1GIT-FORMAT-PATCH(1) Git Manual GIT-FORMAT-PATCH(1)
2
6 git-format-patch - Prepare patches for e-mail submission
7
9 git format-patch [-k] [(-o|--output-directory) <dir> | --stdout]
10 [--no-thread | --thread[=<style>]]
11 [(--attach|--inline)[=<boundary>] | --no-attach]
12 [-s | --signoff]
13 [--signature=<signature> | --no-signature]
14 [--signature-file=<file>]
15 [-n | --numbered | -N | --no-numbered]
16 [--start-number <n>] [--numbered-files]
17 [--in-reply-to=<message id>] [--suffix=.<sfx>]
18 [--ignore-if-in-upstream]
19 [--cover-from-description=<mode>]
20 [--rfc] [--subject-prefix=<subject prefix>]
21 [(--reroll-count|-v) <n>]
22 [--to=<email>] [--cc=<email>]
23 [--[no-]cover-letter] [--quiet]
24 [--[no-]encode-email-headers]
25 [--no-notes | --notes[=<ref>]]
26 [--interdiff=<previous>]
27 [--range-diff=<previous> [--creation-factor=<percent>]]
28 [--filename-max-length=<n>]
29 [--progress]
30 [<common diff options>]
31 [ <since> | <revision range> ]
32
34 Prepare each commit with its patch in one file per commit, formatted to
35 resemble UNIX mailbox format. The output of this command is convenient
36 for e-mail submission or for use with git am.
37 There are two ways to specify which commits to operate on.
39 1. A single commit, <since>, specifies that the commits leading to the
41 tip of the current branch that are not in the history that leads to
42 the <since> to be output.
43 2. Generic <revision range> expression (see "SPECIFYING REVISIONS"
45 section in gitrevisions(7)) means the commits in the specified
46 range.
47 The first rule takes precedence in the case of a single <commit>. To
49 apply the second rule, i.e., format everything since the beginning of
50 history up until <commit>, use the --root option: git format-patch
51 --root <commit>. If you want to format only <commit> itself, you can do
52 this with git format-patch -1 <commit>.
53 By default, each output file is numbered sequentially from 1, and uses
55 the first line of the commit message (massaged for pathname safety) as
56 the filename. With the --numbered-files option, the output file names
57 will only be numbers, without the first line of the commit appended.
58 The names of the output files are printed to standard output, unless
59 the --stdout option is specified.
60 If -o is specified, output files are created in <dir>. Otherwise they
62 are created in the current working directory. The default path can be
63 set with the format.outputDirectory configuration option. The -o option
64 takes precedence over format.outputDirectory. To store patches in the
65 current working directory even when format.outputDirectory points
66 elsewhere, use -o .. All directory components will be created.
67 By default, the subject of a single patch is "[PATCH] " followed by the
69 concatenation of lines from the commit message up to the first blank
70 line (see the DISCUSSION section of git-commit(1)).
71 When multiple patches are output, the subject prefix will instead be
73 "[PATCH n/m] ". To force 1/1 to be added for a single patch, use -n. To
74 omit patch numbers from the subject, use -N.
75 If given --thread, git-format-patch will generate In-Reply-To and
77 References headers to make the second and subsequent patch mails appear
78 as replies to the first mail; this also generates a Message-Id header
79 to reference.
80
82 -p, --no-stat
83 Generate plain patches without any diffstats.
84 -U<n>, --unified=<n>
86 Generate diffs with <n> lines of context instead of the usual
87 three.
88 --output=<file>
90 Output to a specific file instead of stdout.
91 --output-indicator-new=<char>, --output-indicator-old=<char>,
93 --output-indicator-context=<char>
94 Specify the character used to indicate new, old or context lines in
95 the generated patch. Normally they are +, - and ' ' respectively.
96 --indent-heuristic
98 Enable the heuristic that shifts diff hunk boundaries to make
99 patches easier to read. This is the default.
100 --no-indent-heuristic
102 Disable the indent heuristic.
103 --minimal
105 Spend extra time to make sure the smallest possible diff is
106 produced.
107 --patience
109 Generate a diff using the "patience diff" algorithm.
110 --histogram
112 Generate a diff using the "histogram diff" algorithm.
113 --anchored=<text>
115 Generate a diff using the "anchored diff" algorithm.
116 This option may be specified more than once.
118 If a line exists in both the source and destination, exists only
120 once, and starts with this text, this algorithm attempts to prevent
121 it from appearing as a deletion or addition in the output. It uses
122 the "patience diff" algorithm internally.
123 --diff-algorithm={patience|minimal|histogram|myers}
125 Choose a diff algorithm. The variants are as follows:
126 default, myers
128 The basic greedy diff algorithm. Currently, this is the
129 default.
130 minimal
132 Spend extra time to make sure the smallest possible diff is
133 produced.
