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.
77 Microsoft Windows: download `git-for-windows
78 <https://github.com/git-for-windows/git/releases>`_ or `msysGit
79 <https://github.com/msysgit/msysgit/releases>`_.
81 MacOS X: use git installed with `XCode
82 <https://developer.apple.com/xcode/downloads/>`_ or download from
83 `MacPorts <https://www.macports.org/ports.php?by=name&substr=git>`_ or
85 <http://sourceforge.net/projects/git-osx-installer/files/>`_ or
86 install git with `Homebrew <http://brew.sh/>`_: ``brew install git``.
88 `git-cola <https://git-cola.github.io/index.html>`_ is a Git GUI
89 written in Python and GPL licensed. Linux, Windows, MacOS X.
91 `TortoiseGit <https://tortoisegit.org/>`_ is a Windows Shell Interface
92 to Git based on TortoiseSVN; open source.
97 This simple code is often appears in documentation, but it is
98 important so let repeat it here. Git stores author and committer
99 names/emails in every commit, so configure your real name and
102 $ git config --global user.name "User Name"
103 $ git config --global user.email user.name@example.org
109 Examples of git commands in this PEP use the following approach. It is
110 supposed that you, the user, works with a local repository named
111 ``python`` that has an upstream remote repo named ``origin``. Your
112 local repo has two branches ``v1`` and ``v2``. For most examples the
113 currently checked out branch is ``v2``. That is, it's assumed you have
114 done something like that::
116 $ git clone -b v2 http://git.python.org/python.git
118 $ git branch v1 origin/v1
120 The first command clones remote repository into local directory
121 `python``, creates a new local branch v2, sets remotes/origin/v2 as
122 its upstream remote-tracking branch and checks it out into the working
125 The last command creates a new local branch v1 and sets
126 remotes/origin/v1 as its upstream remote-tracking branch.
128 The same result can be achieved with commands::
130 $ git clone -b v1 http://git.python.org/python.git
132 $ git checkout --track origin/v2
134 The last command creates a new local branch v2, sets remotes/origin/v2
135 as its upstream remote-tracking branch and checks it out into the
139 Branches and branches
140 =====================
142 Git terminology can be a bit misleading. Take, for example, the term
143 "branch". In git it has two meanings. A branch is a directed line of
144 commits (possibly with merges). And a branch is a label or a pointer
145 assigned to a line of commits. It is important to distinguish when you
146 talk about commits and when about their labels. Lines of commits are
147 by itself unnamed and are usually only lengthening and merging.
148 Labels, on the other hand, can be created, moved, renamed and deleted
152 Remote repositories and remote branches
153 =======================================
155 Remote-tracking branches are branches (pointers to commits) in your
156 local repository. They are there for you to remember what branches and
157 commits have been pulled from and pushed to what remote repos (you can
158 pull from and push to many remotes). Remote-tracking branches live
159 under ``remotes/$REMOTE`` namespaces, e.g. ``remotes/origin/v2``.
161 To see the status of remote-tracking branches run::
165 To see local and remote-tracking branches (and tags) pointing to
170 You never do your own development on remote-tracking branches. You
171 create a local branch that has a remote branch as upstream and do
172 development on that local branch. On push git pushes commits to the
173 remote repo and updates remote-tracking branches, on pull git fetches
174 commits from the remote repo, updates remote-tracking branches and
175 fast-forwards, merges or rebases local branches.
177 When you do an initial clone like this::
179 $ git clone -b v1 http://git.python.org/python.git
181 git clones remote repository ``http://git.python.org/python.git`` to
182 directory ``python``, creates remote-tracking branches, creates a
183 local branch ``v1``, configure it to track upstream remotes/origin/v1
184 branch and checks out ``v1`` into the working directory.
186 Updating local and remote-tracking branches
187 -------------------------------------------
189 There is a major difference between
193 $ git fetch $REMOTE $BRANCH
199 $ git fetch $REMOTE $BRANCH:$BRANCH
201 The first command fetches commits from the named $BRANCH in the
202 $REMOTE repository that are not in your repository and leaves the id
203 (the hash) of the head commit in file .git/FETCH_HEAD and updates
204 remote-tracking branch.
206 The second command fetches commits from the named $BRANCH in the
207 $REMOTE repository that are not in your repository and updates both
208 the local branch $BRANCH and its upstream remote-tracking branch. But
209 it refuses to update branches in case of non-fast-forward. And it
210 refuses to update the current branch.
