in order for them to
163try out your work.
164
165Finally, note that some parts of the system have dedicated maintainers
166with their own separate source code repositories (see the section
167"Subsystems" below).
168
169[[separate-commits]]
170=== Make separate commits for logically separate changes.
171
172Unless your patch is really trivial, you should not be sending
173out a patch that was generated between your working tree and
174your commit head. Instead, always make a commit with complete
175commit message and generate a series of patches from your
176repository. It is a good discipline.
177
178Give an explanation for the change(s) that is detailed enough so
179that people can judge if it is good thing to do, without reading
180the actual patch text to determine how well the code does what
181the explanation promises to do.
182
183If your description starts to get too long, that's a sign that you
184probably need to split up your commit to finer grained pieces.
185That being said, patches which plainly describe the things that
186help reviewers check the patch, and future maintainers understand
187the code, are the most beautiful patches. Descriptions that summarize
188the point in the subject well, and describe the motivation for the
189change, the approach taken by the change, and if relevant how this
190differs substantially from the prior version, are all good things
191to have.
192
193Make sure that you have tests for the bug you are fixing. See
194`t/README` for guidance.
195
196[[tests]]
197When adding a new feature, make sure that you have new tests to show
198the feature triggers the new behavior when it should, and to show the
199feature does not trigger when it shouldn't. After any code change,
200make sure that the entire test suite passes. When fixing a bug, make
201sure you have new tests that break if somebody else breaks what you
202fixed by accident to avoid regression. Also, try merging your work to
203'next' and 'seen' and make sure the tests still pass; topics by others
204that are still in flight may have unexpected interactions with what
205you are trying to do in your topic.
206
207Pushing to a fork of https://github.com/git/git will use their CI
208integration to test your changes on Linux, Mac and Windows. See the
209<<GHCI,GitHub CI>> section for details.
210
211Do not forget to update the documentation to describe the updated
212behavior and make sure that the resulting documentation set formats
213well (try the Documentation/doc-diff script).
214
215We currently have a liberal mixture of US and UK English norms for
216spelling and grammar, which is somewhat unfortunate. A huge patch that
217touches the files all over the place only to correct the inconsistency
218is not welcome, though. Potential clashes with other changes that can
219result from such a patch are not worth it. We prefer to gradually
220reconcile the inconsistencies in favor of US English, with small and
221easily digestible patches, as a side effect of doing some other real
222work in the vicinity (e.g. rewriting a paragraph for clarity, while
223turning en_UK spelling to en_US). Obvious typographical fixes are much
224more welcomed ("teh -> "the"), preferably submitted as independent
225patches separate from other documentation changes.
226
227[[whitespace-check]]
228Oh, another thing. We are picky about whitespaces. Make sure your
229changes do not trigger errors with the sample pre-commit hook shipped
230in `templates/hooks--pre-commit`. To help ensure this does not happen,
231run `git diff --check` on your changes before you commit.
232
233[[describe-changes]]
234=== Describe your changes well.
235
236The log message that explains your changes is just as important as the
237changes themselves. Your code may be clearly written with in-code
238comment to sufficiently explain how it works with the surrounding
239code, but those who need to fix or enhance your code in the future
240will need to know _why_ your code does what it does, for a few
241reasons:
242
243. Your code may be doing something differently from what you wanted it
244to do. Writing down what you actually wanted to achieve will help
245them fix your code and make it do what it should have been doing
246(also, you often discover your own bugs yourself, while writing the
247log message to summarize the thought behind it).
248
249. Your code may be doing things that were only necessary for your
250immediate needs (e.g. "do X to directories" without implementing or
251even designing what is to be done on files). Writing down why you
252excluded what the code does not do will help guide future developers.
253Writing down "we do X to directories, because directories have
254characteristic Y" would help them infer "oh, files also have the same
255characteristic Y, so perhaps doing X to them would also make sense?".
256Saying "we don't do the same X to files, because ..." will help them
257decide if the reasoning is sound (in which case they do not waste
258time extending your code to cover files), or reason differently (in
259which case, they can explain why they extend your code to cover
260files, too).
261
262The goal of your log message is to convey the _why_ behind your change
263to help future developers. The reviewers will also make sure that
264your proposed log message will serve this purpose well.
265
266The first line of the commit message should be a short description (50
267characters is the soft limit, see DISCUSSION in linkgit:git-commit[1]),
268and should skip the full stop. It is also conventional in most cases to
269prefix the first line with "area: " where the area is a filename or
270identifier for the general area of the code being modified, e.g.