134 patience
136 Use "patience diff" algorithm when generating patches.
137 histogram
139 This algorithm extends the patience algorithm to "support
140 low-occurrence common elements".
141 For instance, if you configured the diff.algorithm variable to a
143 non-default value and want to use the default one, then you have to
144 use --diff-algorithm=default option.
145 --stat[=<width>[,<name-width>[,<count>]]]
147 Generate a diffstat. By default, as much space as necessary will be
148 used for the filename part, and the rest for the graph part.
149 Maximum width defaults to terminal width, or 80 columns if not
150 connected to a terminal, and can be overridden by <width>. The
151 width of the filename part can be limited by giving another width
152 <name-width> after a comma. The width of the graph part can be
153 limited by using --stat-graph-width=<width> (affects all commands
154 generating a stat graph) or by setting diff.statGraphWidth=<width>
155 (does not affect git format-patch). By giving a third parameter
156 <count>, you can limit the output to the first <count> lines,
157 followed by ... if there are more.
158 These parameters can also be set individually with
160 --stat-width=<width>, --stat-name-width=<name-width> and
161 --stat-count=<count>.
162 --compact-summary
164 Output a condensed summary of extended header information such as
165 file creations or deletions ("new" or "gone", optionally "+l" if
166 it’s a symlink) and mode changes ("+x" or "-x" for adding or
167 removing executable bit respectively) in diffstat. The information
168 is put between the filename part and the graph part. Implies
169 --stat.
170 --numstat
172 Similar to --stat, but shows number of added and deleted lines in
173 decimal notation and pathname without abbreviation, to make it more
174 machine friendly. For binary files, outputs two - instead of saying
175 0 0.
176 --shortstat
178 Output only the last line of the --stat format containing total
179 number of modified files, as well as number of added and deleted
180 lines.
181 -X[<param1,param2,...>], --dirstat[=<param1,param2,...>]
183 Output the distribution of relative amount of changes for each
184 sub-directory. The behavior of --dirstat can be customized by
185 passing it a comma separated list of parameters. The defaults are
186 controlled by the diff.dirstat configuration variable (see git-
187 config(1)). The following parameters are available:
188 changes
190 Compute the dirstat numbers by counting the lines that have
191 been removed from the source, or added to the destination. This
192 ignores the amount of pure code movements within a file. In
193 other words, rearranging lines in a file is not counted as much
194 as other changes. This is the default behavior when no
195 parameter is given.
196 lines
198 Compute the dirstat numbers by doing the regular line-based
199 diff analysis, and summing the removed/added line counts. (For
200 binary files, count 64-byte chunks instead, since binary files
201 have no natural concept of lines). This is a more expensive
202 --dirstat behavior than the changes behavior, but it does count
203 rearranged lines within a file as much as other changes. The
204 resulting output is consistent with what you get from the other
205 --*stat options.
206 files
208 Compute the dirstat numbers by counting the number of files
209 changed. Each changed file counts equally in the dirstat
210 analysis. This is the computationally cheapest --dirstat
211 behavior, since it does not have to look at the file contents
212 at all.
213 cumulative
215 Count changes in a child directory for the parent directory as
216 well. Note that when using cumulative, the sum of the
217 percentages reported may exceed 100%. The default
218 (non-cumulative) behavior can be specified with the
219 noncumulative parameter.
220 <limit>
222 An integer parameter specifies a cut-off percent (3% by
223 default). Directories contributing less than this percentage of
224 the changes are not shown in the output.
225 Example: The following will count changed files, while ignoring
227 directories with less than 10% of the total amount of changed
228 files, and accumulating child directory counts in the parent
229 directories: --dirstat=files,10,cumulative.
230 --cumulative
232 Synonym for --dirstat=cumulative
233 --dirstat-by-file[=<param1,param2>...]
235 Synonym for --dirstat=files,param1,param2...
236 --summary
238 Output a condensed summary of extended header information such as
239 creations, renames and mode changes.
240 --no-renames
242 Turn off rename detection, even when the configuration file gives
243 the default to do so.
244 --[no-]rename-empty
246 Whether to use empty blobs as rename source.
247 --full-index
249 Instead of the first handful of characters, show the full pre- and
250 post-image blob object names on the "index" line when generating
251 patch format output.
252 --binary
254 In addition to --full-index, output a binary diff that can be
255 applied with git-apply.
256 --abbrev[=<n>]
258 Instead of showing the full 40-byte hexadecimal object name in
259 diff-raw format output and diff-tree header lines, show the
260 shortest prefix that is at least <n> hexdigits long that uniquely
261 refers the object. In diff-patch output format, --full-index takes
262 higher precedence, i.e. if --full-index is specified, full blob
263 names will be shown regardless of --abbrev. Non default number of
264 digits can be specified with --abbrev=<n>.
