GIT-DIFF-TREE(1)

🌸 NAME 🌸

git-diff-tree — it compares the content and the mode of blobs that it finds via two tree objects 💖 (yes, this is the GIT-DIFF-TREE(1) page from the Git Manual, and you're gonna love her)

🎀 SYNOPSIS 🎀

git diff-tree [--stdin] [-m] [-s] [-v] [--no-commit-id] [--pretty]
              [-t] [-r] [-c | --cc] [--combined-all-paths] [--root] [--merge-base]
              [<common-diff-options>] <tree-ish> [<tree-ish>] [<path>...]

✨ DESCRIPTION ✨

Okay so bestie, the whole vibe here is: compare the content and the mode of blobs found via two tree objects 🦋. That's the gig.

If you only hand it one <tree-ish>, then the commit gets compared with its parents instead — peep --stdin down below for the details 👀.

Also, small but important: git diff-tree can absolutely use the tree that's encapsulated inside a commit object. No cap 🫶.

💅 OPTIONS 💅

-p, -u, --patch — generate a patch (go read the section called "GENERATING PATCH TEXT WITH -P" for the full tea).
-s, --no-patch — suppress all output from the diff machinery 🙈. It's the move when you've got commands like git show that show the patch by default and you want to squelch their output, or when you need to cancel the effect of options like --patch, --stat that showed up earlier on the command line in an alias. Very "not today" energy.
-U<n>, --unified=<n> — generate diffs with <n> lines of context. The number of context lines defaults to diff.context, or 3 if that configuration variable is unset. (-U without <n> is silently accepted as a synonym for -p — it's a historical accident, we don't ask questions 💀). Implies --patch.
--output=<file> — output to a specific file instead of stdout. Fancy.
--output-indicator-new=<char>, --output-indicator-old=<char>, --output-indicator-context=<char> — pick the character used to indicate new, old or context lines in the generated patch ✨. Normally they're +, - and ' ' respectively.
--raw — generate the diff in raw format. This is the default, so she's literally the blueprint 👑.
--patch-with-raw — synonym for -p --raw.
--indent-heuristic — turns on the heuristic that shifts diff hunk boundaries around so patches are easier to read 🎀. This is the default.
--no-indent-heuristic — turns the indent heuristic off. Bye 👋.
--minimal — spend extra time to make sure the smallest possible diff is produced. Slow but thorough, main character energy 🐢✨.
--patience — generate a diff using the "patience diff" algorithm.
--histogram — generate a diff using the "histogram diff" algorithm.
--anchored=<text> — generate a diff using the "anchored diff" algorithm.

This option may be specified more than once 🔁.

If a line exists in both the source and destination, exists only once, and starts with <text>, this algorithm attempts to prevent it from appearing as a deletion or addition in the output. It uses the "patience diff" algorithm internally 🧚.

--diff-algorithm=(patience|minimal|histogram|myers) — choose a diff algorithm. The variants are as follows:
- default, myers — the basic greedy diff algorithm. Currently, this is the default.
- minimal — spend extra time to make sure the smallest possible diff is produced.
- patience — use the "patience diff" algorithm when generating patches.
- histogram — this algorithm extends the patience algorithm to "support low-occurrence common elements".

For instance, if you configured the diff.algorithm variable to a non-default value and want to use the default one, then you have to use --diff-algorithm=default option. iykyk 💫

--stat[=<width>[,<name-width>[,<count>]]] — generate a diffstat 📊. By default, as much space as necessary will be used for the filename part, and the rest for the graph part. Maximum width defaults to terminal width, or 80 columns if not connected to a terminal, and can be overridden by <width>. The width of the filename part can be limited by giving another width <name-width> after a comma, or by setting diff.statNameWidth=<name-width>. The width of the graph part can be limited by using --stat-graph-width=<graph-width> or by setting diff.statGraphWidth=<graph-width>. Using --stat or --stat-graph-width affects all commands generating a stat graph, while setting diff.statNameWidth or diff.statGraphWidth does not affect git format-patch. By giving a third parameter <count>, you can limit the output to the first <count> lines, followed by ... if there are more.

These parameters can also be set individually with --stat-width=<width>, --stat-name-width=<name-width> and --stat-count=<count>.

--compact-summary — output a condensed summary of extended header information, like file creations or deletions ("new" or "gone", optionally +l if it's a symlink) and mode changes (+x or -x for adding or removing the executable bit respectively) in the diffstat 🎯. The information gets put between the filename part and the graph part. Implies --stat.
--numstat — similar to --stat, but shows the number of added and deleted lines in decimal notation and the pathname without abbreviation, so it's way more machine friendly 🤖. For binary files, it outputs two - instead of saying 0 0.
--shortstat — output only the last line of the --stat format, containing the total number of modified files, plus the number of added and deleted lines.
-X [<param>,...], --dirstat[=<param>,...] — output the distribution of the relative amount of changes for each sub-directory 🗂️. The behavior of --dirstat can be customized by passing it a comma separated list of parameters. The defaults are controlled by the diff.dirstat configuration variable (see git-config(1)). The following parameters are available:
- changes — compute the dirstat numbers by counting the lines that have been removed from the source, or added to the destination. This ignores the amount of pure code movements within a file. In other words, rearranging lines in a file is not counted as much as other changes. This is the default behavior when no parameter is given.
- lines — compute the dirstat numbers by doing the regular line-based diff analysis, and summing the removed/added line counts. (For binary files, count 64-byte chunks instead, since binary files have no natural concept of lines). This is a more expensive --dirstat behavior than the changes behavior, but it does count rearranged lines within a file as much as other changes. The resulting output is consistent with what you get from the other --*stat options.
- files — compute the dirstat numbers by counting the number of files changed. Each changed file counts equally in the dirstat analysis. This is the computationally cheapest --dirstat behavior, since it does not have to look at the file contents at all 💤.
- cumulative — count changes in a child directory for the parent directory as well. Note that when using cumulative, the sum of the percentages reported may exceed 100%. The default (non-cumulative) behavior can be specified with the noncumulative parameter.
- <limit> — an integer parameter specifies a cut-off percent (3% by default). Directories contributing less than this percentage of the changes are not shown in the output.

Example: the following will count changed files, while ignoring directories with less than 10% of the total amount of changed files, and accumulating child directory counts in the parent directories: --dirstat=files,10,cumulative 🔥.

--cumulative — synonym for --dirstat=cumulative.
--dirstat-by-file[=<param>,...] — synonym for --dirstat=files,<param>,....
--summary — output a condensed summary of extended header information, like creations, renames and mode changes 📝.
--patch-with-stat — synonym for -p --stat.
-z — when --raw, --numstat, --name-only or --name-status has been given, do not munge pathnames and use NULs as output field terminators ✅.

Without this option, pathnames with "unusual" characters are quoted as explained for the configuration variable core.quotePath (see git-config(1)).

