2 .\" Title: gitformat-index
3 .\" Author: [FIXME: author] [see http://www.docbook.org/tdg5/en/html/author]
4 .\" Generator: DocBook XSL Stylesheets vsnapshot <http://docbook.sf.net/>
7 .\" Source: Git 2.37.3.542.gdd3f6c4cae
10 .TH "GITFORMAT\-INDEX" "5" "09/09/2022" "Git 2\&.37\&.3\&.542\&.gdd3f6c" "Git Manual"
11 .\" -----------------------------------------------------------------
12 .\" * Define some portability stuff
13 .\" -----------------------------------------------------------------
14 .\" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
15 .\" http://bugs.debian.org/507673
16 .\" http://lists.gnu.org/archive/html/groff/2009-02/msg00013.html
17 .\" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
20 .\" -----------------------------------------------------------------
21 .\" * set default formatting
22 .\" -----------------------------------------------------------------
23 .\" disable hyphenation
25 .\" disable justification (adjust text to left margin only)
27 .\" -----------------------------------------------------------------
28 .\" * MAIN CONTENT STARTS HERE *
29 .\" -----------------------------------------------------------------
31 gitformat-index \- Git index format
41 .SH "THE GIT INDEX FILE HAS THE FOLLOWING FORMAT"
47 All binary numbers are in network byte order\&.
48 In a repository using the traditional SHA\-1, checksums and object IDs
49 (object names) mentioned below are all computed using SHA\-1\&. Similarly,
50 in SHA\-256 repositories, these values are computed using SHA\-256\&.
51 Version 2 is described here unless stated otherwise\&.
65 A 12\-byte header consisting of
72 The signature is { \(aqD\(aq, \(aqI\(aq, \(aqR\(aq, \(aqC\(aq } (stands for "dircache")
82 4\-byte version number:
83 The current supported versions are 2, 3 and 4\&.
93 32\-bit number of index entries\&.
108 A number of sorted index entries (see below)\&.
125 Extensions are identified by signature\&. Optional extensions can
126 be ignored if Git does not understand them\&.
136 4\-byte extension signature\&. If the first byte is \(aqA\(aq\&.\&.\(aqZ\(aq the
137 extension is optional and can be ignored\&.
147 32\-bit size of the extension
172 Hash checksum over the content of the index file before this checksum\&.
180 Index entries are sorted in ascending order on the name field,
181 interpreted as a string of unsigned bytes (i\&.e\&. memcmp() order, no
182 localization, no special casing of directory separator \(aq/\(aq)\&. Entries
183 with the same name are sorted by their stage field\&.
193 An index entry typically represents a file\&. However, if sparse\-checkout
194 is enabled in cone mode (`core\&.sparseCheckoutCone` is enabled) and the
195 `extensions\&.sparseIndex` extension is enabled, then the index may
196 contain entries for directories outside of the sparse\-checkout definition\&.
197 These entries have mode `040000`, include the `SKIP_WORKTREE` bit, and
198 the path ends in a directory separator\&.
208 32\-bit ctime seconds, the last time a file\(aqs metadata changed
219 32\-bit ctime nanosecond fractions
230 32\-bit mtime seconds, the last time a file\(aqs data changed
241 32\-bit mtime nanosecond fractions
274 32\-bit mode, split into (high to low bits)
285 valid values in binary are 1000 (regular file), 1010 (symbolic link)
306 9\-bit unix permission\&. Only 0755 and 0644 are valid for regular files\&.
307 Symbolic links and gitlinks have value 0 in this field\&.
340 This is the on\-disk size from stat(2), truncated to 32\-bit\&.
350 Object name for the represented object
360 A 16\-bit \(aqflags\(aq field split into (high to low bits)
370 1\-bit assume\-valid flag
380 1\-bit extended flag (must be zero in version 2)
390 2\-bit stage (during merge)
400 12\-bit name length if the length is less than 0xFFF; otherwise 0xFFF
401 is stored in this field\&.
411 (Version 3 or later) A 16\-bit field, only applicable if the
412 "extended flag" above is 1, split into (high to low bits)\&.
