2 Title: Collecting information about git
5 Author: Oleg Broytman <phd@phdru.name>
8 Content-Type: text/x-rst
15 This Informational PEP collects information about git. There is, of
16 course, a lot of documentation for git, so the PEP concentrates on
17 more complex issues, scenarios and topics.
19 The plan is to extend the PEP in the future collecting information
20 about equivalence of Mercurial and git scenarios to help migrating
21 Python development from Mercurial to git.
23 The author of the PEP doesn't currently plan to write a Process PEP on
24 migration from Mercurial to git.
30 Git is accompanied with a lot of documentation, both online and
33 Documentation for starters
34 --------------------------
37 <https://www.kernel.org/pub/software/scm/git/docs/gittutorial.html>`_,
39 <https://www.kernel.org/pub/software/scm/git/docs/gittutorial-2.html>`_.
42 <https://www.kernel.org/pub/software/scm/git/docs/user-manual.html>`_.
43 `Everyday GIT With 20 Commands Or So
44 <https://www.kernel.org/pub/software/scm/git/docs/everyday.html>`_.
46 <https://www.kernel.org/pub/software/scm/git/docs/gitworkflows.html>`_.
48 Advanced documentation
49 ----------------------
52 <http://www-cs-students.stanford.edu/~blynn/gitmagic/index.html>`_,
53 also with a number of translations.
55 `Pro Git <https://git-scm.com/book>`_. The Book about git. Buy it at
56 Amazon or download in PDF, mobi, or ePub form. Has translations to
57 many different languages. Download Russian translation from `GArik
58 <https://github.com/GArik/progit/wiki>`_.
60 `Git Wiki <https://git.wiki.kernel.org/index.php/Main_Page>`_.
65 Git has builtin help: run ``git help $TOPIC``. For example, run
66 ``git help git`` or ``git help help``.
72 Download and installation
73 -------------------------
75 Unix users: `download and install using your package manager
76 <https://git-scm.com/download/linux>`_.
78 Microsoft Windows: download `git-for-windows
79 <https://github.com/git-for-windows/git/releases>`_ or `msysGit
80 <https://github.com/msysgit/msysgit/releases>`_.
82 MacOS X: use git installed with `XCode
83 <https://developer.apple.com/xcode/downloads/>`_ or download from
84 `MacPorts <https://www.macports.org/ports.php?by=name&substr=git>`_ or
86 <http://sourceforge.net/projects/git-osx-installer/files/>`_ or
87 install git with `Homebrew <http://brew.sh/>`_: ``brew install git``.
89 `git-cola <https://git-cola.github.io/index.html>`_ is a Git GUI
90 written in Python and GPL licensed. Linux, Windows, MacOS X.
92 `TortoiseGit <https://tortoisegit.org/>`_ is a Windows Shell Interface
93 to Git based on TortoiseSVN; open source.
98 This simple code is often appears in documentation, but it is
99 important so let repeat it here. Git stores author and committer
100 names/emails in every commit, so configure your real name and
103 $ git config --global user.name "User Name"
104 $ git config --global user.email user.name@example.org
110 Examples of git commands in this PEP use the following approach. It is
111 supposed that you, the user, works with a local repository named
112 ``python`` that has an upstream remote repo named ``origin``. Your
113 local repo has two branches ``v1`` and ``master``. For most examples
114 the currently checked out branch is ``master``. That is, it's assumed
115 you have done something like that::
117 $ git clone http://git.python.org/python.git
119 $ git branch v1 origin/v1
121 The first command clones remote repository into local directory
122 `python``, creates a new local branch master, sets
123 remotes/origin/master as its upstream remote-tracking branch and
124 checks it out into the working directory.
126 The last command creates a new local branch v1 and sets
127 remotes/origin/v1 as its upstream remote-tracking branch.
129 The same result can be achieved with commands::
131 $ git clone -b v1 http://git.python.org/python.git
133 $ git checkout --track origin/master
135 The last command creates a new local branch master, sets
136 remotes/origin/master as its upstream remote-tracking branch and
137 checks it out into the working directory.
140 Branches and branches
141 =====================
143 Git terminology can be a bit misleading. Take, for example, the term
144 "branch". In git it has two meanings. A branch is a directed line of
145 commits (possibly with merges). And a branch is a label or a pointer
146 assigned to a line of commits. It is important to distinguish when you
147 talk about commits and when about their labels. Lines of commits are
148 by itself unnamed and are usually only lengthening and merging.