--name-only — show only the name of each changed file in the post-image tree. The file names are often encoded in UTF-8 🌈. For more information see the discussion about encoding in the git-log(1) manual page.
--name-status — show only the name(s) and status of each changed file. See the description of the --diff-filter option for what the status letters mean. Just like --name-only, the file names are often encoded in UTF-8.
--submodule[=<format>] — specify how differences in submodules are shown. When specifying --submodule=short the short format is used — this format just shows the names of the commits at the beginning and end of the range. When --submodule or --submodule=log is specified, the log format is used, which lists the commits in the range like git-submodule(1) summary does 📚. When --submodule=diff is specified, the diff format is used, showing an inline diff of the changes in the submodule contents between the commit range. Defaults to diff.submodule, or the short format if the config option is unset.
--color[=<when>] — show colored diff 🎨. --color (i.e. without =<when>) is the same as --color=always. <when> can be one of always, never, or auto.
--no-color — turn off colored diff. It is the same as --color=never.
--color-moved[=<mode>] — moved lines of code get colored differently 🦋. The <mode> defaults to no if the option is not given, and to zebra if the option is given with no mode. The mode must be one of:
- no — moved lines are not highlighted.
- default — is a synonym for zebra. This may change to a more sensible mode in the future 👀.
- plain — any line that is added in one location and was removed in another location will be colored with color.diff.newMoved. Similarly color.diff.oldMoved will be used for removed lines that are added somewhere else in the diff. This mode picks up any moved line, but it is not very useful in a review to determine if a block of code was moved without permutation.
- blocks — blocks of moved text of at least 20 alphanumeric characters are detected greedily. The detected blocks are painted using either the color.diff.(old|new)Moved color. Adjacent blocks cannot be told apart 💀.
zebra — blocks of moved text get detected just like in blocks mode, and then they're painted with either color.diff.(old|new)Moved or color.diff.(old|new)MovedAlternative. The switch between those two colors is your visual cue that a new block just got detected 💅
dimmed-zebra — same energy as zebra, but it also dims the uninteresting bits of the moved code. The lines bordering two adjacent blocks count as interesting; everything else is uninteresting. Also: dimmed_zebra is a deprecated synonym, so like… use the hyphen version bestie 🙈

--no-color-moved — turns move detection off entirely. Handy for overriding your config settings. It's literally the same thing as --color-moved=no.
--color-moved-ws=<mode>,... — this is the "how do we feel about whitespace" knob while doing the move detection for --color-moved. You can pass the modes as a comma separated list:
- no — do not ignore whitespace when performing move detection. Zero chill, full strictness.
- ignore-space-at-eol — ignore changes in whitespace at EOL.
- ignore-space-change — ignore changes in the amount of whitespace. This ignores whitespace at line end, and treats all other sequences of one or more whitespace characters as equivalent.
- ignore-all-space — ignore whitespace when comparing lines. This one ignores differences even if one line has whitespace where the other line has none.
- allow-indentation-change — initially ignore any whitespace in the move detection, then only group the moved code blocks into a block if the change in whitespace is the same per line. Heads up: this is incompatible with the other modes ⚠️
--no-color-moved-ws — do not ignore whitespace when performing move detection. Overrides configuration settings. Same as --color-moved-ws=no.

--word-diff[=<mode>] — by default, words are delimited by whitespace (peep --word-diff-regex below for the spicy version). <mode> defaults to plain and has to be one of these:
- color — highlight changed words using only colors. Implies --color.
- plain — show words as [-removed-] and {added}. It makes no attempt to escape the delimiters if they show up in the input, so the output may be ambiguous 😬
- porcelain — a special line-based format meant for script consumption. Added/removed/unchanged runs are printed in the usual unified diff format, starting with a +/-/ character at the beginning of the line and extending to the end of the line. Newlines in the input are represented by a tilde ~ on a line of its own.
- none — disable word diff again. We're done here.

Ngl, despite the name of the first mode, color gets used to highlight the changed parts in all modes if it's enabled.

Also important: --word-diff works by taking the same line-by-line diff you'd get without the option, then computing word-by-word changes inside each hunk. That means it may produce a larger diff than a dedicated word-diff tool would. And if Git gets a different implementation in the future, the output may change. It's similar to --diff-algorithm, which can also change the output. No cap.

--word-diff-regex=<regex> — use <regex> to decide what counts as a word, instead of just saying "runs of non-whitespace are a word". This also implies --word-diff unless you'd already turned it on.

Every non-overlapping match of the <regex> is considered a word. Anything between those matches counts as whitespace and gets ignored(!) when finding differences. You may want to append |[^[:space:]] to your regular expression to make sure it matches all non-whitespace characters. And a match that contains a newline gets silently truncated(!) at the newline 👀

For example, --word-diff-regex=. will treat each character as a word and, correspondingly, show differences character by character.

The regex can also be set via a diff driver or a configuration option — see gitattributes(5) or git-config(1). Giving it explicitly overrides any diff driver or configuration setting. And diff drivers override configuration settings. Hierarchy, periodt.

--color-words[=<regex>] — equivalent to --word-diff=color plus (if a regex was specified) --word-diff-regex=<regex>. Two-for-one deal 🎀

--no-renames — turn off rename detection, even when the config file says it should be on by default.
--rename-empty, --no-rename-empty — whether to use empty blobs as rename source. That's it, that's the whole vibe.
--check — warn if changes introduce conflict markers or whitespace errors. What counts as a whitespace error is controlled by the core.whitespace configuration. By default, trailing whitespaces (including lines that consist solely of whitespaces) and a space character immediately followed by a tab character inside the initial indent of the line are considered whitespace errors. Exits with non-zero status if problems are found. ⚠️ Not compatible with --exit-code.
--ws-error-highlight=<kind> — highlight whitespace errors in the context, old or new lines of the diff. Multiple values are separated by comma, none resets previous values, default resets the list to new, and all is a shorthand for old,new,context. When this option isn't given, and the configuration variable diff.wsErrorHighlight isn't set, only whitespace errors in new lines get highlighted. These whitespace errors are colored with color.diff.whitespace.
--full-index — instead of just the first handful of characters, show the full pre- and post-image blob object names on the index line when generating patch format output. Show it all, hide nothing ✨
--binary — in addition to --full-index, output a binary diff that can be applied with git-apply. Implies --patch.
--abbrev[=<n>] — instead of the full 40-byte hexadecimal object name in diff-raw format output and diff-tree header lines, show the shortest prefix that's at least <n> hexdigits long and still uniquely refers to the object. In diff-patch output format, --full-index takes higher precedence — i.e. if --full-index is specified, full blob names are shown regardless of --abbrev. A non-default number of digits can be specified with --abbrev=<n>.

-B[<n>][/<m>], --break-rewrites[=[<n>][/<m>]] — break complete rewrite changes into pairs of delete and create. This serves two purposes:

It affects the way a change that amounts to a total rewrite of a file is treated — not as a series of deletion and insertion mixed together with a very few lines that happen to match textually as the context, but as a single deletion of everything old followed by a single insertion of everything new. The number <m> controls this aspect of the -B option (defaults to 60%). -B/70% specifies that less than 30% of the original should remain in the result for Git to consider it a total rewrite (i.e. otherwise the resulting patch will be a series of deletion and insertion mixed together with context lines).