422 1\-bit reserved for future
432 1\-bit skip\-worktree flag (used by sparse checkout)
442 1\-bit intent\-to\-add flag (used by "git add \-N")
452 13\-bit unused, must be zero
462 Entry path name (variable length) relative to top level directory
463 (without leading slash)\&. \(aq/\(aq is used as path separator\&. The special
464 path components "\&.", "\&.\&." and "\&.git" (without quotes) are disallowed\&.
465 Trailing slash is also disallowed\&.
475 The exact encoding is undefined, but the \(aq\&.\(aq and \(aq/\(aq characters
476 are encoded in 7\-bit ASCII and the encoding cannot contain a NUL
477 byte (iow, this is a UNIX pathname)\&.
487 (Version 4) In version 4, the entry path name is prefix\-compressed
488 relative to the path name for the previous entry (the very first
489 entry is encoded as if the path name for the previous entry is an
490 empty string)\&. At the beginning of an entry, an integer N in the
491 variable width encoding (the same encoding as the offset is encoded
492 for OFS_DELTA pack entries; see linkgit:gitformat\-pack[5]) is stored, followed
493 by a NUL\-terminated string S\&. Removing N bytes from the end of the
494 path name for the previous entry, and replacing it with the string S
495 yields the path name for this entry\&.
505 1\-8 nul bytes as necessary to pad the entry to a multiple of eight bytes
506 while keeping the name NUL\-terminated\&.
516 (Version 4) In version 4, the padding after the pathname does not
527 Interpretation of index entries in split index mode is completely
528 different\&. See below for details\&.
540 Since the index does not record entries for directories, the cache
541 entries cannot describe tree objects that already exist in the object
542 database for regions of the index that are unchanged from an existing
543 commit\&. The cache tree extension stores a recursive tree structure that
544 describes the trees that already exist and completely match sections of
545 the cache entries\&. This speeds up tree object generation from the index
546 for a new commit by only computing the trees that are "new" to that
547 commit\&. It also assists when comparing the index to another tree, such
548 as `HEAD^{tree}`, since sections of the index can be skipped when a tree
549 comparison demonstrates equality\&.
559 The recursive tree structure uses nodes that store a number of cache
560 entries, a list of subnodes, and an object ID (OID)\&. The OID references
561 the existing tree for that node, if it is known to exist\&. The subnodes
562 correspond to subdirectories that themselves have cache tree nodes\&. The
563 number of cache entries corresponds to the number of cache entries in
564 the index that describe paths within that tree\(aqs directory\&.
574 The extension tracks the full directory structure in the cache tree
575 extension, but this is generally smaller than the full cache entry list\&.
585 When a path is updated in index, Git invalidates all nodes of the
586 recursive cache tree corresponding to the parent directories of that
587 path\&. We store these tree nodes as being "invalid" by using "\-1" as the
588 number of cache entries\&. Invalid nodes still store a span of index
589 entries, allowing Git to focus its efforts when reconstructing a full
600 The signature for this extension is { \(aqT\(aq, \(aqR\(aq, \(aqE\(aq, \(aqE\(aq }\&.
610 A series of entries fill the entire extension; each of which
625 NUL\-terminated path component (relative to its parent directory);
636 ASCII decimal number of entries in the index that is covered by the tree this entry represents (entry_count);
658 ASCII decimal number that represents the number of subtrees this tree has;
669 A newline (ASCII 10); and
680 Object name for the object that would result from writing this span of index as a tree\&.
686 An entry can be in an invalidated state and is represented by having
687 a negative number in the entry_count field\&. In this case, there is no
688 object name and the next entry starts immediately after the newline\&.
689 When writing an invalid entry, \-1 should always be used as entry_count\&.
699 The entries are written out in the top\-down, depth\-first order\&. The
700 first entry represents the root level of the repository, followed by the
701 first subtree\-\-let\(aqs call this A\-\-of the root level (with its name
702 relative to the root level), followed by the first subtree of A (with
703 its name relative to A), and so on\&. The specified number of subtrees
704 indicates when the current level of the recursive stack is complete\&.