149 Labels, on the other hand, can be created, moved, renamed and deleted
153 Remote repositories and remote branches
154 =======================================
156 Remote-tracking branches are branches (pointers to commits) in your
157 local repository. They are there for you to remember what branches and
158 commits have been pulled from and pushed to what remote repos (you can
159 pull from and push to many remotes). Remote-tracking branches live
160 under ``remotes/$REMOTE`` namespaces, e.g. ``remotes/origin/master``.
162 To see the status of remote-tracking branches run::
166 To see local and remote-tracking branches (and tags) pointing to
171 You never do your own development on remote-tracking branches. You
172 create a local branch that has a remote branch as upstream and do
173 development on that local branch. On push git pushes commits to the
174 remote repo and updates remote-tracking branches, on pull git fetches
175 commits from the remote repo, updates remote-tracking branches and
176 fast-forwards, merges or rebases local branches.
178 When you do an initial clone like this::
180 $ git clone -b v1 http://git.python.org/python.git
182 git clones remote repository ``http://git.python.org/python.git`` to
183 directory ``python``, creates remote-tracking branches, creates a
184 local branch ``v1``, configure it to track upstream remotes/origin/v1
185 branch and checks out ``v1`` into the working directory.
187 Updating local and remote-tracking branches
188 -------------------------------------------
190 There is a major difference between
194 $ git fetch $REMOTE $BRANCH
200 $ git fetch $REMOTE $BRANCH:$BRANCH
202 The first command fetches commits from the named $BRANCH in the
203 $REMOTE repository that are not in your repository, updates
204 remote-tracking branch and leaves the id (the hash) of the head commit
205 in file .git/FETCH_HEAD.
207 The second command fetches commits from the named $BRANCH in the
208 $REMOTE repository that are not in your repository and updates both
209 the local branch $BRANCH and its upstream remote-tracking branch. But
210 it refuses to update branches in case of non-fast-forward. And it
211 refuses to update the current branch.
213 The first command is used internally by ``git pull``.
217 $ git pull $REMOTE $BRANCH
223 $ git fetch $REMOTE $BRANCH
224 $ git merge FETCH_HEAD
226 Certainly, $BRANCH in that case should be your current branch. If you
227 want to merge a different branch into your current branch first update
228 that non-current branch and then merge::
230 $ git fetch origin v1:v1 # Update v1
231 $ git pull --rebase origin master # Update the current branch master
232 # using rebase instead of merge
235 If you have not yet pushed commits on ``v1``, though, the scenario has
236 to become a bit more complex. Git refuses to update
237 non-fast-forwardable branch, and you don't want to do force-pull
238 because that would remove your non-pushed commits and you would need
239 to recover. So you want to rebase ``v1`` but you cannot rebase
240 non-current branch. Hence, checkout ``v1`` and rebase it before
244 $ git pull --rebase origin v1
245 $ git checkout master
246 $ git pull --rebase origin master
249 It is possible to configure git to make it fetch/pull a few branches
250 or all branches at once, so you can simply run
262 Default remote repository for fetching/pulling is origin. Default set
263 of references to fetch is calculated using matching algorithm: git
264 fetches all branches having the same name on both ends.
269 Pushing is a bit simpler. There is only one command ``push``. When you
274 $ git push origin v1 master
276 git pushes local v1 to remote v1 and local master to remote master.
279 $ git push origin v1:v1 master:master
281 Git pushes commits to the remote repo and updates remote-tracking
282 branches. Git refuses to push commits that aren't fast-forwardable.
283 You can force-push anyway, but please remember - you can force-push to
284 your own repositories but don't force-push to public or shared repos.
285 If you find git refuses to push commits that aren't fast-forwardable,
286 better fetch and merge commits from the remote repo (or rebase your
287 commits on top of the fetched commits), then push. Only force-push if
288 you know what you do and why you do it. See the section `Commit
289 editing and caveats`_ below.
291 It is possible to configure git to make it push a few branches or all
292 branches at once, so you can simply run
304 Default remote repository for pushing is origin. Default set
305 of references to push in git before 2.0 is calculated using matching
306 algorithm: git pushes all branches having the same name on both ends.
307 Default set of references to push in git 2.0+ is calculated using
308 simple algorithm: git pushes the current branch back to its
311 To configure git before 2.0 to the new behaviour run::
313 $ git config push.default simple
315 To configure git 2.0+ to the old behaviour run::
317 $ git config push.default matching
319 Git refuses to push a branch if it's the current branch in the remote
320 non-bare repository: git refuses to update remote working directory.