When used with -M, a totally-rewritten file is also considered as the source of a rename (usually -M only considers a file that disappeared as the source of a rename), and the number <n> controls this aspect of the -B option (defaults to 50%). -B20% specifies that a change with addition and deletion compared to 20% or more of the file's size are eligible for being picked up as a possible source of a rename to another file. 🔥

-M[<n>], --find-renames[=<n>] — detect renames. If <n> is specified, it's a threshold on the similarity index (i.e. amount of addition/deletions compared to the file's size). For example, -M90% means Git should consider a delete/add pair to be a rename if more than 90% of the file hasn't changed. Without a % sign, the number is read as a fraction, with a decimal point before it — i.e. -M5 becomes 0.5, and is thus the same as -M50%. Similarly, -M05 is the same as -M5%. To limit detection to exact renames, use -M100%. The default similarity index is 50%.
-C[<n>], --find-copies[=<n>] — detect copies as well as renames. Also peep --find-copies-harder. If <n> is specified, it means the same thing as it does for -M<n>.
--find-copies-harder — for performance reasons, by default the -C option finds copies only if the original file of the copy was modified in the same changeset. This flag makes the command inspect unmodified files as candidates for the source of a copy. ⚠️ This is a very expensive operation for large projects, so use it with caution. Giving more than one -C option has the same effect.
-D, --irreversible-delete — omit the preimage for deletes, i.e. print only the header but not the diff between the preimage and /dev/null. The resulting patch is not meant to be applied with patch or git apply — this is solely for people who want to concentrate on reviewing the text after the change. Plus the output obviously lacks enough information to apply such a patch in reverse, even manually, hence the name of the option 💀

When used together with -B, omit also the preimage in the deletion part of a delete/create pair.

-l<num> — the -M and -C options involve some preliminary steps that can detect subsets of renames/copies cheaply, followed by an exhaustive fallback portion that compares all remaining unpaired destinations to all relevant sources. (For renames, only remaining unpaired sources are relevant; for copies, all original sources are relevant.) For N sources and destinations, this exhaustive check is O(N^2). This option prevents the exhaustive portion of rename/copy detection from running if the number of source/destination files involved exceeds the specified number. Defaults to diff.renameLimit. Note that a value of 0 is treated as unlimited 🌀
--diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]] — select only files that are Added (A), Copied (C), Deleted (D), Modified (M), Renamed (R), have their type (i.e. regular file, symlink, submodule, ...) changed (T), are Unmerged (U), are Unknown (X), or have had their pairing Broken (B). Any combination of the filter characters can be used (including none at all). And when * (All-or-none) is added to the combination, all paths are selected if there's any file that matches other criteria in the comparison; if there's no file that matches other criteria, nothing is selected. Iconic behavior honestly 👑

okay so also — those upper-case letters? you can downcase them to exclude. like --diff-filter=ad excludes added and deleted paths. 🎀

and a heads-up bestie: not every diff can feature every type. for instance, copied and renamed entries literally cannot appear if detection for those types is disabled. no cap.

-S<string> — looks for differences that change the number of occurrences of the specified <string> in a file (i.e. addition/deletion). made for the scripter's use, fr.

it's genuinely the move when you're hunting for an exact block of code (like a struct) and wanna know the history of that block since it first came into being: use the feature iteratively to feed the interesting block in the preimage back into -S, and keep going until you get the very first version of the block. 🕵️‍♀️

binary files are searched as well. yes, even those.

-G<regex> — looks for differences whose patch text contains added/removed lines that match <regex>.

to illustrate the difference between -S<regex> --pickaxe-regex and -G<regex>, consider a commit with the following diff in the same file:

    +    return frotz(nitfol, two->ptr, 1, 0);
    ...
    -    hit = frotz(nitfol, mf2.ptr, 1, 0);

so while git log -G"frotz\(nitfol" will show this commit, git log -S"frotz\(nitfol" --pickaxe-regex will not (because the number of occurrences of that string did not change). the way that works 💅

unless --text is supplied, patches of binary files without a textconv filter will be ignored.

see the pickaxe entry in gitdiffcore(7) for more information.

--find-object=<object-id> — looks for differences that change the number of occurrences of the specified object. similar to -S, just the argument is different in that it doesn't search for a specific string but for a specific object id. iconic.

the object can be a blob or a submodule commit. it implies the -t option in git-log to also find trees.

--pickaxe-all — when -S or -G finds a change, show all the changes in that changeset, not just the files that contain the change in <string>. main character energy for the whole commit.
--pickaxe-regex — treat the <string> given to -S as an extended POSIX regular expression to match.
-O<orderfile> — controls the order in which files appear in the output. this overrides the diff.orderFile configuration variable (see git-config(1)). to cancel diff.orderFile, use -O/dev/null. 🗂️

the output order is determined by the order of glob patterns in <orderfile>. all files with pathnames that match the first pattern are output first, all files with pathnames that match the second pattern (but not the first) are output next, and so on. all files with pathnames that do not match any pattern are output last, as if there was an implicit match-all pattern at the end of the file. if multiple pathnames have the same rank (they match the same pattern but no earlier patterns), their output order relative to each other is the normal order.

<orderfile> is parsed as follows:

blank lines are ignored, so they can be used as separators for readability.
lines starting with a hash (#) are ignored, so they can be used for comments. add a backslash (\) to the beginning of the pattern if it starts with a hash.
each other line contains a single pattern.

patterns have the same syntax and semantics as patterns used for fnmatch(3) without the FNM_PATHNAME flag, except a pathname also matches a pattern if removing any number of the final pathname components matches the pattern. for example, the pattern "foo*bar" matches "fooasdfbar" and "foo/bar/baz/asdf" but not "foobarx". 🔥