716 A conflict is represented in the index as a set of higher stage entries\&.
717 When a conflict is resolved (e\&.g\&. with "git add path"), these higher
718 stage entries will be removed and a stage\-0 entry with proper resolution
729 When these higher stage entries are removed, they are saved in the
730 resolve undo extension, so that conflicts can be recreated (e\&.g\&. with
731 "git checkout \-m"), in case users want to redo a conflict resolution
742 The signature for this extension is { \(aqR\(aq, \(aqE\(aq, \(aqU\(aq, \(aqC\(aq }\&.
752 A series of entries fill the entire extension; each of which
767 NUL\-terminated pathname the entry describes (relative to the root of the repository, i\&.e\&. full pathname);
778 Three NUL\-terminated ASCII octal numbers, entry mode of entries in stage 1 to 3 (a missing stage is represented by "0" in this field); and
789 At most three object names of the entry in stages from 1 to 3 (nothing is written for a missing stage)\&.
797 In split index mode, the majority of index entries could be stored
798 in a separate file\&. This extension records the changes to be made on
799 top of that to produce the final index\&.
809 The signature for this extension is { \(aql\(aq, \(aqi\(aq, \(aqn\(aq, \(aqk\(aq }\&.
819 The extension consists of:
833 Hash of the shared index file\&. The shared index file path is $GIT_DIR/sharedindex\&.<hash>\&. If all bits are zero, the index does not require a shared index file\&.
844 An ewah\-encoded delete bitmap, each bit represents an entry in the shared index\&. If a bit is set, its corresponding entry in the shared index will be removed from the final index\&. Note, because a delete operation changes index entry positions, but we do need original positions in replace phase, it\(cqs best to just mark entries for removal, then do a mass deletion after replacement\&.
855 An ewah\-encoded replace bitmap, each bit represents an entry in the shared index\&. If a bit is set, its corresponding entry in the shared index will be replaced with an entry in this index file\&. All replaced entries are stored in sorted order in this index\&. The first "1" bit in the replace bitmap corresponds to the first index entry, the second "1" bit to the second entry and so on\&. Replaced entries may have empty path names to save space\&.
861 The remaining index entries after replaced ones will be added to the
862 final index\&. These added entries are also sorted by entry name then
869 .SH "UNTRACKED CACHE"
875 Untracked cache saves the untracked file list and necessary data to
876 verify the cache\&. The signature for this extension is { \(aqU\(aq, \(aqN\(aq,
877 \(aqT\(aq, \(aqR\(aq }\&.
887 The extension starts with
901 A sequence of NUL\-terminated strings, preceded by the size of the sequence in variable width encoding\&. Each string describes the environment where the cache can be used\&.
912 Stat data of $GIT_DIR/info/exclude\&. See "Index entry" section from ctime field until "file size"\&.
923 Stat data of core\&.excludesFile
934 32\-bit dir_flags (see struct dir_struct)
945 Hash of $GIT_DIR/info/exclude\&. A null hash means the file does not exist\&.
956 Hash of core\&.excludesFile\&. A null hash means the file does not exist\&.
967 NUL\-terminated string of per\-dir exclude file name\&. This usually is "\&.gitignore"\&.
978 The number of following directory blocks, variable width encoding\&. If this number is zero, the extension ends here with a following NUL\&.
989 A number of directory blocks in depth\-first\-search order, each consists of
1000 The number of untracked entries, variable width encoding\&.
1005 \h'-04'\(bu\h'+03'\c
1011 The number of sub\-directory blocks, variable width encoding\&.
1016 \h'-04'\(bu\h'+03'\c
1022 The directory name terminated by NUL\&.
1027 \h'-04'\(bu\h'+03'\c
1033 A number of untracked file/dir names terminated by NUL\&.
1036 The remaining data of each directory block is grouped by type:
1040 \h'-04'\(bu\h'+03'\c
1046 An ewah bitmap, the n\-th bit marks whether the n\-th directory has valid untracked cache entries\&.
1051 \h'-04'\(bu\h'+03'\c
1057 An ewah bitmap, the n\-th bit records "check\-only" bit of read_directory_recursive() for the n\-th directory\&.
1062 \h'-04'\(bu\h'+03'\c
1068 An ewah bitmap, the n\-th bit indicates whether hash and stat data is valid for the n\-th directory and exists in the next data\&.