265 -B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]]
267 Break complete rewrite changes into pairs of delete and create.
268 This serves two purposes:
269 It affects the way a change that amounts to a total rewrite of a
271 file not as a series of deletion and insertion mixed together with
272 a very few lines that happen to match textually as the context, but
273 as a single deletion of everything old followed by a single
274 insertion of everything new, and the number m controls this aspect
275 of the -B option (defaults to 60%). -B/70% specifies that less
276 than 30% of the original should remain in the result for Git to
277 consider it a total rewrite (i.e. otherwise the resulting patch
278 will be a series of deletion and insertion mixed together with
279 context lines).
280 When used with -M, a totally-rewritten file is also considered as
282 the source of a rename (usually -M only considers a file that
283 disappeared as the source of a rename), and the number n controls
284 this aspect of the -B option (defaults to 50%). -B20% specifies
285 that a change with addition and deletion compared to 20% or more of
286 the file’s size are eligible for being picked up as a possible
287 source of a rename to another file.
288 -M[<n>], --find-renames[=<n>]
290 Detect renames. If n is specified, it is a threshold on the
291 similarity index (i.e. amount of addition/deletions compared to the
292 file’s size). For example, -M90% means Git should consider a
293 delete/add pair to be a rename if more than 90% of the file hasn’t
294 changed. Without a % sign, the number is to be read as a fraction,
295 with a decimal point before it. I.e., -M5 becomes 0.5, and is thus
296 the same as -M50%. Similarly, -M05 is the same as -M5%. To limit
297 detection to exact renames, use -M100%. The default similarity
298 index is 50%.
299 -C[<n>], --find-copies[=<n>]
301 Detect copies as well as renames. See also --find-copies-harder. If
302 n is specified, it has the same meaning as for -M<n>.
303 --find-copies-harder
305 For performance reasons, by default, -C option finds copies only if
306 the original file of the copy was modified in the same changeset.
307 This flag makes the command inspect unmodified files as candidates
308 for the source of copy. This is a very expensive operation for
309 large projects, so use it with caution. Giving more than one -C
310 option has the same effect.
311 -D, --irreversible-delete
313 Omit the preimage for deletes, i.e. print only the header but not
314 the diff between the preimage and /dev/null. The resulting patch is
315 not meant to be applied with patch or git apply; this is solely for
316 people who want to just concentrate on reviewing the text after the
317 change. In addition, the output obviously lacks enough information
318 to apply such a patch in reverse, even manually, hence the name of
319 the option.
320 When used together with -B, omit also the preimage in the deletion
322 part of a delete/create pair.
323 -l<num>
325 The -M and -C options require O(n^2) processing time where n is the
326 number of potential rename/copy targets. This option prevents
327 rename/copy detection from running if the number of rename/copy
328 targets exceeds the specified number.
329 -O<orderfile>
331 Control the order in which files appear in the output. This
332 overrides the diff.orderFile configuration variable (see git-
333 config(1)). To cancel diff.orderFile, use -O/dev/null.
334 The output order is determined by the order of glob patterns in
336 <orderfile>. All files with pathnames that match the first pattern
337 are output first, all files with pathnames that match the second
338 pattern (but not the first) are output next, and so on. All files
339 with pathnames that do not match any pattern are output last, as if
340 there was an implicit match-all pattern at the end of the file. If
341 multiple pathnames have the same rank (they match the same pattern
342 but no earlier patterns), their output order relative to each other
343 is the normal order.
344 <orderfile> is parsed as follows:
346 • Blank lines are ignored, so they can be used as separators for
348 readability.
349 • Lines starting with a hash ("#") are ignored, so they can be
351 used for comments. Add a backslash ("\") to the beginning of
352 the pattern if it starts with a hash.
353 • Each other line contains a single pattern.
355 Patterns have the same syntax and semantics as patterns used for
357 fnmatch(3) without the FNM_PATHNAME flag, except a pathname also
358 matches a pattern if removing any number of the final pathname
359 components matches the pattern. For example, the pattern "foo*bar"
360 matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx".
361 --skip-to=<file>, --rotate-to=<file>
363 Discard the files before the named <file> from the output (i.e.
364 skip to), or move them to the end of the output (i.e. rotate to).
365 These were invented primarily for use of the git difftool command,
366 and may not be very useful otherwise.
367 --relative[=<path>], --no-relative
369 When run from a subdirectory of the project, it can be told to
370 exclude changes outside the directory and show pathnames relative
371 to it with this option. When you are not in a subdirectory (e.g. in
372 a bare repository), you can name which subdirectory to make the
373 output relative to by giving a <path> as an argument.
374 --no-relative can be used to countermand both diff.relative config
375 option and previous --relative.