--skip-to=<file>, --rotate-to=<file> — discard the files before the named <file> from the output (i.e. skip to), or move them to the end of the output (i.e. rotate to). these options were invented primarily for the use of the git difftool command, and may not be very useful otherwise. ngl, niche but cute.
-R — swap two inputs; that is, show differences from index or on-disk file to tree contents. flip it and reverse it. 🔄
--relative[=<path>], --no-relative — when run from a subdirectory of the project, it can be told to exclude changes outside the directory and show pathnames relative to it with this option. when you are not in a subdirectory (e.g. in a bare repository), you can name which subdirectory to make the output relative to by giving a <path> as an argument. --no-relative can be used to countermand both the diff.relative config option and a previous --relative.
-a, --text — treat all files as text. all of them. periodt.
--ignore-cr-at-eol — ignore carriage-return at the end of line when doing a comparison.
--ignore-space-at-eol — ignore changes in whitespace at EOL.
-b, --ignore-space-change — ignore changes in amount of whitespace. this ignores whitespace at line end, and considers all other sequences of one or more whitespace characters to be equivalent.
-w, --ignore-all-space — ignore whitespace when comparing lines. this ignores differences even if one line has whitespace where the other line has none. literally unbothered. ✨
--ignore-blank-lines — ignore changes whose lines are all blank.
-I<regex>, --ignore-matching-lines=<regex> — ignore changes whose all lines match <regex>. this option may be specified more than once. stack 'em 💅
--inter-hunk-context=<number> — show the context between diff hunks, up to the specified <number> of lines, thereby fusing hunks that are close to each other. defaults to diff.interHunkContext or 0 if the config option is unset.
-W, --function-context — show whole function as context lines for each change. the function names are determined in the same way as git diff works out patch hunk headers (see "Defining a custom hunk-header" in gitattributes(5)). the whole function, bestie. 🫶
--exit-code — make the program exit with codes similar to diff(1). that is, it exits with 1 if there were differences and 0 means no differences.
--quiet — disable all output of the program. implies --exit-code. disables execution of external diff helpers whose exit code is not trusted, i.e. their respective configuration option diff.trustExitCode or diff.<driver>.trustExitCode or environment variable GIT_EXTERNAL_DIFF_TRUST_EXIT_CODE is false. 🤫
--ext-diff — allow an external diff helper to be executed. if you set an external diff driver with gitattributes(5), you need to use this option with git-log(1) and friends.
--no-ext-diff — disallow external diff drivers. the opposite vibe.
--textconv, --no-textconv — allow (or disallow) external text conversion filters to be run when comparing binary files. see gitattributes(5) for details. because textconv filters are typically a one-way conversion, the resulting diff is suitable for human consumption, but cannot be applied. for this reason, textconv filters are enabled by default only for git-diff(1) and git-log(1), but not for git-format-patch(1) or diff plumbing commands. important!! read it twice 🚨
--ignore-submodules[=(none|untracked|dirty|all)] — ignore changes to submodules in the diff generation. all is the default. using none will consider the submodule modified when it either contains untracked or modified files or its HEAD differs from the commit recorded in the superproject, and can be used to override any settings of the ignore option in git-config(1) or gitmodules(5). when untracked is used, submodules are not considered dirty when they only contain untracked content (but they are still scanned for modified content). using dirty ignores all changes to the work tree of submodules, only changes to the commits stored in the superproject are shown (this was the behavior until 1.7.0). using all hides all changes to submodules. so many flavours 🍓
--src-prefix=<prefix> — show the given source <prefix> instead of "a/".
--dst-prefix=<prefix> — show the given destination <prefix> instead of "b/".
--no-prefix — do not show any source or destination prefix. bare. minimal. 💀
--default-prefix — use the default source and destination prefixes ("a/" and "b/"). this overrides configuration variables such as diff.noprefix, diff.srcPrefix, diff.dstPrefix, and diff.mnemonicPrefix (see git-config(1)).
--line-prefix=<prefix> — prepend an additional <prefix> to every line of output.
--ita-invisible-in-index — by default entries added by git add -N appear as an existing empty file in git diff and a new file in git diff --cached. this option makes the entry appear as a new file in git diff and non-existent in git diff --cached. this option could be reverted with --ita-visible-in-index. both options are experimental and could be removed in future. 🙈 you've been warned.
--max-depth=<depth> — for each pathspec given on command line, descend at most <depth> levels of directories. a value of -1 means no limit. cannot be combined with wildcards in the pathspec. given a tree containing foo/bar/baz, the following list shows the matches generated by each set of options:
- --max-depth=0 -- foo: foo
- --max-depth=1 -- foo: foo/bar
- --max-depth=1 -- foo/bar: foo/bar/baz
- --max-depth=1 -- foo foo/bar: foo/bar/baz
- --max-depth=2 -- foo: foo/bar/baz

if no pathspec is given, the depth is measured as if all top-level entries were specified. note that this is different than measuring from the root, in that --max-depth=0 would still return foo. this allows you to still limit depth while asking for a subset of the top-level entries. 🌈

note that this option is only supported for diffs between tree objects, not against the index or working tree. don't try it, bestie.

for more detailed explanation on these common options, see also gitdiffcore(7).

<tree-ish> — the id of a tree object.
<path>... — if provided, the results are limited to a subset of files matching one of the provided pathspecs.
-r — recurse into sub-trees. go deeper 🌀
-t — show tree entry itself as well as subtrees. implies -r.
--root — when --root is specified the initial commit will be shown as a big creation event. this is equivalent to a diff against the NULL tree. genesis era ✨
--merge-base — instead of comparing the <tree-ish>s directly, use the merge base between the two <tree-ish>s as the "before" side. there must be two <tree-ish>s given and they must both be commits.
--stdin — okay so with this one the command doesn't take <tree-ish> args from the command line at all. Instead it reads lines from standard input, and each line is either two <tree>s, one <commit>, or a whole list of *<commit>*s. (Separate them with a single space, bestie — one. space. 💅)

When you hand it two trees, it compares the first tree to the second. When you hand it a single commit, it compares that commit against its parents. And any remaining commits after the first one? They get treated as if they were parents of that first commit. 🫶

When it's comparing two trees, the ID of both trees gets printed before the difference (separated by a space, terminated by a newline). When it's comparing commits, you get the ID of the first (or only) commit, followed by a newline.

The following flags only change stuff when comparing commits — not trees. Keep that in mind ✨

-m — by default git diff-tree --stdin is not showing differences for merge commits (rude, honestly). With this flag it shows differences to that commit from all of its parents. See also -c. 🌈
-s — by default git diff-tree --stdin shows differences, either machine-readable (no -p) or patch form (with -p). This flag suppresses that output entirely. Ngl it's only actually useful alongside the -v flag. 💀
-v — this one makes git diff-tree --stdin also show the commit message before the differences. Main character energy: message first, diff after. 👑
--pretty[=<format>], --format=<format> — pretty-prints the contents of the commit logs in whatever format you want, where <format> is one of oneline, short, medium, full, fuller, reference, email, raw, format:<string> and tformat:<string>. And if <format> is none of those but does contain %<placeholder>, it acts like --pretty=tformat:<format> was passed. Sneaky but iconic.

Peep the PRETTY FORMATS section for extra deets on each format. And if the =<format> part is just... omitted? It defaults to medium. 🎀

Note — you can set the default pretty format in the repository configuration (see git-config(1)).

--abbrev-commit — instead of showing the full 40-byte hexadecimal commit object name, show a prefix that names the object uniquely. You can use --abbrev=<n> to set the minimum length of that prefix (and yes, it also modifies diff output if a diff is displayed). This makes --pretty=oneline so much more readable for anyone on an 80-column terminal. Our eyes thank you 🙈
--no-abbrev-commit — show the full 40-byte hexadecimal commit object name. This negates --abbrev-commit, whether you set it explicitly or it got implied by other options like --oneline. It also overrides the log.abbrevCommit variable. Full name, no shortening, periodt. ⭐
--oneline — a lil shorthand for --pretty=oneline --abbrev-commit used together. Two-for-one, slay. 🦋
--encoding=<encoding> — commit objects record the character encoding used for the log message in their encoding header, and this option tells the command to re-code the commit log message into the encoding you (the user) prefer. For non plumbing commands this defaults to UTF-8. 💖

Heads up though: if an object claims to be encoded in X and we're outputting in X, we output the object verbatim — which means invalid sequences in the original commit may get copied straight to the output. Same deal if iconv(3) fails to convert the commit: we quietly output the original object verbatim. No drama, no warning, just vibes (and possibly broken characters 😭).

--expand-tabs=<n>, --expand-tabs, --no-expand-tabs — performs tab expansion in the log message before showing it in the output: each tab gets replaced with enough spaces to fill to the next display column that's a multiple of <n>. --expand-tabs is shorthand for --expand-tabs=8, and --no-expand-tabs is shorthand for --expand-tabs=0, which disables tab expansion completely. Math but make it aesthetic ✨

By default, tabs get expanded in pretty formats that indent the log message by 4 spaces — that's medium (the default), full, and fuller. 🌸

--notes[=<ref>] — show the notes (see git-notes(1)) that annotate the commit when showing the commit log message. This is the default for git log, git show and git whatchanged commands when there's no --pretty, --format, or --oneline option on the command line.