271
272* doc: clarify distinction between sign-off and pgp-signing
273* githooks.txt: improve the intro section
274
275If in doubt which identifier to use, run `git log --no-merges` on the
276files you are modifying to see the current conventions.
277
278[[summary-section]]
279The title sentence after the "area:" prefix omits the full stop at the
280end, and its first word is not capitalized (the omission
281of capitalization applies only to the word after the "area:"
282prefix of the title) unless there is a reason to
283capitalize it other than because it is the first word in the sentence.
284E.g. "doc: clarify...", not "doc: Clarify...", or "githooks.txt:
285improve...", not "githooks.txt: Improve...". But "refs: HEAD is also
286treated as a ref" is correct, as we spell `HEAD` in all caps even when
287it appears in the middle of a sentence.
288
289[[meaningful-message]]
290The body should provide a meaningful commit message, which:
291
292. explains the problem the change tries to solve, i.e. what is wrong
293with the current code without the change.
294
295. justifies the way the change solves the problem, i.e. why the
296result with the change is better.
297
298. alternate solutions considered but discarded, if any.
299
300[[present-tense]]
301The problem statement that describes the status quo is written in the
302present tense. Write "The code does X when it is given input Y",
303instead of "The code used to do Y when given input X". You do not
304have to say "Currently"---the status quo in the problem statement is
305about the code _without_ your change, by project convention.
306
307[[imperative-mood]]
308Describe your changes in imperative mood, e.g. "make xyzzy do frotz"
309instead of "[This patch] makes xyzzy do frotz" or "[I] changed xyzzy
310to do frotz", as if you are giving orders to the codebase to change
311its behavior. Try to make sure your explanation can be understood
312without external resources. Instead of giving a URL to a mailing list
313archive, summarize the relevant points of the discussion.
314
315[[commit-reference]]
316
317There are a few reasons why you may want to refer to another commit in
318the "more stable" part of the history (i.e. on branches like `maint`,
319`master`, and `next`):
320
321. A commit that introduced the root cause of a bug you are fixing.
322
323. A commit that introduced a feature that you are enhancing.
324
325. A commit that conflicts with your work when you made a trial merge
326of your work into `next` and `seen` for testing.
327
328When you reference a commit on a more stable branch (like `master`,
329`maint` and `next`), use the format "abbreviated hash (subject,
330date)", like this:
331
332....
333Commit f86a374 (pack-bitmap.c: fix a memleak, 2015-03-30)
334noticed that ...
335....
336
337The "Copy commit reference" command of gitk can be used to obtain this
338format (with the subject enclosed in a pair of double-quotes), or this
339invocation of `git show`:
340
341....
342git show -s --pretty=reference <commit>
343....
344
345or, on an older version of Git without support for --pretty=reference:
346
347....
348git show -s --date=short --pretty='format:%h (%s, %ad)' <commit>
349....
350
351[[sign-off]]
352=== Certify your work by adding your `Signed-off-by` trailer
353
354To improve tracking of who did what, we ask you to certify that you
355wrote the patch or have the right to pass it on under the same license
356as ours, by "signing off" your patch. Without sign-off, we cannot
357accept your patches.
358
359If (and only if) you certify the below D-C-O:
360
361[[dco]]
362.Developer's Certificate of Origin 1.1
363____
364By making a contribution to this project, I certify that:
365
366a. The contribution was created in whole or in part by me and I
367have the right to submit it under the open source license
368indicated in the file; or
369
370b. The contribution is based upon previous work that, to the best
371of my knowledge, is covered under an appropriate open source
372license and I have the right under that license to submit that
373work with modifications, whether created in whole or in part
374by me, under the same open source license (unless I am
375permitted to submit under a different license), as indicated
376in the file; or
377
378c. The contribution was provided directly to me by some other
379person who certified (a), (b) or (c) and I have not modified
380it.
381
382d. I understand and agree that this project and the contribution
383are public and that a record of the contribution (including all
384personal information I submit with it, including my sign-off) is
385maintained indefinitely and may be redistributed consistent with
386this project or the open source license(s) involved.
387____
388
389you add a "Signed-off-by" trailer to your commit, that looks like
390this:
391
392....
393Signed-off-by: Random J Developer <random@developer.example.org>
394....
395
396This line can be added by Git if you run the git-commit command with
397the -s option.