376 -a, --text
378 Treat all files as text.
379 --ignore-cr-at-eol
381 Ignore carriage-return at the end of line when doing a comparison.
382 --ignore-space-at-eol
384 Ignore changes in whitespace at EOL.
385 -b, --ignore-space-change
387 Ignore changes in amount of whitespace. This ignores whitespace at
388 line end, and considers all other sequences of one or more
389 whitespace characters to be equivalent.
390 -w, --ignore-all-space
392 Ignore whitespace when comparing lines. This ignores differences
393 even if one line has whitespace where the other line has none.
394 --ignore-blank-lines
396 Ignore changes whose lines are all blank.
397 -I<regex>, --ignore-matching-lines=<regex>
399 Ignore changes whose all lines match <regex>. This option may be
400 specified more than once.
401 --inter-hunk-context=<lines>
403 Show the context between diff hunks, up to the specified number of
404 lines, thereby fusing hunks that are close to each other. Defaults
405 to diff.interHunkContext or 0 if the config option is unset.
406 -W, --function-context
408 Show whole function as context lines for each change. The function
409 names are determined in the same way as git diff works out patch
410 hunk headers (see Defining a custom hunk-header in
411 gitattributes(5)).
412 --ext-diff
414 Allow an external diff helper to be executed. If you set an
415 external diff driver with gitattributes(5), you need to use this
416 option with git-log(1) and friends.
417 --no-ext-diff
419 Disallow external diff drivers.
420 --textconv, --no-textconv
422 Allow (or disallow) external text conversion filters to be run when
423 comparing binary files. See gitattributes(5) for details. Because
424 textconv filters are typically a one-way conversion, the resulting
425 diff is suitable for human consumption, but cannot be applied. For
426 this reason, textconv filters are enabled by default only for git-
427 diff(1) and git-log(1), but not for git-format-patch(1) or diff
428 plumbing commands.
429 --ignore-submodules[=<when>]
431 Ignore changes to submodules in the diff generation. <when> can be
432 either "none", "untracked", "dirty" or "all", which is the default.
433 Using "none" will consider the submodule modified when it either
434 contains untracked or modified files or its HEAD differs from the
435 commit recorded in the superproject and can be used to override any
436 settings of the ignore option in git-config(1) or gitmodules(5).
437 When "untracked" is used submodules are not considered dirty when
438 they only contain untracked content (but they are still scanned for
439 modified content). Using "dirty" ignores all changes to the work
440 tree of submodules, only changes to the commits stored in the
441 superproject are shown (this was the behavior until 1.7.0). Using
442 "all" hides all changes to submodules.
443 --src-prefix=<prefix>
445 Show the given source prefix instead of "a/".
446 --dst-prefix=<prefix>
448 Show the given destination prefix instead of "b/".
449 --no-prefix
451 Do not show any source or destination prefix.
452 --line-prefix=<prefix>
454 Prepend an additional prefix to every line of output.
455 --ita-invisible-in-index
457 By default entries added by "git add -N" appear as an existing
458 empty file in "git diff" and a new file in "git diff --cached".
459 This option makes the entry appear as a new file in "git diff" and
460 non-existent in "git diff --cached". This option could be reverted
461 with --ita-visible-in-index. Both options are experimental and
462 could be removed in future.
463 For more detailed explanation on these common options, see also
465 gitdiffcore(7).
466 -<n>
468 Prepare patches from the topmost <n> commits.
469 -o <dir>, --output-directory <dir>
471 Use <dir> to store the resulting files, instead of the current
472 working directory.
473 -n, --numbered
475 Name output in [PATCH n/m] format, even with a single patch.
476 -N, --no-numbered
478 Name output in [PATCH] format.
479 --start-number <n>
481 Start numbering the patches at <n> instead of 1.
482 --numbered-files
484 Output file names will be a simple number sequence without the
485 default first line of the commit appended.
486 -k, --keep-subject
488 Do not strip/add [PATCH] from the first line of the commit log
489 message.
490 -s, --signoff
492 Add a Signed-off-by trailer to the commit message, using the
493 committer identity of yourself. See the signoff option in git-
494 commit(1) for more information.
495 --stdout
497 Print all commits to the standard output in mbox format, instead of
498 creating a file for each one.
499 --attach[=<boundary>]
501 Create multipart/mixed attachment, the first part of which is the
502 commit message and the patch itself in the second part, with
503 Content-Disposition: attachment.
504 --no-attach
506 Disable the creation of an attachment, overriding the configuration
507 setting.
508 --inline[=<boundary>]
510 Create multipart/mixed attachment, the first part of which is the
511 commit message and the patch itself in the second part, with
512 Content-Disposition: inline.
513 --thread[=<style>], --no-thread
515 Controls addition of In-Reply-To and References headers to make the
516 second and subsequent mails appear as replies to the first. Also
517 controls generation of the Message-Id header to reference.
518 The optional <style> argument can be either shallow or deep.
520 shallow threading makes every mail a reply to the head of the
521 series, where the head is chosen from the cover letter, the
522 --in-reply-to, and the first patch mail, in this order. deep
523 threading makes every mail a reply to the previous one.