321 You really should push only to bare repositories. For non-bare
322 repositories git prefers pull-based workflow.
324 When you want to deploy code on a remote host and can only use push
325 (because your workstation is behind a firewall and you cannot pull
326 from it) you do that in two steps using two repositories: you push
327 from the workstation to a bare repo on the remote host, ssh to the
328 remote host and pull from the bare repo to a non-bare deployment repo.
330 That changed in git 2.3, but see `the blog post
331 <https://github.com/blog/1957-git-2-3-has-been-released#push-to-deploy>`_
332 for caveats; in 2.4 the push-to-deploy feature was `further improved
333 <https://github.com/blog/1994-git-2-4-atomic-pushes-push-to-deploy-and-more#push-to-deploy-improvements>`_.
338 Git automatically fetches tags that point to commits being fetched
339 during fetch/pull. To fetch all tags (and commits they point to) run
340 ``git fetch --tags origin``. To fetch some specific tags fetch them
343 $ git fetch origin tag $TAG1 tag $TAG2...
347 $ git fetch origin tag 1.4.2
348 $ git fetch origin v1:v1 tag 2.1.7
350 Git doesn't automatically pushes tags. That allows you to have private
351 tags. To push tags list them explicitly::
353 $ git push origin tag 1.4.2
354 $ git push origin v1 master tag 2.1.7
356 Don't move tags with ``git tag -f`` or remove tags with ``git tag -d``
357 after they have been published.
362 When cloning/fetching/pulling/pushing git copies only database objects
363 (commits, trees, files and tags) and symbolic references (branches and
364 lightweight tags). Everything else is private to the repository and
365 never cloned, updated or pushed. It's your config, your hooks, your
366 private exclude file.
368 If you want to distribute hooks, copy them to the working tree, add,
369 commit, push and instruct the team to update ind install the hook
373 Commit editing and caveats
374 ==========================
376 A warning not to edit published (pushed) commits also appears in
377 documentation but it's repeated here anyway as it's very important.
379 It is possible to recover from forced push but it's PITA for the
380 entire team. Please avoid it.
382 To see what commits have not been published yet compare the head of the
383 branch with its upstream remote-tracking branch::
385 $ git log origin/master..
386 $ git log origin/v1..v1
388 For every branch that has an upstream remote-tracking branch git
389 maintains an alias @{upstream} (short version @{u}), so the commands
390 above can be given as::
395 To see the status of all branches::
399 To compare the status of local branches with a remote repo::
401 $ git remote show origin
403 Read `how to recover from upstream rebase
404 <https://git-scm.com/docs/git-rebase#_recovering_from_upstream_rebase>`_.
405 It is in ``git help rebase``.
407 On the other hand don't be too afraid about commit editing. You can
408 safely edit, remove, reorder, combine and split commits that haven't
409 been pushed yet. You can even push commits to your own (backup) repo,
410 edit them later and force-push edited commits to replace what have
411 already been pushed. Not a problem until commits are in a public
412 or shared repository.
418 Whatever you do, don't panic. Almost anything in git can be undone.
420 git checkout: restore file's content
421 ------------------------------------
423 ``git checkout``, for example, can be used to restore the content of
424 file(s) to that one of a commit. Like this::
426 git checkout HEAD~ README
428 The commands restores the contents of README file to the last but one
429 commit in the current branch. By default the commit ID is simply HEAD;
430 i.e. ``git checkout README`` restores README to the latest commit.
432 (Do not use ``git checkout`` to view a content of a file in a commit,
433 use ``git cat-file -p``; e.g. ``git cat-file -p HEAD~:path/to/README``).
435 git reset: remove (non-pushed) commits
436 --------------------------------------
438 ``git reset`` moves the head of the current branch. The head can be
439 moved to point to any commit but it's often used to remove a commit or
440 a few (preferably, non-pushed ones) from the top of the branch - that
441 is, to move the branch backward in order to undo a few (non-pushed)
444 ``git reset`` has three modes of operation - soft, hard and mixed.
445 Default is mixed. ProGit `explains
446 <https://git-scm.com/book/en/Git-Tools-Reset-Demystified>`_ the
447 difference very clearly. Bare repositories don't have indices or
448 working trees so in a bare repo only soft reset is possible.