1073 \h'-04'\(bu\h'+03'\c
1079 An array of stat data\&. The n\-th data corresponds with the n\-th "one" bit in the previous ewah bitmap\&.
1084 \h'-04'\(bu\h'+03'\c
1090 An array of hashes\&. The n\-th hash corresponds with the n\-th "one" bit in the previous ewah bitmap\&.
1095 \h'-04'\(bu\h'+03'\c
1103 .SH "FILE SYSTEM MONITOR CACHE"
1109 The file system monitor cache tracks files for which the core\&.fsmonitor
1110 hook has told us about changes\&. The signature for this extension is
1111 { \(aqF\(aq, \(aqS\(aq, \(aqM\(aq, \(aqN\(aq }\&.
1121 The extension starts with
1129 \h'-04'\(bu\h'+03'\c
1135 32\-bit version number: the current supported versions are 1 and 2\&.
1140 \h'-04'\(bu\h'+03'\c
1146 (Version 1) 64\-bit time: the extension data reflects all changes through the given time which is stored as the nanoseconds elapsed since midnight, January 1, 1970\&.
1151 \h'-04'\(bu\h'+03'\c
1157 (Version 2) A null terminated string: an opaque token defined by the file system monitor application\&. The extension data reflects all changes relative to that token\&.
1162 \h'-04'\(bu\h'+03'\c
1168 32\-bit bitmap size: the size of the CE_FSMONITOR_VALID bitmap\&.
1173 \h'-04'\(bu\h'+03'\c
1179 An ewah bitmap, the n\-th bit indicates whether the n\-th index entry is not CE_FSMONITOR_VALID\&.
1181 .SH "END OF INDEX ENTRY"
1187 The End of Index Entry (EOIE) is used to locate the end of the variable
1188 length index entries and the beginning of the extensions\&. Code can take
1189 advantage of this to quickly locate the index extensions without having
1190 to parse through all of the index entries\&.
1200 Because it must be able to be loaded before the variable length cache
1201 entries and other index extensions, this extension must be written last\&.
1202 The signature for this extension is { \(aqE\(aq, \(aqO\(aq, \(aqI\(aq, \(aqE\(aq }\&.
1212 The extension consists of:
1220 \h'-04'\(bu\h'+03'\c
1226 32\-bit offset to the end of the index entries
1231 \h'-04'\(bu\h'+03'\c
1237 Hash over the extension types and their sizes (but not their contents)\&. E\&.g\&. if we have "TREE" extension that is N\-bytes long, "REUC" extension that is M\-bytes long, followed by "EOIE", then the hash would be:
1243 Hash("TREE" + <binary representation of N> +
1244 "REUC" + <binary representation of M>)
1250 .SH "INDEX ENTRY OFFSET TABLE"
1256 The Index Entry Offset Table (IEOT) is used to help address the CPU
1257 cost of loading the index by enabling multi\-threading the process of
1258 converting cache entries from the on\-disk format to the in\-memory format\&.
1259 The signature for this extension is { \(aqI\(aq, \(aqE\(aq, \(aqO\(aq, \(aqT\(aq }\&.
1269 The extension consists of:
1277 \h'-04'\(bu\h'+03'\c
1283 32\-bit version (currently 1)
1288 \h'-04'\(bu\h'+03'\c
1294 A number of index offset entries each consisting of:
1299 \h'-04'\(bu\h'+03'\c
1305 32\-bit offset from the beginning of the file to the first cache entry in this block of entries\&.
1310 \h'-04'\(bu\h'+03'\c
1316 32\-bit count of cache entries in this block
1318 .SH "SPARSE DIRECTORY ENTRIES"
1324 When using sparse\-checkout in cone mode, some entire directories within
1325 the index can be summarized by pointing to a tree object instead of the
1326 entire expanded list of paths within that tree\&. An index containing such
1327 entries is a "sparse index"\&. Index format versions 4 and less were not
1328 implemented with such entries in mind\&. Thus, for these versions, an
1329 index containing sparse directory entries will include this extension
1330 with signature { \(aqs\(aq, \(aqd\(aq, \(aqi\(aq, \(aqr\(aq }\&. Like the split\-index extension,
1331 tools should avoid interacting with a sparse index unless they understand
1339 Part of the \fBgit\fR(1) suite