Documentation/revisions.txt

   1 SPECIFYING REVISIONS
   2 --------------------
   3
   4 A revision parameter '<rev>' typically, but not necessarily, names a
   5 commit object.  It uses what is called an 'extended SHA-1'
   6 syntax.  Here are various ways to spell object names.  The
   7 ones listed near the end of this list name trees and
   8 blobs contained in a commit.
   9
  10 '<sha1>', e.g. 'dae86e1950b1277e545cee180551750029cfe735', 'dae86e'::
  11   The full SHA-1 object name (40-byte hexadecimal string), or
  12   a leading substring that is unique within the repository.
  13   E.g. dae86e1950b1277e545cee180551750029cfe735 and dae86e both
  14   name the same commit object if there is no other object in
  15   your repository whose object name starts with dae86e.
  16
  17 '<describeOutput>', e.g. 'v1.7.4.2-679-g3bee7fb'::
  18   Output from `git describe`; i.e. a closest tag, optionally
  19   followed by a dash and a number of commits, followed by a dash, a
  20   'g', and an abbreviated object name.
  21
  22 '<refname>', e.g. 'master', 'heads/master', 'refs/heads/master'::
  23   A symbolic ref name.  E.g. 'master' typically means the commit
  24   object referenced by 'refs/heads/master'.  If you
  25   happen to have both 'heads/master' and 'tags/master', you can
  26   explicitly say 'heads/master' to tell Git which one you mean.
  27   When ambiguous, a '<refname>' is disambiguated by taking the
  28   first match in the following rules:
  29
  30   . If '$GIT_DIR/<refname>' exists, that is what you mean (this is usually
  31     useful only for 'HEAD', 'FETCH_HEAD', 'ORIG_HEAD', 'MERGE_HEAD'
  32     and 'CHERRY_PICK_HEAD');
  33
  34   . otherwise, 'refs/<refname>' if it exists;
  35
  36   . otherwise, 'refs/tags/<refname>' if it exists;
  37
  38   . otherwise, 'refs/heads/<refname>' if it exists;
  39
  40   . otherwise, 'refs/remotes/<refname>' if it exists;
  41
  42   . otherwise, 'refs/remotes/<refname>/HEAD' if it exists.
  43 +
  44 'HEAD' names the commit on which you based the changes in the working tree.
  45 'FETCH_HEAD' records the branch which you fetched from a remote repository
  46 with your last `git fetch` invocation.
  47 'ORIG_HEAD' is created by commands that move your 'HEAD' in a drastic
  48 way, to record the position of the 'HEAD' before their operation, so that
  49 you can easily change the tip of the branch back to the state before you ran
  50 them.
  51 'MERGE_HEAD' records the commit(s) which you are merging into your branch
  52 when you run `git merge`.
  53 'CHERRY_PICK_HEAD' records the commit which you are cherry-picking
  54 when you run `git cherry-pick`.
  55 +
  56 Note that any of the 'refs/*' cases above may come either from
  57 the '$GIT_DIR/refs' directory or from the '$GIT_DIR/packed-refs' file.
  58 While the ref name encoding is unspecified, UTF-8 is preferred as
  59 some output processing may assume ref names in UTF-8.
  60
  61 '@'::
  62   '@' alone is a shortcut for 'HEAD'.
  63
  64 '<refname>@\{<date>\}', e.g. 'master@\{yesterday\}', 'HEAD@\{5 minutes ago\}'::
  65   A ref followed by the suffix '@' with a date specification
  66   enclosed in a brace
  67   pair (e.g. '\{yesterday\}', '\{1 month 2 weeks 3 days 1 hour 1
  68   second ago\}' or '\{1979-02-26 18:30:00\}') specifies the value
  69   of the ref at a prior point in time.  This suffix may only be
  70   used immediately following a ref name and the ref must have an
  71   existing log ('$GIT_DIR/logs/<ref>'). Note that this looks up the state
  72   of your *local* ref at a given time; e.g., what was in your local
  73   'master' branch last week. If you want to look at commits made during
  74   certain times, see '--since' and '--until'.
  75
  76 '<refname>@\{<n>\}', e.g. 'master@\{1\}'::
  77   A ref followed by the suffix '@' with an ordinal specification
  78   enclosed in a brace pair (e.g. '\{1\}', '\{15\}') specifies
  79   the n-th prior value of that ref.  For example 'master@\{1\}'
  80   is the immediate prior value of 'master' while 'master@\{5\}'
  81   is the 5th prior value of 'master'. This suffix may only be used
  82   immediately following a ref name and the ref must have an existing
  83   log ('$GIT_DIR/logs/<refname>').