524 The default is --no-thread, unless the format.thread configuration
526 is set. If --thread is specified without a style, it defaults to
527 the style specified by format.thread if any, or else shallow.
528 Beware that the default for git send-email is to thread emails
530 itself. If you want git format-patch to take care of threading, you
531 will want to ensure that threading is disabled for git send-email.
532 --in-reply-to=<message id>
534 Make the first mail (or all the mails with --no-thread) appear as a
535 reply to the given <message id>, which avoids breaking threads to
536 provide a new patch series.
537 --ignore-if-in-upstream
539 Do not include a patch that matches a commit in <until>..<since>.
540 This will examine all patches reachable from <since> but not from
541 <until> and compare them with the patches being generated, and any
542 patch that matches is ignored.
543 --cover-from-description=<mode>
545 Controls which parts of the cover letter will be automatically
546 populated using the branch’s description.
547 If <mode> is message or default, the cover letter subject will be
549 populated with placeholder text. The body of the cover letter will
550 be populated with the branch’s description. This is the default
551 mode when no configuration nor command line option is specified.
552 If <mode> is subject, the first paragraph of the branch description
554 will populate the cover letter subject. The remainder of the
555 description will populate the body of the cover letter.
556 If <mode> is auto, if the first paragraph of the branch description
558 is greater than 100 bytes, then the mode will be message, otherwise
559 subject will be used.
560 If <mode> is none, both the cover letter subject and body will be
562 populated with placeholder text.
563 --subject-prefix=<subject prefix>
565 Instead of the standard [PATCH] prefix in the subject line, instead
566 use [<subject prefix>]. This allows for useful naming of a patch
567 series, and can be combined with the --numbered option.
568 --filename-max-length=<n>
570 Instead of the standard 64 bytes, chomp the generated output
571 filenames at around <n> bytes (too short a value will be silently
572 raised to a reasonable length). Defaults to the value of the
573 format.filenameMaxLength configuration variable, or 64 if
574 unconfigured.
575 --rfc
577 Alias for --subject-prefix="RFC PATCH". RFC means "Request For
578 Comments"; use this when sending an experimental patch for
579 discussion rather than application.
580 -v <n>, --reroll-count=<n>
582 Mark the series as the <n>-th iteration of the topic. The output
583 filenames have v<n> prepended to them, and the subject prefix
584 ("PATCH" by default, but configurable via the --subject-prefix
585 option) has ` v<n>` appended to it. E.g. --reroll-count=4 may
586 produce v4-0001-add-makefile.patch file that has "Subject: [PATCH
587 v4 1/20] Add makefile" in it.
588 --to=<email>
590 Add a To: header to the email headers. This is in addition to any
591 configured headers, and may be used multiple times. The negated
592 form --no-to discards all To: headers added so far (from config or
593 command line).
594 --cc=<email>
596 Add a Cc: header to the email headers. This is in addition to any
597 configured headers, and may be used multiple times. The negated
598 form --no-cc discards all Cc: headers added so far (from config or
599 command line).
600 --from, --from=<ident>
602 Use ident in the From: header of each commit email. If the author
603 ident of the commit is not textually identical to the provided
604 ident, place a From: header in the body of the message with the
605 original author. If no ident is given, use the committer ident.
606 Note that this option is only useful if you are actually sending
608 the emails and want to identify yourself as the sender, but retain
609 the original author (and git am will correctly pick up the in-body
610 header). Note also that git send-email already handles this
611 transformation for you, and this option should not be used if you
612 are feeding the result to git send-email.
613 --add-header=<header>
615 Add an arbitrary header to the email headers. This is in addition
616 to any configured headers, and may be used multiple times. For
617 example, --add-header="Organization: git-foo". The negated form
618 --no-add-header discards all (To:, Cc:, and custom) headers added
619 so far from config or command line.
620 --[no-]cover-letter
622 In addition to the patches, generate a cover letter file containing
623 the branch description, shortlog and the overall diffstat. You can
624 fill in a description in the file before sending it out.
625 --encode-email-headers, --no-encode-email-headers
627 Encode email headers that have non-ASCII characters with
628 "Q-encoding" (described in RFC 2047), instead of outputting the
629 headers verbatim. Defaults to the value of the
630 format.encodeEmailHeaders configuration variable.
631 --interdiff=<previous>
633 As a reviewer aid, insert an interdiff into the cover letter, or as
634 commentary of the lone patch of a 1-patch series, showing the
635 differences between the previous version of the patch series and
636 the series currently being formatted. previous is a single
637 revision naming the tip of the previous series which shares a
638 common base with the series being formatted (for example git
639 format-patch --cover-letter --interdiff=feature/v1 -3 feature/v2).