398
399Notice that you can place your own `Signed-off-by` trailer when
400forwarding somebody else's patch with the above rules for
401D-C-O. Indeed you are encouraged to do so. Do not forget to
402place an in-body "From: " line at the beginning to properly attribute
403the change to its true author (see (2) above).
404
405This procedure originally came from the Linux kernel project, so our
406rule is quite similar to theirs, but what exactly it means to sign-off
407your patch differs from project to project, so it may be different
408from that of the project you are accustomed to.
409
410[[real-name]]
411Also notice that a real name is used in the `Signed-off-by` trailer. Please
412don't hide your real name.
413
414[[commit-trailers]]
415If you like, you can put extra tags at the end:
416
417. `Reported-by:` is used to credit someone who found the bug that
418the patch attempts to fix.
419. `Acked-by:` says that the person who is more familiar with the area
420the patch attempts to modify liked the patch.
421. `Reviewed-by:`, unlike the other tags, can only be offered by the
422reviewers themselves when they are completely satisfied with the
423patch after a detailed analysis.
424. `Tested-by:` is used to indicate that the person applied the patch
425and found it to have the desired effect.
426. `Co-authored-by:` is used to indicate that people exchanged drafts
427of a patch before submitting it.
428. `Helped-by:` is used to credit someone who suggested ideas for
429changes without providing the precise changes in patch form.
430. `Mentored-by:` is used to credit someone with helping develop a
431patch as part of a mentorship program (e.g., GSoC or Outreachy).
432. `Suggested-by:` is used to credit someone with suggesting the idea
433for a patch.
434
435While you can also create your own trailer if the situation warrants it, we
436encourage you to instead use one of the common trailers in this project
437highlighted above.
438
439Only capitalize the very first letter of tags, i.e. favor
440"Signed-off-by" over "Signed-Off-By" and "Acked-by:" over "Acked-By".
441
442[[git-tools]]
443=== Generate your patch using Git tools out of your commits.
444
445Git based diff tools generate unidiff which is the preferred format.
446
447You do not have to be afraid to use `-M` option to `git diff` or
448`git format-patch`, if your patch involves file renames. The
449receiving end can handle them just fine.
450
451[[review-patch]]
452Please make sure your patch does not add commented out debugging code,
453or include any extra files which do not relate to what your patch
454is trying to achieve. Make sure to review
455your patch after generating it, to ensure accuracy. Before
456sending out, please make sure it cleanly applies to the starting point you
457have chosen in the "Choose a starting point" section.
458
459NOTE: From the perspective of those reviewing your patch, the `master`
460branch is the default expected starting point. So if you have chosen a
461different starting point, please communicate this choice in your cover
462letter.
463
464
465[[send-patches]]
466=== Sending your patches.
467
468==== Choosing your reviewers
469
470:security-ml: footnoteref:[security-ml,The Git Security mailing list: git-security@googlegroups.com]
471
472NOTE: Patches that may be
473security relevant should be submitted privately to the Git Security
474mailing list{security-ml}, instead of the public mailing list.
475
476:contrib-scripts: footnoteref:[contrib-scripts,Scripts under `contrib/` are +
477not part of the core `git` binary and must be called directly. Clone the Git +
478codebase and run `perl contrib/contacts/git-contacts`.]
479
480Send your patch with "To:" set to the mailing list, with "cc:" listing
481people who are involved in the area you are touching (the `git-contacts`
482script in `contrib/contacts/`{contrib-scripts} can help to
483identify them), to solicit comments and reviews. Also, when you made
484trial merges of your topic to `next` and `seen`, you may have noticed
485work by others conflicting with your changes. There is a good possibility
486that these people may know the area you are touching well.
487
488If you are using `send-email`, you can feed it the output of `git-contacts` like
489this:
490
491....
492git send-email --cc-cmd='perl contrib/contacts/git-contacts' feature/*.patch
493....
494
495:current-maintainer: footnote:[The current maintainer: gitster@pobox.com]
496:git-ml: footnote:[The mailing list: git@vger.kernel.org]
497
498After the list reached a consensus that it is a good idea to apply the
499patch, re-send it with "To:" set to the maintainer{current-maintainer}
500and "cc:" the list{git-ml} for inclusion. This is especially relevant
501when the maintainer did not heavily participate in the discussion and
502instead left the review to trusted others.