  84
  85 '@\{<n>\}', e.g. '@\{1\}'::
  86   You can use the '@' construct with an empty ref part to get at a
  87   reflog entry of the current branch. For example, if you are on
  88   branch 'blabla' then '@\{1\}' means the same as 'blabla@\{1\}'.
  89
  90 '@\{-<n>\}', e.g. '@\{-1\}'::
  91   The construct '@\{-<n>\}' means the <n>th branch checked out
  92   before the current one.
  93
  94 '<branchname>@\{upstream\}', e.g. 'master@\{upstream\}', '@\{u\}'::
  95   The suffix '@\{upstream\}' to a branchname (short form '<branchname>@\{u\}')
  96   refers to the branch that the branch specified by branchname is set to build on
  97   top of.  A missing branchname defaults to the current one.
  98
  99 '<rev>{caret}', e.g. 'HEAD{caret}, v1.5.1{caret}0'::
 100   A suffix '{caret}' to a revision parameter means the first parent of
 101   that commit object.  '{caret}<n>' means the <n>th parent (i.e.
 102   '<rev>{caret}'
 103   is equivalent to '<rev>{caret}1').  As a special rule,
 104   '<rev>{caret}0' means the commit itself and is used when '<rev>' is the
 105   object name of a tag object that refers to a commit object.
 106
 107 '<rev>{tilde}<n>', e.g. 'master{tilde}3'::
 108   A suffix '{tilde}<n>' to a revision parameter means the commit
 109   object that is the <n>th generation ancestor of the named
 110   commit object, following only the first parents.  I.e. '<rev>{tilde}3' is
 111   equivalent to '<rev>{caret}{caret}{caret}' which is equivalent to
 112   '<rev>{caret}1{caret}1{caret}1'.  See below for an illustration of
 113   the usage of this form.
 114
 115 '<rev>{caret}\{<type>\}', e.g. 'v0.99.8{caret}\{commit\}'::
 116   A suffix '{caret}' followed by an object type name enclosed in
 117   brace pair means the object
 118   could be a tag, and dereference the tag recursively until an
 119   object of that type is found or the object cannot be
 120   dereferenced anymore (in which case, barf).  '<rev>{caret}0'
 121   is a short-hand for '<rev>{caret}\{commit\}'.
 122 +
 123 'rev{caret}\{object\}' can be used to make sure 'rev' names an
 124 object that exists, without requiring 'rev' to be a tag, and
 125 without dereferencing 'rev'; because a tag is already an object,
 126 it does not have to be dereferenced even once to get to an object.
 127
 128 '<rev>{caret}\{\}', e.g. 'v0.99.8{caret}\{\}'::
 129   A suffix '{caret}' followed by an empty brace pair
 130   means the object could be a tag,
 131   and dereference the tag recursively until a non-tag object is
 132   found.
 133
 134 '<rev>{caret}\{/<text>\}', e.g. 'HEAD^{/fix nasty bug}'::
 135   A suffix '{caret}' to a revision parameter, followed by a brace
 136   pair that contains a text led by a slash,
 137   is the same as the ':/fix nasty bug' syntax below except that
 138   it returns the youngest matching commit which is reachable from
 139   the '<rev>' before '{caret}'.
 140
 141 ':/<text>', e.g. ':/fix nasty bug'::
 142   A colon, followed by a slash, followed by a text, names
 143   a commit whose commit message matches the specified regular expression.
 144   This name returns the youngest matching commit which is
 145   reachable from any ref.  If the commit message starts with a
 146   '!' you have to repeat that;  the special sequence ':/!',
 147   followed by something else than '!', is reserved for now.
 148   The regular expression can match any part of the commit message. To
 149   match messages starting with a string, one can use e.g. ':/^foo'.
 150
 151 '<rev>:<path>', e.g. 'HEAD:README', ':README', 'master:./README'::
 152   A suffix ':' followed by a path names the blob or tree
 153   at the given path in the tree-ish object named by the part
 154   before the colon.
 155   ':path' (with an empty part before the colon)
 156   is a special case of the syntax described next: content
 157   recorded in the index at the given path.
 158   A path starting with './' or '../' is relative to the current working directory.
 159   The given path will be converted to be relative to the working tree's root directory.
 160   This is most useful to address a blob or tree from a commit or tree that has
 161   the same tree structure as the working tree.
 162
 163 ':<n>:<path>', e.g. ':0:README', ':README'::
 164   A colon, optionally followed by a stage number (0 to 3) and a
 165   colon, followed by a path, names a blob object in the
 166   index at the given path. A missing stage number (and the colon
 167   that follows it) names a stage 0 entry. During a merge, stage
 168   1 is the common ancestor, stage 2 is the target branch's version
 169   (typically the current branch), and stage 3 is the version from
 170   the branch which is being merged.