640 --range-diff=<previous>
642 As a reviewer aid, insert a range-diff (see git-range-diff(1)) into
643 the cover letter, or as commentary of the lone patch of a 1-patch
644 series, showing the differences between the previous version of the
645 patch series and the series currently being formatted. previous
646 can be a single revision naming the tip of the previous series if
647 it shares a common base with the series being formatted (for
648 example git format-patch --cover-letter --range-diff=feature/v1 -3
649 feature/v2), or a revision range if the two versions of the series
650 are disjoint (for example git format-patch --cover-letter
651 --range-diff=feature/v1~3..feature/v1 -3 feature/v2).
652 Note that diff options passed to the command affect how the primary
654 product of format-patch is generated, and they are not passed to
655 the underlying range-diff machinery used to generate the
656 cover-letter material (this may change in the future).
657 --creation-factor=<percent>
659 Used with --range-diff, tweak the heuristic which matches up
660 commits between the previous and current series of patches by
661 adjusting the creation/deletion cost fudge factor. See git-range-
662 diff(1)) for details.
663 --notes[=<ref>], --no-notes
665 Append the notes (see git-notes(1)) for the commit after the
666 three-dash line.
667 The expected use case of this is to write supporting explanation
669 for the commit that does not belong to the commit log message
670 proper, and include it with the patch submission. While one can
671 simply write these explanations after format-patch has run but
672 before sending, keeping them as Git notes allows them to be
673 maintained between versions of the patch series (but see the
674 discussion of the notes.rewrite configuration options in git-
675 notes(1) to use this workflow).
676 The default is --no-notes, unless the format.notes configuration is
678 set.
679 --[no-]signature=<signature>
681 Add a signature to each message produced. Per RFC 3676 the
682 signature is separated from the body by a line with '-- ' on it. If
683 the signature option is omitted the signature defaults to the Git
684 version number.
685 --signature-file=<file>
687 Works just like --signature except the signature is read from a
688 file.
689 --suffix=.<sfx>
691 Instead of using .patch as the suffix for generated filenames, use
692 specified suffix. A common alternative is --suffix=.txt. Leaving
693 this empty will remove the .patch suffix.
694 Note that the leading character does not have to be a dot; for
696 example, you can use --suffix=-patch to get
697 0001-description-of-my-change-patch.
698 -q, --quiet
700 Do not print the names of the generated files to standard output.
701 --no-binary
703 Do not output contents of changes in binary files, instead display
704 a notice that those files changed. Patches generated using this
705 option cannot be applied properly, but they are still useful for
706 code review.
707 --zero-commit
709 Output an all-zero hash in each patch’s From header instead of the
710 hash of the commit.
711 --[no-]base[=<commit>]
713 Record the base tree information to identify the state the patch
714 series applies to. See the BASE TREE INFORMATION section below for
715 details. If <commit> is "auto", a base commit is automatically
716 chosen. The --no-base option overrides a format.useAutoBase
717 configuration.
718 --root
720 Treat the revision argument as a <revision range>, even if it is
721 just a single commit (that would normally be treated as a <since>).
722 Note that root commits included in the specified range are always
723 formatted as creation patches, independently of this flag.
724 --progress
726 Show progress reports on stderr as patches are generated.
727
729 You can specify extra mail header lines to be added to each message,
730 defaults for the subject prefix and file suffix, number patches when
731 outputting more than one patch, add "To:" or "Cc:" headers, configure
732 attachments, change the patch output directory, and sign off patches
733 with configuration variables.
734 [format]
736 headers = "Organization: git-foo\n"
737 subjectPrefix = CHANGE
738 suffix = .txt
739 numbered = auto
740 to = <email>
741 cc = <email>
742 attach [ = mime-boundary-string ]
743 signOff = true
744 outputDirectory = <directory>
745 coverLetter = auto
746 coverFromDescription = auto
747
749 The patch produced by git format-patch is in UNIX mailbox format, with
750 a fixed "magic" time stamp to indicate that the file is output from
751 format-patch rather than a real mailbox, like so:
752 From 8f72bad1baf19a53459661343e21d6491c3908d3 Mon Sep 17 00:00:00 2001
754 From: Tony Luck <tony.luck@intel.com>
755 Date: Tue, 13 Jul 2010 11:42:54 -0700
756 Subject: [PATCH] =?UTF-8?q?[IA64]=20Put=20ia64=20config=20files=20on=20the=20?=
757 =?UTF-8?q?Uwe=20Kleine-K=C3=B6nig=20diet?=
758 MIME-Version: 1.0
759 Content-Type: text/plain; charset=UTF-8
760 Content-Transfer-Encoding: 8bit
761 arch/arm config files were slimmed down using a python script
763 (See commit c2330e286f68f1c408b4aa6515ba49d57f05beae comment)
764 Do the same for ia64 so we can have sleek & trim looking
766 ...