503
504Do not forget to add trailers such as `Acked-by:`, `Reviewed-by:` and
505`Tested-by:` lines as necessary to credit people who helped your
506patch, and "cc:" them when sending such a final version for inclusion.
507
508==== `format-patch` and `send-email`
509
510Learn to use `format-patch` and `send-email` if possible. These commands
511are optimized for the workflow of sending patches, avoiding many ways
512your existing e-mail client (often optimized for "multipart/*" MIME
513type e-mails) might render your patches unusable.
514
515NOTE: Here we outline the procedure using `format-patch` and
516`send-email`, but you can instead use GitGitGadget to send in your
517patches (see link:MyFirstContribution.html[MyFirstContribution]).
518
519People on the Git mailing list need to be able to read and
520comment on the changes you are submitting. It is important for
521a developer to be able to "quote" your changes, using standard
522e-mail tools, so that they may comment on specific portions of
523your code. For this reason, each patch should be submitted
524"inline" in a separate message.
525
526All subsequent versions of a patch series and other related patches should be
527grouped into their own e-mail thread to help readers find all parts of the
528series. To that end, send them as replies to either an additional "cover
529letter" message (see below), the first patch, or the respective preceding patch.
530Here is a link:MyFirstContribution.html#v2-git-send-email[step-by-step guide] on
531how to submit updated versions of a patch series.
532
533If your log message (including your name on the
534`Signed-off-by` trailer) is not writable in ASCII, make sure that
535you send off a message in the correct encoding.
536
537WARNING: Be wary of your MUAs word-wrap
538corrupting your patch. Do not cut-n-paste your patch; you can
539lose tabs that way if you are not careful.
540
541It is a common convention to prefix your subject line with
542[PATCH]. This lets people easily distinguish patches from other
543e-mail discussions. Use of markers in addition to PATCH within
544the brackets to describe the nature of the patch is also
545encouraged. E.g. [RFC PATCH] (where RFC stands for "request for
546comments") is often used to indicate a patch needs further
547discussion before being accepted, [PATCH v2], [PATCH v3] etc.
548are often seen when you are sending an update to what you have
549previously sent.
550
551The `git format-patch` command follows the best current practice to
552format the body of an e-mail message. At the beginning of the
553patch should come your commit message, ending with the
554`Signed-off-by` trailers, and a line that consists of three dashes,
555followed by the diffstat information and the patch itself. If
556you are forwarding a patch from somebody else, optionally, at
557the beginning of the e-mail message just before the commit
558message starts, you can put a "From: " line to name that person.
559To change the default "[PATCH]" in the subject to "[<text>]", use
560`git format-patch --subject-prefix=<text>`. As a shortcut, you
561can use `--rfc` instead of `--subject-prefix="RFC PATCH"`, or
562`-v <n>` instead of `--subject-prefix="PATCH v<n>"`.
563
564You often want to add additional explanation about the patch,
565other than the commit message itself. Place such "cover letter"
566material between the three-dash line and the diffstat. For
567patches requiring multiple iterations of review and discussion,
568an explanation of changes between each iteration can be kept in
569Git-notes and inserted automatically following the three-dash
570line via `git format-patch --notes`.
571
572[[the-topic-summary]]
573*This is EXPERIMENTAL*.
574
575When sending a topic, you can propose a one-paragraph summary that
576should appear in the "What's cooking" report when it is picked up to
577explain the topic. If you choose to do so, please write a 2-5 line
578paragraph that will fit well in our release notes (see many bulleted
579entries in the Documentation/RelNotes/* files for examples), and make
580it the first paragraph of the cover letter. For a single-patch
581series, use the space between the three-dash line and the diffstat, as
582described earlier.
583
584[[attachment]]
585Do not attach the patch as a MIME attachment, compressed or not.
586Do not let your e-mail client send quoted-printable. Do not let
587your e-mail client send format=flowed which would destroy
588whitespaces in your patches. Many
589popular e-mail applications will not always transmit a MIME
590attachment as plain text, making it impossible to comment on
591your code. A MIME attachment also takes a bit more time to
592process. This does not decrease the likelihood of your
593MIME-attached change being accepted, but it makes it more likely
594that it will be postponed.
595
596Exception: If your mailer is mangling patches then someone may ask
597you to re-send them using MIME, that is OK.
598
599[[pgp-signature]]
600Do not PGP sign your patch. Most likely, your maintainer or other people on the
601list would not have your PGP key and would not bother obtaining it anyway.