453 Mixed mode reset with a path or paths can be used to unstage changes -
454 that is, to remove from index changes added with ``git add`` for
455 committing. See `The Book
456 <https://git-scm.com/book/en/Git-Basics-Undoing-Things>`_ for details
457 about unstaging and other undo tricks.
459 git reflog: reference log
460 -------------------------
462 Removing commits with ``git reset`` or moving the head of a branch
463 sounds dangerous and it is. But there is a way to undo: another
464 reset back to the original commit. Git doesn't remove commits
465 immediately; unreferenced commits (in git terminology they are called
466 "dangling commits") stay in the database for some time (default is two
467 weeks) so you can reset back to it or create a new branch pointing to
470 For every move of a branch's head - with ``git commit``, ``git
471 checkout``, ``git fetch``, ``git pull``, ``git rebase``, ``git reset``
472 and so on - git stores a reference log (reflog for short). For every
473 move git stores where the head was. Command ``git reflog`` can be used
474 to view (and manipulate) the log.
476 In addition to the moves of the head of every branch git stores the
477 moves of the HEAD - a symbolic reference that (usually) names the
478 current branch. HEAD is changed with ``git checkout $BRANCH``.
480 By default ``git reflog`` shows the moves of the HEAD, i.e. the
481 command is equivalent to ``git reflog HEAD``. To show the moves of the
482 head of a branch use the command ``git reflog $BRANCH``.
484 So to undo a ``git reset`` lookup the original commit in ``git
485 reflog``, verify it with ``git show`` or ``git log`` and run ``git
486 reset $COMMIT_ID``. Git stores the move of the branch's head in
487 reflog, so you can undo that undo later again.
489 In a more complex situation you'd want to move some commits along with
490 resetting the head of the branch. Cherry-pick them to the new branch.
491 For example, if you want to reset the branch ``master`` back to the
492 original commit but preserve two commits created in the current branch
495 $ git branch save-master # create a new branch saving master
496 $ git reflog # find the original place of master
497 $ git reset $COMMIT_ID
498 $ git cherry-pick save-master~ save-master
499 $ git branch -D save-master # remove temporary branch
501 git revert: revert a commit
502 ---------------------------
504 ``git revert`` reverts a commit or commits, that is, it creates a new
505 commit or commits that revert(s) the effects of the given commits.
506 It's the only way to undo published commits (``git commit --amend``,
507 ``git rebase`` and ``git reset`` change the branch in
508 non-fast-forwardable ways so they should only be used for non-pushed
511 There is a problem with reverting a merge commit. ``git revert`` can
512 undo the code created by the merge commit but it cannot undo the fact
513 of merge. See the discussion `How to revert a faulty merge
514 <https://www.kernel.org/pub/software/scm/git/docs/howto/revert-a-faulty-merge.html>`_.
516 One thing that cannot be undone
517 -------------------------------
519 Whatever you undo, there is one thing that cannot be undone -
520 overwritten uncommitted changes. Uncommitted changes don't belong to
521 git so git cannot help preserving them.
523 Most of the time git warns you when you're going to execute a command
524 that overwrites uncommitted changes. Git warns you when you try to
525 switch branches with ``git checkout``. It warns you when you're going
526 to rebase with non-clean working tree. It refuses to pull new commits
527 over non-committed files.
529 But there are commands that do exactly that - overwrite files in the
530 working tree. Commands like ``git checkout $PATHs`` or ``git reset
531 --hard`` silently overwrite files including your uncommitted changes.
533 With that in mind you can understand the stance "commit early, commit
534 often". Commit as often as possible. Commit on every save in your
535 editor or IDE. You can edit your commits before pushing - change,
536 reorder, combine, remove. But save your changes in git database,
537 either commit changes or at least stash them with ``git stash``.
543 Internet is full of heated discussions on the topic: "merge or
544 rebase?" Most of them are meaningless. When a DVCS is being used in a
545 big team with a big and complex project with many branches there is
546 simply no way to avoid merges. So the question's diminished to
547 "whether to use rebase, and if yes - when to use rebase?" Considering
548 that it is very much recommended not to rebase published commits the
549 question's diminished even further: "whether to use rebase on
552 That small question is for the team to decide. The author of the PEP
553 recommends to use rebase when pulling, i.e. always do ``git pull
554 --rebase`` or even configure automatic setup of rebase for every new
557 $ git config branch.autosetuprebase always
559 and configure rebase for existing branches::
561 $ git config branch.$NAME.rebase true
565 $ git config branch.v1.rebase true
566 $ git config branch.master.rebase true
568 After that ``git pull origin master`` becomes equivalent to ``git pull
569 --rebase origin master``.