767 Typically it will be placed in a MUA’s drafts folder, edited to add
769 timely commentary that should not go in the changelog after the three
770 dashes, and then sent as a message whose body, in our example, starts
771 with "arch/arm config files were...". On the receiving end, readers can
772 save interesting patches in a UNIX mailbox and apply them with git-
773 am(1).
774 When a patch is part of an ongoing discussion, the patch generated by
776 git format-patch can be tweaked to take advantage of the git am
777 --scissors feature. After your response to the discussion comes a line
778 that consists solely of "-- >8 --" (scissors and perforation), followed
779 by the patch with unnecessary header fields removed:
780 ...
782 > So we should do such-and-such.
783 Makes sense to me. How about this patch?
785 -- >8 --
787 Subject: [IA64] Put ia64 config files on the Uwe Kleine-König diet
788 arch/arm config files were slimmed down using a python script
790 ...
791 When sending a patch this way, most often you are sending your own
793 patch, so in addition to the "From $SHA1 $magic_timestamp" marker you
794 should omit From: and Date: lines from the patch file. The patch title
795 is likely to be different from the subject of the discussion the patch
796 is in response to, so it is likely that you would want to keep the
797 Subject: line, like the example above.
798 Checking for patch corruption
800 Many mailers if not set up properly will corrupt whitespace. Here are
801 two common types of corruption:
802 • Empty context lines that do not have any whitespace.
804 • Non-empty context lines that have one extra whitespace at the
806 beginning.
807 One way to test if your MUA is set up correctly is:
809 • Send the patch to yourself, exactly the way you would, except with
811 To: and Cc: lines that do not contain the list and maintainer
812 address.
813 • Save that patch to a file in UNIX mailbox format. Call it a.patch,
815 say.
816 • Apply it:
818 $ git fetch <project> master:test-apply
820 $ git switch test-apply
821 $ git restore --source=HEAD --staged --worktree :/
822 $ git am a.patch
823 If it does not apply correctly, there can be various reasons.
825 • The patch itself does not apply cleanly. That is bad but does not
827 have much to do with your MUA. You might want to rebase the patch
828 with git-rebase(1) before regenerating it in this case.
829 • The MUA corrupted your patch; "am" would complain that the patch
831 does not apply. Look in the .git/rebase-apply/ subdirectory and see
832 what patch file contains and check for the common corruption
833 patterns mentioned above.
834 • While at it, check the info and final-commit files as well. If what
836 is in final-commit is not exactly what you would want to see in the
837 commit log message, it is very likely that the receiver would end
838 up hand editing the log message when applying your patch. Things
839 like "Hi, this is my first patch.\n" in the patch e-mail should
840 come after the three-dash line that signals the end of the commit
841 message.
842
844 Here are some hints on how to successfully submit patches inline using
845 various mailers.
846 GMail
848 GMail does not have any way to turn off line wrapping in the web
849 interface, so it will mangle any emails that you send. You can however
850 use "git send-email" and send your patches through the GMail SMTP
851 server, or use any IMAP email client to connect to the google IMAP
852 server and forward the emails through that.
853 For hints on using git send-email to send your patches through the
855 GMail SMTP server, see the EXAMPLE section of git-send-email(1).
856 For hints on submission using the IMAP interface, see the EXAMPLE
858 section of git-imap-send(1).
859 Thunderbird
861 By default, Thunderbird will both wrap emails as well as flag them as
862 being format=flowed, both of which will make the resulting email
863 unusable by Git.
864 There are three different approaches: use an add-on to turn off line
866 wraps, configure Thunderbird to not mangle patches, or use an external
867 editor to keep Thunderbird from mangling the patches.
868 Approach #1 (add-on)
870 Install the Toggle Word Wrap add-on that is available from
871 https://addons.mozilla.org/thunderbird/addon/toggle-word-wrap/ It
872 adds a menu entry "Enable Word Wrap" in the composer’s "Options"
873 menu that you can tick off. Now you can compose the message as you
874 otherwise do (cut + paste, git format-patch | git imap-send, etc),
875 but you have to insert line breaks manually in any text that you
876 type.
877 Approach #2 (configuration)
879 Three steps:
880 1. Configure your mail server composition as plain text: Edit...
882 Account Settings...Composition & Addressing, uncheck "Compose
883 Messages in HTML".
884 2. Configure your general composition window to not wrap.
886 In Thunderbird 2: Edit..Preferences..Composition, wrap plain
888 text messages at 0
889 In Thunderbird 3: Edit..Preferences..Advanced..Config Editor.
891 Search for "mail.wrap_long_lines". Toggle it to make sure it is
892 set to false. Also, search for "mailnews.wraplength" and set
893 the value to 0.
894 3. Disable the use of format=flowed:
896 Edit..Preferences..Advanced..Config Editor. Search for
897 "mailnews.send_plaintext_flowed". Toggle it to make sure it is
898 set to false.