602Your patch is not judged by who you are; a good patch from an unknown origin
603has a far better chance of being accepted than a patch from a known, respected
604origin that is done poorly or does incorrect things.
605
606If you really really really really want to do a PGP signed
607patch, format it as "multipart/signed", not a text/plain message
608that starts with `-----BEGIN PGP SIGNED MESSAGE-----`. That is
609not a text/plain, it's something else.
610
611=== Handling Conflicts and Iterating Patches
612
613When revising changes made to your patches, it's important to
614acknowledge the possibility of conflicts with other ongoing topics. To
615navigate these potential conflicts effectively, follow the recommended
616steps outlined below:
617
618. Build on a suitable base branch, see the <<choose-starting-point, section above>>,
619and format-patch the series. If you are doing "rebase -i" in-place to
620update from the previous round, this will reuse the previous base so
621(2) and (3) may become trivial.
622
623. Find the base of where the last round was queued
624+
625$ mine='kn/ref-transaction-symref'
626$ git checkout "origin/seen^{/^Merge branch '$mine'}...master"
627
628. Apply your format-patch result. There are two cases
629.. Things apply cleanly and tests fine. Go to (4).
630.. Things apply cleanly but does not build or test fails, or things do
631not apply cleanly.
632+
633In the latter case, you have textual or semantic conflicts coming from
634the difference between the old base and the base you used to build in
635(1). Identify what caused the breakages (e.g., a topic or two may have
636merged since the base used by (2) until the base used by (1)).
637+
638Check out the latest 'origin/master' (which may be newer than the base
639used by (2)), "merge --no-ff" the topics you newly depend on in there,
640and use the result of the merge(s) as the base, rebuild the series and
641test again. Run format-patch from the last such merges to the tip of
642your topic. If you did
643+
644$ git checkout origin/master
645$ git merge --no-ff --into-name kn/ref-transaction-symref fo/obar
646$ git merge --no-ff --into-name kn/ref-transaction-symref ba/zqux
647... rebuild the topic ...
648+
649Then you'd just format your topic above these "preparing the ground"
650merges, e.g.
651+
652$ git format-patch "HEAD^{/^Merge branch 'ba/zqux'}"..HEAD
653+
654Do not forget to write in the cover letter you did this, including the
655topics you have in your base on top of 'master'. Then go to (4).
656
657. Make a trial merge of your topic into 'next' and 'seen', e.g.
658+
659$ git checkout --detach 'origin/seen'
660$ git revert -m 1 <the merge of the previous iteration into seen>
661$ git merge kn/ref-transaction-symref
662+
663The "revert" is needed if the previous iteration of your topic is
664already in 'seen' (like in this case). You could choose to rebuild
665master..origin/seen from scratch while excluding your previous
666iteration, which may emulate what happens on the maintainers end more
667closely.
668+
669This trial merge may conflict. It is primarily to see what conflicts
670_other_ topics may have with your topic. In other words, you do not
671have to depend on it to make your topic work on 'master'. It may
672become the job of the other topic owners to resolve conflicts if your
673topic goes to 'next' before theirs.
674+
675Make a note on what conflict you saw in the cover letter. You do not
676necessarily have to resolve them, but it would be a good opportunity to
677learn what others are doing in related areas.
678+
679$ git checkout --detach 'origin/next'
680$ git merge kn/ref-transaction-symref
681+
682This is to see what conflicts your topic has with other topics that are
683already cooking. This should not conflict if (3)-2 prepared a base on
684top of updated master plus dependent topics taken from 'next'. Unless
685the context is severe (one way to tell is try the same trial merge with
686your old iteration, which may conflict in a similar way), expect that it
687will be handled on maintainers end (if it gets unmanageable, I'll ask to
688rebase when I receive your patches).
689
690== Subsystems with dedicated maintainers
691
692Some parts of the system have dedicated maintainers with their own
693repositories.
694
695- `git-gui/` comes from git-gui project, maintained by Johannes Sixt:
696
697https://github.com/j6t/git-gui
698
699- `gitk-git/` comes from Paul Mackerras's gitk project:
700
701git://git.ozlabs.org/~paulus/gitk
702
703Those who are interested in improving gitk can volunteer to help Paul
704maintain it, cf. <YntxL/fTplFm8lr6@cleo>.
705
706- `po/` comes from the localization coordinator, Jiang Xin:
707
708https://github.com/git-l10n/git-po/
709
710Patches to these parts should be based on their trees.