By default the notes shown come from the notes refs listed in the core.notesRef and notes.displayRef variables (or corresponding environment overrides). See git-config(1) for more deets.

With the optional <ref> argument, use that ref to find the notes to display. The ref can specify the full refname when it begins with refs/notes/; when it begins with notes/, refs/ gets prefixed to it; otherwise refs/notes/ gets prefixed to form the full name of the ref. ✨

You can stack multiple --notes options to control which notes get displayed. Examples, because bestie loves examples: "--notes=foo" shows only notes from refs/notes/foo; "--notes=foo --notes" shows both notes from "refs/notes/foo" and from the default notes ref(s). Combo moves 🫶

--no-notes — do not show notes. This negates the --notes option above by resetting the list of notes refs from which notes are shown. Options get parsed in the order given on the command line, so like... "--notes --notes=foo --no-notes --notes=bar" will only show notes from refs/notes/bar. Order matters, queen 👑
--show-notes-by-default — show the default notes unless options for displaying specific notes are given. 💅
--show-notes[=<ref>], --standard-notes, --no-standard-notes — these are deprecated (iykyk, they had their era 💀). Use the --notes/--no-notes options above instead.
--show-signature — checks the validity of a signed commit object by passing the signature to gpg --verify and shows the output. Detective mode: activated 🔥
--no-commit-id — git diff-tree outputs a line with the commit ID when applicable. This flag suppresses that commit ID output. Gone. Poof. 🦋
-c — this flag changes the way a merge commit is displayed (which means it's only useful when the command is given one <tree-ish>, or --stdin). It shows the differences from each of the parents to the merge result simultaneously, instead of showing a pairwise diff between a parent and the result one at a time (which is what -m does). On top of that, it lists only files which were modified from all parents. Simultaneous slay 🌈
--cc — changes the way a merge commit patch is displayed, similar to -c. It implies the -c and -p options, and further compresses the patch output by omitting uninteresting hunks — the ones whose contents in the parents have only two variants and the merge result just picks one of them without modification. And when all hunks are uninteresting? The commit itself and the commit log message aren't shown at all, just like in any other "empty diff" case. Nothing to see here, literally 💀
--combined-all-paths — makes combined diffs (the merge-commit kind) list the name of the file from all parents. So it only has any effect when -c or --cc are specified, and it's likely only useful if filename changes are detected (i.e. when either rename or copy detection have been requested). Niche but cute 🎀
--always — show the commit itself and the commit log message even if the diff itself is empty. Empty diff, full commitment. 🫶

✨ PRETTY FORMATS ✨

If the commit is a merge, and if the pretty-format is not oneline, email or raw, an extra line gets inserted before the Author: line. That line starts with "Merge: " and then the hashes of ancestral commits get printed, separated by spaces. 🌸

Little caveat bestie: the listed commits may not necessarily be the list of the direct parent commits if you've limited your view of history — like, say, if you're only interested in changes related to a certain directory or file. 💅

There are several built-in formats, and you can define extra formats by setting a pretty.<name> config option to either another format name, or a format: string, as described below (see git-config(1)). Here's the tea on the built-in formats:

`oneline`

<hash> <title-line>

Designed to be as compact as possible. Minimalist queen ⭐

`short`

commit <hash>
Author: <author>

    <title-line>

`medium`

commit <hash>
Author: <author>
Date:   <author-date>

    <title-line>

    <full-commit-message>

`full`

commit <hash>
Author: <author>
Commit: <committer>

    <title-line>

    <full-commit-message>

`fuller`

commit <hash>
Author:     <author>
AuthorDate: <author-date>
Commit:     <committer>
CommitDate: <committer-date>

     <title-line>

     <full-commit-message>

`reference`

<abbrev-hash> (<title-line>, <short-author-date>)

This format is used to refer to another commit in a commit message, and it's the same as --pretty='format:%C(auto)%h (%s, %ad)'. By default the date is formatted with --date=short unless another --date option is explicitly specified. And just like any format: with format placeholders, its output is not affected by other options like --decorate and --walk-reflogs. Untouchable 💅

`email`

From <hash> <date>
From: <author>
Date: <author-date>
Subject: [PATCH] <title-line>

<full-commit-message>

`mboxrd`

Like email, but lines in the commit message starting with "From " (preceded by zero or more ">") are quoted with ">" so they aren't confused as starting a new commit. Protecting the timeline 🙈

`raw`

The raw format shows the entire commit exactly as stored in the commit object. Notably, the hashes are displayed in full, regardless of whether --abbrev or --no-abbrev are used, and parents information shows the true parent commits, without taking grafts or history simplification into account. 🫶

Note that this format affects the way commits are displayed, but not the way the diff is shown — e.g. with git log --raw. To get full object names in a raw diff format, use --no-abbrev.

`format:<format-string>`

The format:<format-string> format lets you specify exactly which information you want to show. It works a little bit like printf format, with the notable exception that you get a newline with %n instead of \n. 🎀

E.g, format:"The author of %h was %an, %ar%nThe title was >>%s<<%n" would show something like this:

The author of fe6e0ee was Junio C Hamano, 23 hours ago
The title was >>t4119: test autocomputing -p<n> for traditional diff input.<<

The placeholders are:

Placeholders that expand to a single literal character:
- %n — newline
- %% — a raw %
%x00 — %x plus two hexadecimal digits gets swapped out for the byte those digits spell out (we're calling this a "literal formatting code" for the rest of this doc, bestie, so don't forget it 💅).
Placeholders that change up how the later placeholders look 👀:
- %Cred — flips the color to red
- %Cgreen — flips the color to green
- %Cblue — flips the color to blue
- %Creset — resets the color back to normal, phew
- %C(<spec>) — a color spec, exactly like the ones described under Values in the "CONFIGURATION FILE" section of git-config(1). By default, colors only show up when color is enabled for log output (via color.diff, color.ui, or --color, and respecting the auto settings of those first ones if we're heading to a terminal). %C(auto,<spec>) is accepted as a historical synonym for the default (e.g., %C(auto,red)). If you specify %C(always,<spec>) the colors show even when color isn't otherwise enabled — though honestly consider just using --color=always to turn on color for the whole entire output, including this format and anything else git might color 🎨. And auto all by itself (i.e. %C(auto)) turns on auto coloring for the next placeholders until the color gets switched again, periodt.
- %m — the mark: left (<), right (>), or boundary (-)
- %w([<w>[,<i1>[,<i2>]]]) — switches up the line wrapping, just like the -w option of git-shortlog(1)
- %<(<n>[,(trunc|ltrunc|mtrunc)]) — makes the next placeholder take at least N column widths, padding with spaces on the right if it needs to. Optionally truncate (with an ellipsis ..) at the left (ltrunc) so it's ..ft, the middle (mtrunc) so it's mi..le, or the end (trunc) so it's rig.., if the output runs longer than <n> columns. Note 1: truncating only works correctly with <n> >= 2. Note 2: the spaces around the <n> and <m> (see below) values are optional. Note 3: emojis and other wide characters take up two display columns, which may over-run column boundaries 😭. Note 4: decomposed character combining marks may end up misplaced at padding boundaries.
- %<|(<m> ) — makes the next placeholder take at least until the <m>th display column, padding spaces on the right if necessary. Use negative <m> values for column positions measured from the right hand edge of the terminal window 🖥️
- %>(<n>), %>|(<m>) — same vibe as %<(<n>) and %<|(<m>) respectively, except the padding spaces go on the left
- %>>(<n>), %>>|(<m>) — same as %>(<n>) and %>|(<m>) respectively, except if the next placeholder needs more spaces than given and there are spaces on its left, use those spaces instead
- %><(<n>), %><|(<m>) — same as %<(<n>) and %<|(<m>) respectively, but padding both sides so the text is centered 💅 iconic
Placeholders that expand into info pulled straight from the commit 🫶:
- %H — commit hash
- %h — abbreviated commit hash
- %T — tree hash
- %t — abbreviated tree hash
- %P — parent hashes
- %p — abbreviated parent hashes
- %an — author name
- %aN — author name (respecting .mailmap, see git-shortlog(1) or git-blame(1))
- %ae — author email
- %aE — author email (respecting .mailmap, see git-shortlog(1) or git-blame(1))
- %al — author email local-part (the part before the @ sign)
- %aL — author local-part (see %al) respecting .mailmap, see git-shortlog(1) or git-blame(1))
- %ad — author date (format respects the --date= option)
- %aD — author date, RFC2822 style
- %ar — author date, relative
- %at — author date, UNIX timestamp
- %ai — author date, ISO 8601-like format
- %aI — author date, strict ISO 8601 format
- %as — author date, short format (YYYY-MM-DD)
- %ah — author date, human style (like the --date=human option of git-rev-list(1))
- %cn — committer name
- %cN — committer name (respecting .mailmap, see git-shortlog(1) or git-blame(1))
- %ce — committer email
- %cE — committer email (respecting .mailmap, see git-shortlog(1) or git-blame(1))
- %cl — committer email local-part (the part before the @ sign)
- %cL — committer local-part (see %cl) respecting .mailmap, see git-shortlog(1) or git-blame(1))
- %cd — committer date (format respects the --date= option)
- %cD — committer date, RFC2822 style
- %cr — committer date, relative
- %ct — committer date, UNIX timestamp
- %ci — committer date, ISO 8601-like format
- %cI — committer date, strict ISO 8601 format
- %cs — committer date, short format (YYYY-MM-DD)
- %ch — committer date, human style (like the --date=human option of git-rev-list(1))
- %d — ref names, like the --decorate option of git-log(1)
- %D — ref names without the " (" and ")" wrapping
- %(count) — the number of a patch within a patch series. Used only in --commit-list-format in format-patch
- %(total) — the total number of patches in a patch series. Used only in --commit-list-format in format-patch
- %(decorate[:<option>,...]) — ref names with custom decorations ✨. The decorate string can be followed by a colon and zero or more comma-separated options. Option values may contain literal formatting codes — and those must be used for commas (%x2C) and closing parentheses (%x29), because of their role in the option syntax, no cap.
  - prefix=<value> — shown before the list of ref names. Defaults to " (".
  - suffix=<value> — shown after the list of ref names. Defaults to ")".
  - separator=<value> — shown between ref names. Defaults to ", ".
  - pointer=<value> — shown between HEAD and the branch it points to, if any. Defaults to " → ".
  - tag=<value> — shown before tag names. Defaults to "tag: ".

For example, to produce decorations with no wrapping or tag annotations, and spaces as separators:

%(decorate:prefix=,suffix=,tag=,separator= )

%(describe[:<option>,...]) — human-readable name, like git-describe(1); an empty string for undescribable commits. The describe string can be followed by a colon and zero or more comma-separated options. Descriptions can be inconsistent when tags are added or removed at the same time 🫠.
- tags[=<bool-value>] — instead of only considering annotated tags, consider lightweight tags as well.
- abbrev=<number> — instead of using the default number of hexadecimal digits (which varies according to the number of objects in the repository, with a default of 7) of the abbreviated object name, use <number> digits, or as many digits as needed to form a unique object name.
- match=<pattern> — only consider tags matching the given glob(7) <pattern>, excluding the refs/tags/ prefix.
- exclude=<pattern> — do not consider tags matching the given glob(7) <pattern>, excluding the refs/tags/ prefix.
%S — ref name given on the command line by which the commit was reached (like git log --source); only works with git log
%e — encoding
%s — subject
%f — sanitized subject line, suitable for a filename
%b — body
%B — raw body (unwrapped subject and body)
%N — commit notes
%GG — raw verification message from GPG for a signed commit
%G? — shows "G" for a good (valid) signature, "B" for a bad signature, "U" for a good signature with unknown validity, "X" for a good signature that has expired, "Y" for a good signature made by an expired key, "R" for a good signature made by a revoked key, "E" if the signature cannot be checked (e.g. missing key), and "N" for no signature
%GS — shows the name of the signer for a signed commit
%GK — shows the key used to sign a signed commit
%GF — shows the fingerprint of the key used to sign a signed commit
%GP — shows the fingerprint of the primary key whose subkey was used to sign a signed commit
%GT — shows the trust level for the key used to sign a signed commit
%gD — reflog selector, e.g. refs/stash@{1} or refs/stash@{2 minutes ago}; the format follows the rules described for the -g option. The portion before the @ is the refname as given on the command line (so git log -g refs/heads/master would yield refs/heads/master@{0}).
%gd — shortened reflog selector; same as %gD, but the refname portion gets shortened for human readability (so refs/heads/master becomes just master).
%gn — reflog identity name
%gN — reflog identity name (respecting .mailmap, see git-shortlog(1) or git-blame(1))
%ge — reflog identity email

%gE — the reflog identity email (and yeah, it respects .mailmap, peek git-shortlog(1) or git-blame(1) for the deets) 💌
%gs — the reflog subject, aka the vibe of that reflog entry ✨
%(trailers[:<option>,...]) — shows the trailers of the body exactly how git-interpret-trailers(1) reads them 👑. You can slap a colon after it and then zero or more comma-separated options. If you give the same option more than once, the last one wins — no cap, that's just how it is.
- key=<key> — only show trailers with that <key>. Matching is case-insensitive and the trailing colon is optional 🎀. Give it multiple times and it shows trailer lines matching any of the keys. Lowkey sneaky: this option auto-enables only, so non-trailer lines in the trailer block get hidden. Not the vibe? Turn it off with only=false. Example: %(trailers:key=Reviewed-by) shows the trailer lines with key Reviewed-by.
- only[=<bool>] — picks whether non-trailer lines from the trailer block get included or not.
- separator=<sep> — sets the separator dropped between trailer lines. Defaults to a line feed character. <sep> can contain the literal formatting codes from above 🫶. Wanna use a comma as separator? You literally have to write %x2C, otherwise it gets parsed as the next option. Example: %(trailers:key=Ticket,separator=%x2C ) shows all trailer lines whose key is Ticket, separated by a comma and a space.
- unfold[=<bool>] — makes it act like interpret-trailer's --unfold option was given. Example: %(trailers:only,unfold=true) unfolds and shows all the trailer lines 🦋.
- keyonly[=<bool>] — only show the key part of the trailer.
- valueonly[=<bool>] — only show the value part of the trailer.
- key_value_separator=<sep> — sets the separator between the key and value of each trailer. Defaults to ": ". Otherwise it's the same semantics as separator=<sep> above 💅.