899 After that is done, you should be able to compose email as you
901 otherwise would (cut + paste, git format-patch | git imap-send,
902 etc), and the patches will not be mangled.
903 Approach #3 (external editor)
905 The following Thunderbird extensions are needed: AboutConfig from
906 http://aboutconfig.mozdev.org/ and External Editor from
907 http://globs.org/articles.php?lng=en&pg=8
908 1. Prepare the patch as a text file using your method of choice.
910 2. Before opening a compose window, use Edit→Account Settings to
912 uncheck the "Compose messages in HTML format" setting in the
913 "Composition & Addressing" panel of the account to be used to
914 send the patch.
915 3. In the main Thunderbird window, before you open the compose
917 window for the patch, use Tools→about:config to set the
918 following to the indicated values:
919 mailnews.send_plaintext_flowed => false
921 mailnews.wraplength => 0
922 4. Open a compose window and click the external editor icon.
924 5. In the external editor window, read in the patch file and exit
926 the editor normally.
927 Side note: it may be possible to do step 2 with about:config and
929 the following settings but no one’s tried yet.
930 mail.html_compose => false
932 mail.identity.default.compose_html => false
933 mail.identity.id?.compose_html => false
934 There is a script in contrib/thunderbird-patch-inline which can
936 help you include patches with Thunderbird in an easy way. To use
937 it, do the steps above and then use the script as the external
938 editor.
939 KMail
941 This should help you to submit patches inline using KMail.
942 1. Prepare the patch as a text file.
944 2. Click on New Mail.
946 3. Go under "Options" in the Composer window and be sure that "Word
948 wrap" is not set.
949 4. Use Message → Insert file... and insert the patch.
951 5. Back in the compose window: add whatever other text you wish to the
953 message, complete the addressing and subject fields, and press
954 send.
955
957 The base tree information block is used for maintainers or third party
958 testers to know the exact state the patch series applies to. It
959 consists of the base commit, which is a well-known commit that is part
960 of the stable part of the project history everybody else works off of,
961 and zero or more prerequisite patches, which are well-known patches in
962 flight that is not yet part of the base commit that need to be applied
963 on top of base commit in topological order before the patches can be
964 applied.
965 The base commit is shown as "base-commit: " followed by the 40-hex of
967 the commit object name. A prerequisite patch is shown as
968 "prerequisite-patch-id: " followed by the 40-hex patch id, which can be
969 obtained by passing the patch through the git patch-id --stable
970 command.
971 Imagine that on top of the public commit P, you applied well-known
973 patches X, Y and Z from somebody else, and then built your three-patch
974 series A, B, C, the history would be like:
975 ---P---X---Y---Z---A---B---C
977 With git format-patch --base=P -3 C (or variants thereof, e.g. with
979 --cover-letter or using Z..C instead of -3 C to specify the range), the
980 base tree information block is shown at the end of the first message
981 the command outputs (either the first patch, or the cover letter), like
982 this:
983 base-commit: P
985 prerequisite-patch-id: X
986 prerequisite-patch-id: Y
987 prerequisite-patch-id: Z
988 For non-linear topology, such as
990 ---P---X---A---M---C
992 \ /
993 Y---Z---B
994 You can also use git format-patch --base=P -3 C to generate patches for
996 A, B and C, and the identifiers for P, X, Y, Z are appended at the end
997 of the first message.
998 If set --base=auto in cmdline, it will track base commit automatically,
1000 the base commit will be the merge base of tip commit of the
1001 remote-tracking branch and revision-range specified in cmdline. For a
1002 local branch, you need to track a remote branch by git branch
1003 --set-upstream-to before using this option.
1004
1006 • Extract commits between revisions R1 and R2, and apply them on top
1007 of the current branch using git am to cherry-pick them:
1008 $ git format-patch -k --stdout R1..R2 | git am -3 -k
1010 • Extract all commits which are in the current branch but not in the
1012 origin branch:
1013 $ git format-patch origin
1015 For each commit a separate file is created in the current
1017 directory.
1018 • Extract all commits that lead to origin since the inception of the
1020 project:
1021 $ git format-patch --root origin
1023 • The same as the previous one:
1025 $ git format-patch -M -B origin
1027 Additionally, it detects and handles renames and complete rewrites
1029 intelligently to produce a renaming patch. A renaming patch reduces
1030 the amount of text output, and generally makes it easier to review.
1031 Note that non-Git "patch" programs won’t understand renaming
1032 patches, so use it only when you know the recipient uses Git to
1033 apply your patch.
1034 • Extract three topmost commits from the current branch and format
1036 them as e-mailable patches:
1037 $ git format-patch -3
1039
1041 git-am(1), git-send-email(1)
1042
1044 Part of the git(1) suite
1045Git 2.31.1 2021-03-26 GIT-FORMAT-PATCH(1)