711
712- The "Git documentation translations" project, led by Jean-Noël
713Avila, translates our documentation pages. Their work products are
714maintained separately from this project, not as part of our tree:
715
716https://github.com/jnavila/git-manpages-l10n/
717
718
719== GitHub CI[[GHCI]]
720
721With an account at GitHub, you can use GitHub CI to test your changes
722on Linux, Mac and Windows. See
723https://github.com/git/git/actions/workflows/main.yml for examples of
724recent CI runs.
725
726Follow these steps for the initial setup:
727
728. Fork https://github.com/git/git to your GitHub account.
729You can find detailed instructions how to fork here:
730https://help.github.com/articles/fork-a-repo/
731
732After the initial setup, CI will run whenever you push new changes
733to your fork of Git on GitHub. You can monitor the test state of all your
734branches here: `https://github.com/<Your GitHub handle>/git/actions/workflows/main.yml`
735
736If a branch does not pass all test cases then it will be marked with a
737red +x+, instead of a green check. In that case, you can click on the
738failing job and navigate to "ci/run-build-and-tests.sh" and/or
739"ci/print-test-failures.sh". You can also download "Artifacts" which
740are zip archives containing tarred (or zipped) archives with test data
741relevant for debugging.
742
743Then fix the problem and push your fix to your GitHub fork. This will
744trigger a new CI build to ensure all tests pass.
745
746[[mua]]
747== MUA specific hints
748
749Some of the patches I receive or pick up from the list share common
750patterns of breakage. Please make sure your MUA is set up
751properly not to corrupt whitespaces.
752
753See the DISCUSSION section of linkgit:git-format-patch[1] for hints on
754checking your patch by mailing it to yourself and applying with
755linkgit:git-am[1].
756
757While you are at it, check the resulting commit log message from
758a trial run of applying the patch. If what is in the resulting
759commit is not exactly what you would want to see, it is very
760likely that your maintainer would end up hand editing the log
761message when he applies your patch. Things like "Hi, this is my
762first patch.\n", if you really want to put in the patch e-mail,
763should come after the three-dash line that signals the end of the
764commit message.
765
766
767=== Pine
768
769(Johannes Schindelin)
770
771....
772I don't know how many people still use pine, but for those poor
773souls it may be good to mention that the quell-flowed-text is
774needed for recent versions.
775
776... the "no-strip-whitespace-before-send" option, too. AFAIK it
777was introduced in 4.60.
778....
779
780(Linus Torvalds)
781
782....
783And 4.58 needs at least this.
784
785diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1)
786Author: Linus Torvalds <torvalds@g5.osdl.org>
787Date: Mon Aug 15 17:23:51 2005 -0700
788
789Fix pine whitespace-corruption bug
790
791There's no excuse for unconditionally removing whitespace from
792the pico buffers on close.
793
794diff --git a/pico/pico.c b/pico/pico.c
795--- a/pico/pico.c
796+++ b/pico/pico.c
797@@ -219,7 +219,9 @@ PICO *pm;
798switch(pico_all_done){ /* prepare for/handle final events */
799case COMP_EXIT : /* already confirmed */
800packheader();
801+#if 0
802stripwhitespace();
803+#endif
804c |= COMP_EXIT;
805break;
806....
807
808(Daniel Barkalow)
809
810....
811> A patch to SubmittingPatches, MUA specific help section for
812> users of Pine 4.63 would be very much appreciated.
813
814Ah, it looks like a recent version changed the default behavior to do the
815right thing, and inverted the sense of the configuration option. (Either
816that or Gentoo did it.) So you need to set the
817"no-strip-whitespace-before-send" option, unless the option you have is
818"strip-whitespace-before-send", in which case you should avoid checking
819it.
820....
821
822=== Thunderbird, KMail, GMail
823
824See the MUA-SPECIFIC HINTS section of linkgit:git-format-patch[1].
825
826=== Gnus
827
828"|" in the `*Summary*` buffer can be used to pipe the current
829message to an external program, and this is a handy way to drive
830`git am`. However, if the message is MIME encoded, what is
831piped into the program is the representation you see in your
832`*Article*` buffer after unwrapping MIME. This is often not what
833you would want for two reasons. It tends to screw up non-ASCII
834characters (most notably in people's names), and also
835whitespaces (fatal in patches). Running "C-u g" to display the
836message in raw form before using "|" to run the pipe can work
837this problem around.
838