571 In case when merge is preferred it is recommended to create new
572 commits in a separate feature or topic branch while using rebase to
573 update the mainline branch. When the topic branch is ready merge it
574 into mainline. To avoid a tedious task of resolving large number of
575 conflicts at once you can merge the topic branch to the mainline from
576 time to time and switch back to the topic branch to continue working
577 on it. The entire workflow would be something like::
579 $ git checkout -b issue-42 # create a new issue branch and switch to it
580 ...edit/test/commit...
581 $ git checkout master
582 $ git pull --rebase origin master # update master from the upstream
584 $ git branch -d issue-42 # delete the topic branch
585 $ git push origin master
587 When the topic branch is deleted only the label is removed, commits
588 are stayed in the database, they are now merged into master::
590 o--o--o--o--o--M--< master - the mainline branch
592 --*--*--* - the topic branch, now unnamed
594 The topic branch is deleted to avoid cluttering branch namespace with
595 small topic branches. Information on what issue was fixed or what
596 feature was implemented should be in the commit messages.
602 Git has a builtin merge strategy for what Python core developers call
605 $ git merge -s ours v1 # null-merge v1 into master
608 Advanced configuration
609 ======================
614 Git has builtin mechanisms to handle line endings between platforms
615 with different EOL styles. To allow git to do CRLF conversion assign
616 ``text`` attribute to files using `.gitattributes
617 <https://www.kernel.org/pub/software/scm/git/docs/gitattributes.html>`_.
618 For files that have to have specific line ending assign ``eol``
619 attribute. For binary files the attribute is, naturally, ``binary``.
629 To check what attributes git uses for files use ``git check-attr``
630 command. For example::
632 $ git check-attr -a -- \*.py
641 Staging area aka index aka cache is a distinguishing feature of git.
642 Staging area is where git collects patches before committing them.
643 Separation between collecting patches and commit phases provides a
644 very useful feature of git: one can review collected patches before
645 commit and even edit them - remove some hunks, add new hunks and
648 To add files to the index use ``git add``. Collecting patches before
649 committing means you need to do that for every change, not only to add
650 new (untracked) files. To simplify committing in case you just want to
651 commit everything without reviewing run ``git commit --all`` (or just
652 ``-a``) - the command adds every changed tracked file to the index and
655 To add hunks of patches to the index use ``git add --patch`` (or just
656 ``-p``). To remove collected files from the index use ``git reset HEAD
657 -- $FILE...`` To add/inspect/remove collected hunks use ``git add
658 --interactive`` (``-i``).
660 To see the diff between the index and the last commit (i.e., collected
661 patches) use ``git diff --cached``. To see the diff between the
662 working tree and the index (i.e., uncollected patches) use just ``git
663 diff``. To see the diff between the working tree and the last commit
664 (i.e., both collected and uncollected patches) use ``git diff HEAD``.
667 <https://git.wiki.kernel.org/index.php/WhatIsTheIndex>`_ and
668 `IndexCommandQuickref
669 <https://git.wiki.kernel.org/index.php/IndexCommandQuickref>`_ in Git
676 https://git-scm.com/book/en/Git-Tools-Rerere
682 TODO: dangling objects, git gc, git repack.
684 https://gcc.gnu.org/ml/gcc/2007-12/msg00165.html
686 http://vcscompare.blogspot.ru/2008/06/git-repack-parameters.html
692 TODO: sticky options; example: git grep -O.
694 TODO: tricky options; example: git log -p3.
696 TODO: bash/zsh completion, bash/zsh prompt.
697 https://git.kernel.org/cgit/git/git.git/tree/contrib/completion
703 TODO: anonymous access; git over ssh; gitolite; gitweb; cgit; gitlab.
705 http://gitolite.com/gitolite/index.html
707 https://git.kernel.org/cgit/git/git.git/tree/gitweb
709 http://git.zx2c4.com/cgit/
711 From Mercurial to git
712 =====================
714 Mercurial for Git users https://mercurial.selenic.com/wiki/GitConcepts
716 https://github.com/felipec/git-remote-hg
718 https://hg-git.github.io/
730 This document has been placed in the public domain.
737 indent-tabs-mode: nil
738 sentence-end-double-space: t
742 vim: set fenc=us-ascii tw=70 :