212 The first command is used internally by ``git pull``.
216 $ git pull $REMOTE $BRANCH
222 $ git fetch $REMOTE $BRANCH
223 $ git merge FETCH_HEAD
225 Certainly, $BRANCH in that case should be your current branch. If you
226 want to merge a different branch into your current branch first update
227 that non-current branch and then merge::
229 $ git fetch origin v1:v1 # Update v1
230 $ git pull --rebase origin v2 # Update the current branch v2 using
231 # rebase instead of merge
234 If you have not yet pushed commits on ``v1``, though, the scenario has
235 to become a bit more complex. Git refuses to update
236 non-fast-forwardable branch, and you don't want to do force-pull
237 because that would remove your non-pushed commits and you would need
238 to recover. So you want to rebase ``v1`` but you cannot rebase
239 non-current branch. Hence, checkout ``v1`` and rebase it before
243 $ git pull --rebase origin v1
245 $ git pull --rebase origin v2
248 It is possible to configure git to make it fetch/pull a few branches
249 or all branches at once, so you can simply run
261 Default remote repository for fetching/pulling is origin. Default set
262 of references to fetch is calculated using matching algorithm: git
263 fetches all branches having the same name on both ends.
268 Pushing is a bit simpler. There is only one command ``push``. When you
273 $ git push origin v1 v2
275 git pushes local v1 to remote v1 and local v2 to remote v2. The same
278 $ git push origin v1:v1 v2:v2
280 Git pushes commits to the remote repo and updates remote-tracking
281 branches. Git refuses to push commits that aren't fast-forwardable.
282 You can force-push anyway, but please remember - you can force-push to
283 your own repositories but don't force-push to public or shared repos.
284 If you find git refuses to push commits that aren't fast-forwardable,
285 better fetch and merge commits from the remote repo (or rebase your
286 commits on top of the fetched commits), then push. Only force-push if
287 you know what you do and why you do it. See the section `Commit
288 editing and caveats`_ below.
290 It is possible to configure git to make it push a few branches or all
291 branches at once, so you can simply run
303 Default remote repository for pushing is origin. Default set
304 of references to push in git before 2.0 is calculated using matching
305 algorithm: git pushes all branches having the same name on both ends.
306 Default set of references to push in git 2.0+ is calculated using
307 simple algorithm: git pushes the current branch back to its
310 To configure git before 2.0 to the new behaviour run::
312 $ git config push.default simple
314 To configure git 2.0+ to the old behaviour run::
316 $ git config push.default matching
318 Git refuses to push a branch if it's the current branch in the remote
319 non-bare repository: git refuses to update remote working directory.
320 You really should push only to bare repositories. For non-bare
321 repositories git prefers pull-based workflow.
323 When you want to deploy code on a remote host and can only use push
324 (because your workstation is behind a firewall and you cannot pull
325 from it) you do that in two steps using two repositories: you push
326 from the workstation to a bare repo on the remote host, ssh to the
327 remote host and pull from the bare repo to a non-bare deployment repo.
329 That changed in git 2.3, but see `the blog post
330 <https://github.com/blog/1957-git-2-3-has-been-released#push-to-deploy>`_
331 for caveats; in 2.4 the push-to-deploy feature was `further improved
332 <https://github.com/blog/1994-git-2-4-atomic-pushes-push-to-deploy-and-more#push-to-deploy-improvements>`_.
337 Git automatically fetches tags that point to commits being fetched
338 during fetch/pull. To fetch all tags (and commits they point to) run
339 ``git fetch --tags origin``. To fetch some specific tags fetch them
342 $ git fetch origin tag $TAG1 tag $TAG2...
346 $ git fetch origin tag 1.4.2 tag 2.1.7
348 Git doesn't automatically pushes tags. That allows you to have private
349 tags (lightweight tags are also private for a repo, they cannot be
350 pushed). To push tags list them explicitly::
352 $ git push origin tag 1.4.2
353 $ git push origin v1 v2 tag 2.1.7
355 Don't move tags with ``git tag -f`` or remove tags with ``git tag -d``
356 after they have been published.
359 Commit editing and caveats
360 ==========================
362 A warning not to edit published (pushed) commits also appears in
363 documentation but it's repeated here anyway as it's very important.
365 It is possible to recover from forced push but it's PITA for the
366 entire team. Please avoid it.
368 To see what commits have not been published yet compare the head of the
369 branch with its upstream remote-tracking branch::
371 $ git log origin/v2..
372 $ git log origin/v1..v1
374 For every branch that has an upstream remote-tracking branch git
375 maintains an alias @{upstream} (short version @{u}), so the commands
376 above can be given as::
381 To see the status of all branches::
385 To compare the status of local branches with a remote repo::
387 $ git remote show origin
389 Read `how to recover from upstream rebase
390 <https://git-scm.com/docs/git-rebase#_recovering_from_upstream_rebase>`_.