Note 📌 Some placeholders depend on other options you hand the revision traversal engine. For example, the %g* reflog options insert an empty string unless we're traversing reflog entries (like with git log -g). And %d / %D use the "short" decoration format if --decorate wasn't already passed on the command line. iykyk.

The boolean options all accept an optional value [=<bool-value>]. Every value that --type=bool git-config(1) accepts — stuff like yes and off — is fair game. Giving a boolean option with no =<value> is the same as giving it =true. Slay 💅

If you put a + (plus sign) after the % of a placeholder, a line-feed gets inserted right before the expansion if and only if the placeholder expands to a non-empty string 🌸

If you put a - (minus sign) after the % of a placeholder, all the consecutive line-feeds immediately before the expansion get deleted if and only if the placeholder expands to an empty string ✂️

If you put a ' ' (space) after the % of a placeholder, a space gets inserted right before the expansion if and only if the placeholder expands to a non-empty string 🎀

tformat:

The tformat: format works exactly like format:, except it gives you "terminator" semantics instead of "separator" semantics. Translation: each commit gets the message terminator character (usually a newline) appended, instead of a separator being placed between entries. Which means the final entry of a single-line format is properly terminated with a newline, just like the "oneline" format does 💖 For example:

$ git log -2 --pretty=format:%h 4da45bef \
  | perl -pe '$_ .= " -- NO NEWLINE\n" unless /\n/'
4da45be
7134973 -- NO NEWLINE

$ git log -2 --pretty=tformat:%h 4da45bef \
  | perl -pe '$_ .= " -- NO NEWLINE\n" unless /\n/'
4da45be
7134973

Also: any unrecognized string that has a % in it is interpreted as if it has tformat: in front of it 👀 So these two are straight-up equivalent:

$ git log -2 --pretty=tformat:%h 4da45bef
$ git log -2 --pretty=%h 4da45bef

🌈 RAW OUTPUT FORMAT 🌈

So the raw output format from git-diff-index, git-diff-tree, git-diff-files and git diff --raw are all very similar — like siblings, not twins 💫

All these commands compare two sets of things; what gets compared is where they differ:

git-diff-index <tree-ish> — compares the <tree-ish> and the files on the filesystem.
git-diff-index --cached <tree-ish> — compares the <tree-ish> and the index.
git-diff-tree [-r] <tree-ish-1> <tree-ish-2> [<pattern>...] — compares the trees named by the two arguments.
git-diff-files [<pattern>...] — compares the index and the files on the filesystem.

All the commands print one output line per changed file — except git diff-files in the case of an unmerged file, which prints both an "unmerged" and an "in-place edit" line 😮‍💨

An output line is formatted like this:

in-place edit  :100644 100644 bcd1234 0123456 M file0
copy-edit      :100644 100644 abcd123 1234567 C68 file1 file2
rename-edit    :100644 100644 abcd123 1234567 R86 file1 file3
create         :000000 100644 0000000 1234567 A file4
delete         :100644 000000 1234567 0000000 D file5
unmerged       :000000 000000 0000000 0000000 U file6

So like, reading it left to right 🫶

a colon.
mode for "src"; 000000 if creation, or if "src" is from the index and is unmerged.
a space.
mode for "dst"; 000000 if deletion, or if "dst" is from the index and is unmerged.
a space.
sha1 for "src"; 0{40} if creation, or if "src" is from the index and is unmerged.
a space.
sha1 for "dst"; 0{40} if deletion, if "dst" is from the index and is unmerged, or if "dst" is from the work tree and is out of sync with the index.
a space.
status, followed by optional "score" number.
a tab or a NUL when -z option is used.
path for "src"
a tab or a NUL when -z option is used; only exists for C or R.
path for "dst"; only exists for C or R.
an LF or a NUL when -z option is used, to terminate the record.

Possible status letters are:

A: addition of a file ✨
C: copy of a file into a new one
D: deletion of a file 💀
M: modification of the contents or mode of a file
R: renaming of a file
T: change in the type of the file (regular file, symbolic link or submodule)
U: file is unmerged (you must complete the merge before it can be committed — no skipping, bestie 🚨)
X: "unknown" change type (most probably a bug, please report it 🙈)

Status letters C and R are always followed by a score (that's the percentage of similarity between the source and target of the move or copy). Status letter M may be followed by a score (denoting the percentage of dissimilarity) for file rewrites 👑

The sha1 for "dst" is shown as all 0's if a file on the filesystem is out of sync with the index.

Example:

:100644 100644 5be4a4a 0000000 M file.c

Without the -z option, pathnames with "unusual" characters get quoted as explained for the configuration variable core.quotePath (see git-config(1)). With -z the filename is output verbatim and the line is terminated by a NUL byte 🌸

💥 DIFF FORMAT FOR MERGES 💥

git-diff-tree, git-diff-files and git-diff --raw can take the -c or --cc option to generate diff output for merge commits too — iconic behaviour, honestly 👑 The output differs from the format above in these ways:

there is a colon for each parent
there are more "src" modes and "src" sha1
status is concatenated status characters for each parent
no optional "score" number
tab-separated pathname(s) of the file

For -c and --cc, only the destination or final path is shown even if the file was renamed on any side of history. With --combined-all-paths, the name of the path in each parent is shown followed by the name of the path in the merge commit 🦋

Examples for -c and --cc without --combined-all-paths:

::100644 100644 100644 fabadb8 cc95eb0 4866510 MM       desc.c
::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM       bar.sh
::100644 100644 100644 e07d6c5 9042e82 ee91881 RR       phooey.c

Examples when --combined-all-paths is added to either -c or --cc:

::100644 100644 100644 fabadb8 cc95eb0 4866510 MM       desc.c  desc.c  desc.c
::100755 100755 100755 52b7a2d 6d1ac04 d2ac7d7 RM       foo.sh  bar.sh  bar.sh
::100644 100644 100644 e07d6c5 9042e82 ee91881 RR       fooey.c fuey.c  phooey.c

Heads up: combined diff lists only files which were modified from all parents. Not some of them. All of them. Periodt 💅

✨ GENERATING PATCH TEXT WITH `-P` ✨

Okay so bestie, when you run git-diff(1), git-log(1), git-show(1), git-diff-index(1), git-diff-tree(1), or git-diff-files(1) with the -p option, it produces patch text. Iconic. You can customize how that patch text gets created via the GIT_EXTERNAL_DIFF and GIT_DIFF_OPTS environment variables (peep git(1) for those), and also the diff attribute (see gitattributes(5)).

What -p produces is slightly different from the traditional diff format, and here’s the tea:

It’s preceded by a "git diff" header that looks like this:

diff --git a/file1 b/file2

The a/ and b/ filenames are the same unless rename/copy is involved. Especially, even for a creation or a deletion, /dev/null is not used in place of the a/ or b/ filenames. When a rename/copy is involved, file1 and file2 show the name of the source file of the rename/copy and the name of the file that the rename/copy produces, respectively. Slay.