 171
 172 Here is an illustration, by Jon Loeliger.  Both commit nodes B
 173 and C are parents of commit node A.  Parent commits are ordered
 174 left-to-right.
 175
 176 ........................................
 177 G   H   I   J
 178  \ /     \ /
 179   D   E   F
 180    \  |  / \
 181     \ | /   |
 182      \|/    |
 183       B     C
 184        \   /
 185         \ /
 186          A
 187 ........................................
 188
 189     A =      = A^0
 190     B = A^   = A^1     = A~1
 191     C = A^2  = A^2
 192     D = A^^  = A^1^1   = A~2
 193     E = B^2  = A^^2
 194     F = B^3  = A^^3
 195     G = A^^^ = A^1^1^1 = A~3
 196     H = D^2  = B^^2    = A^^^2  = A~2^2
 197     I = F^   = B^3^    = A^^3^
 198     J = F^2  = B^3^2   = A^^3^2
 199
 200
 201 SPECIFYING RANGES
 202 -----------------
 203
 204 History traversing commands such as `git log` operate on a set
 205 of commits, not just a single commit.  To these commands,
 206 specifying a single revision with the notation described in the
 207 previous section means the set of commits reachable from that
 208 commit, following the commit ancestry chain.
 209
 210 To exclude commits reachable from a commit, a prefix '{caret}'
 211 notation is used.  E.g. '{caret}r1 r2' means commits reachable
 212 from 'r2' but exclude the ones reachable from 'r1'.
 213
 214 This set operation appears so often that there is a shorthand
 215 for it.  When you have two commits 'r1' and 'r2' (named according
 216 to the syntax explained in SPECIFYING REVISIONS above), you can ask
 217 for commits that are reachable from r2 excluding those that are reachable
 218 from r1 by '{caret}r1 r2' and it can be written as 'r1..r2'.
 219
 220 A similar notation 'r1\...r2' is called symmetric difference
 221 of 'r1' and 'r2' and is defined as
 222 'r1 r2 --not $(git merge-base --all r1 r2)'.
 223 It is the set of commits that are reachable from either one of
 224 'r1' or 'r2' but not from both.
 225
 226 In these two shorthands, you can omit one end and let it default to HEAD.
 227 For example, 'origin..' is a shorthand for 'origin..HEAD' and asks "What
 228 did I do since I forked from the origin branch?"  Similarly, '..origin'
 229 is a shorthand for 'HEAD..origin' and asks "What did the origin do since
 230 I forked from them?"  Note that '..' would mean 'HEAD..HEAD' which is an
 231 empty range that is both reachable and unreachable from HEAD.
 232
 233 Two other shorthands for naming a set that is formed by a commit
 234 and its parent commits exist.  The 'r1{caret}@' notation means all
 235 parents of 'r1'.  'r1{caret}!' includes commit 'r1' but excludes
 236 all of its parents.
 237
 238 To summarize:
 239
 240 '<rev>'::
 241         Include commits that are reachable from (i.e. ancestors of)
 242         <rev>.
 243
 244 '{caret}<rev>'::
 245         Exclude commits that are reachable from (i.e. ancestors of)
 246         <rev>.
 247
 248 '<rev1>..<rev2>'::
 249         Include commits that are reachable from <rev2> but exclude
 250         those that are reachable from <rev1>.  When either <rev1> or
 251         <rev2> is omitted, it defaults to 'HEAD'.
 252
 253 '<rev1>\...<rev2>'::
 254         Include commits that are reachable from either <rev1> or
 255         <rev2> but exclude those that are reachable from both.  When
 256         either <rev1> or <rev2> is omitted, it defaults to 'HEAD'.
 257
 258 '<rev>{caret}@', e.g. 'HEAD{caret}@'::
 259   A suffix '{caret}' followed by an at sign is the same as listing
 260   all parents of '<rev>' (meaning, include anything reachable from
 261   its parents, but not the commit itself).
 262
 263 '<rev>{caret}!', e.g. 'HEAD{caret}!'::
 264   A suffix '{caret}' followed by an exclamation mark is the same
 265   as giving commit '<rev>' and then all its parents prefixed with
 266   '{caret}' to exclude them (and their ancestors).
 267
 268 Here are a handful of examples:
 269
 270    D                G H D
 271    D F              G H I J D F
 272    ^G D             H D
 273    ^D B             E I J F B
 274    B..C             C
 275    B...C            G H D E B C
 276    ^D B C           E I J F B C
 277    C                I J F C
 278    C^@              I J F
 279    C^!              C
 280    F^! D            G H D F