It’s followed by one or more extended header lines:

old mode <mode>
new mode <mode>
deleted file mode <mode>
new file mode <mode>
copy from <path>
copy to <path>
rename from <path>
rename to <path>
similarity index <number>
dissimilarity index <number>
index <hash>..<hash> <mode>

File modes <mode> are printed as 6-digit octal numbers including the file type and file permission bits. Path names in extended headers do not include the a/ and b/ prefixes. The similarity index is the percentage of unchanged lines, and the dissimilarity index is the percentage of changed lines. It’s a rounded down integer, followed by a percent sign. The similarity index value of 100% is thus reserved for two equal files, while 100% dissimilarity means that no line from the old file made it into the new one. The index line includes the blob object names before and after the change. The <mode> is included if the file mode does not change; otherwise, separate lines indicate the old and the new mode. Periodt.

Pathnames with "unusual" characters are quoted as explained for the configuration variable core.quotePath (see git-config(1)).
All the file1 files in the output refer to files before the commit, and all the file2 files refer to files after the commit. It is incorrect to apply each change to each file sequentially. For example, this patch will swap a and b:

diff --git a/a b/b
rename from a
rename to b
diff --git a/b b/a
rename from b
rename to a

Bestie, do not go applying those one by one. That’s not the vibe. 🙈

Hunk headers mention the name of the function to which the hunk applies. See "Defining a custom hunk-header" in gitattributes(5) for details of how to tailor this to specific languages.

💅 COMBINED DIFF FORMAT 💅

Any diff-generating command can take the -c or --cc option to produce a combined diff when showing a merge. This is the default format when showing merges with git-diff(1) or git-show(1). Also note you can give a suitable --diff-merges option to any of these commands to force generation of diffs in a specific format. The way that works is lowkey iconic.

A "combined diff" format looks like this:

diff --combined describe.c
index fabadb8,cc95eb0..4866510
--- a/describe.c
+++ b/describe.c
@@@ -98,20 -98,12 +98,20 @@@
        return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;
  }

- static void describe(char *arg)
 -static void describe(struct commit *cmit, int last_one)
++static void describe(char *arg, int last_one)
  {
 +      unsigned char sha1[20];
 +      struct commit *cmit;
        struct commit_list *list;
        static int initialized = 0;
        struct commit_name *n;

 +      if (get_sha1(arg, sha1) < 0)
 +              usage(describe_usage);
 +      cmit = lookup_commit_reference(sha1);
 +      if (!cmit)
 +              usage(describe_usage);
 +
        if (!initialized) {
                initialized = 1;
                for_each_ref(get_name);

It’s preceded by a "git diff" header that looks like this (when the -c option is used):

diff --combined file

or like this (when the --cc option is used):

diff --cc file

It’s followed by one or more extended header lines (this example shows a merge with two parents):

index <hash>,<hash>..<hash>
mode <mode>,<mode>..<mode>
new file mode <mode>
deleted file mode <mode>,<mode>

The mode <mode>,<mode>..<mode> line appears only if at least one of the <mode> is different from the rest. Extended headers with information about detected content movement (renames and copying detection) are designed to work with the diff of two <tree-ish> and are not used by combined diff format.

It’s followed by a two-line from-file/to-file header:

--- a/file
+++ b/file

Similar to the two-line header for the traditional unified diff format, /dev/null is used to signal created or deleted files. However, if the --combined-all-paths option is provided, instead of a two-line from-file/to-file, you get an N+1 line from-file/to-file header, where N is the number of parents in the merge commit:

--- a/file
--- a/file
--- a/file
+++ b/file

This extended format can be useful if rename or copy detection is active, to allow you to see the original name of the file in different parents. Obsessed.

Chunk header format is modified to prevent people from accidentally feeding it to patch -p1. Combined diff format was created for review of merge commit changes, and was not meant to be applied. The change is similar to the change in the extended index header:

@@@ <from-file-range> <from-file-range> <to-file-range> @@@

There are (number of parents + 1) @ characters in the chunk header for combined diff format.

Okay so this next part is the whole vibe of combined diff, and it’s a little different from the traditional unified diff format. The traditional one shows two files A and B with a single column that has - (minus — appears in A but removed in B), + (plus — missing in A but added to B), or " " (space — unchanged) prefix. This format compares two or more files file1, file2,... with one file X, and shows how X differs from each of fileN. One column for each of fileN is prepended to the output line to note how X’s line is different from it.

A - character in the column N means that the line appears in fileN but it does not appear in the result. A + character in the column N means that the line appears in the result, and fileN does not have that line (in other words, the line was added, from the point of view of that parent).

In the above example output, the function signature was changed from both files (hence two - removals from both file1 and file2, plus ++ to mean one line that was added does not appear in either file1 or file2). Also, eight other lines are the same from file1 but do not appear in file2 (hence prefixed with +). The way that works is kind of a lot, but we move.

When shown by git diff-tree -c, it compares the parents of a merge commit with the merge result (i.e. file1..fileN are the parents). When shown by git diff-files -c, it compares the two unresolved merge parents with the working tree file (i.e. file1 is stage 2 aka "our version", file2 is stage 3 aka "their version"). Iykyk.

🌸 OTHER DIFF FORMATS 🌸

The --summary option describes newly added, deleted, renamed and copied files. The --stat option adds diffstat(1) graph to the output. These options can be combined with other options, such as -p, and are meant for human consumption. Cute, readable, slay.

When showing a change that involves a rename or a copy, --stat output formats the pathnames compactly by combining common prefix and suffix of the pathnames. For example, a change that moves arch/i386/Makefile to arch/x86/Makefile while modifying 4 lines will be shown like this:

arch/{i386 => x86}/Makefile    |   4 +--

The --numstat option gives the diffstat(1) information but is designed for easier machine consumption. An entry in --numstat output looks like this:

1       2       README
3       1       arch/{i386 => x86}/Makefile

That is, from left to right:

the number of added lines;
a tab;
the number of deleted lines;
a tab;
pathname (possibly with rename/copy information);
a newline.

When -z output option is in effect, the output is formatted this way:

1       2       README NUL
3       1       NUL arch/i386/Makefile NUL arch/x86/Makefile NUL

That is:

the number of added lines;
a tab;
the number of deleted lines;
a tab;
a NUL (only exists if renamed/copied);
pathname in preimage;
a NUL (only exists if renamed/copied);
pathname in postimage (only exists if renamed/copied);
a NUL.

The extra NUL before the preimage path in renamed case is to allow scripts that read the output to tell if the current record being read is a single-path record or a rename/copy record without reading ahead. After reading added and deleted lines, reading up to NUL would yield the pathname, but if that is NUL, the record will show two paths. Fr fr, that’s the whole point.

🌈 GIT 🌈

Part of the git(1) suite. Obviously.

And that’s the whole page, bestie! This was GIT-DIFF-TREE(1) from Git 2.56.0, dated 2026-09-27. Go off and diff responsibly. 💖✨🦋

📦 where this page comes from

arch rolling: git 2.56.0-1 · GPL-2.0-only
💖 girlified by deepseek-ai/DeepSeek-V4.1-Flash-1017 on 2026-11-06: a language model's rewrite of the original, under its terms

GIT-DIFF-TREE(1)

🌸 NAME 🌸

🎀 SYNOPSIS 🎀

✨ DESCRIPTION ✨

💅 OPTIONS 💅

✨ PRETTY FORMATS ✨

oneline

short

medium

full

fuller

reference

email

mboxrd

raw

format:<format-string>