391 It is in ``git help rebase``.
393 On the other hand don't be too afraid about commit editing. You can
394 safely edit, remove, reorder, combine and split commits that hasn't
395 been pushed yet. You can even push commits to your own (backup) repo,
396 edit them later and force-push edited commits to replace what has
397 already been pushed. Not a problem until commits are in a public
398 or shared repository.
404 Whatever you do, don't panic. Almost anything in git can be undone.
406 git checkout: restore file's content
407 ------------------------------------
409 ``git checkout``, for example, can be used to restore the content of
410 file(s) to that one of a commit. Like this::
412 git checkout HEAD~ README
414 The commands restores the contents of README file to the last but one
415 commit in the current branch. By default the commit ID is simply HEAD;
416 i.e. ``git checkout README`` restores README to the latest commit.
418 (Do not use ``git checkout`` to view a content of a file in a commit,
419 use ``git cat-file -p``; e.g. ``git cat-file -p HEAD~:path/to/README``).
421 git reset: remove (non-pushed) commits
422 --------------------------------------
424 ``git reset`` moves the head of the current branch. The head can be
425 moved to point to any commit but it's often used to remove a commit or
426 a few (preferably, non-pushed ones) from the top of the branch - that
427 is, to move the branch backward in order to undo a few (non-pushed)
430 ``git reset`` has three modes of operation - soft, hard and mixed.
431 Default is mixed. ProGit `explains
432 <https://git-scm.com/book/en/Git-Tools-Reset-Demystified>`_ the
433 difference very clearly. Bare repositories don't have indices or
434 working trees so in a bare repo only soft reset is possible.
439 Mixed mode reset with a path or paths can be used to unstage changes -
440 that is, to remove from index changes added with ``git add`` for
441 committing. See `The Book
442 <https://git-scm.com/book/en/Git-Basics-Undoing-Things>`_ for details
443 about unstaging and other undo tricks.
445 git reflog: reference log
446 -------------------------
448 Removing commits with ``git reset`` or moving the head of a branch
449 sounds dangerous and it is. But there is a way to undo: another
450 reset back to the original commit. Git doesn't remove commits
451 immediately; unreferenced commits (in git terminology they are called
452 "dangling commits") stay in the database for some time (default is two
453 weeks) so you can reset back to it or create a new branch pointing to
456 For every move of a branch's head - with ``git commit``, ``git
457 checkout``, ``git fetch``, ``git pull``, ``git rebase``, ``git reset``
458 and so on - git stores a reference log (reflog for short). For every
459 move git stores where the head was. Command ``git reflog`` can be used
460 to view (and manipulate) the log.
462 In addition to the moves of the head of every branch git stores the
463 moves of the HEAD - a symbolic reference that (usually) names the
464 current branch. HEAD is changed with ``git checkout $BRANCH``.
466 By default ``git reflog`` shows the moves of the HEAD, i.e. the
467 command is equivalent to ``git reflog HEAD``. To show the moves of the
468 head of a branch use the command ``git reflog $BRANCH``.
470 So to undo a ``git reset`` lookup the original commit in ``git
471 reflog``, verify it with ``git show`` or ``git log`` and run ``git
472 reset $COMMIT_ID``. Git stores the move of the branch's head in
473 reflog, so you can undo that undo later again.
475 In a more complex situation you'd want to move some commits along with
476 resetting the head of the branch. Cherry-pick them to the new branch.
477 For example, if you want to reset the branch ``v2`` back to the
478 original commit but preserve two commits created in the current branch
481 $ git branch save-v2 # create a new branch saving v2
482 $ git reflog # find the original place of v2
483 $ git reset $COMMIT_ID
484 $ git cherry-pick save-v2~ save-v2
485 $ git branch -D save-v2 # remove temporary branch
487 git revert: revert a commit
488 ---------------------------
490 ``git revert`` reverts a commit or commits, that is, it creates a new
491 commit or commits that reverts the effects of the given commits. It's
492 the only way to undo published commits (``git commit --amend``, ``git
493 rebase`` and ``git reset`` change the branch in non-fast-forwardable
494 ways so they should only be used for non-pushed commits.)
496 There is a problem with reverting a merge commit. ``git revert`` can
497 undo the code created by the merge commit but it cannot undo the fact
498 of merge. See the discussion `How to revert a faulty merge
499 <https://www.kernel.org/pub/software/scm/git/docs/howto/revert-a-faulty-merge.html>`_.
501 One thing that cannot be undone
502 -------------------------------
504 Whatever you undo, there is one thing that cannot be undone -
505 overwritten uncommitted changes. Uncommitted changes don't belong to
506 git so git cannot help preserving them.
508 Most of the time git warns you when you're going to execute a command
509 that overwrites uncommitted changes. Git warns you when you try to
510 switch branches with ``git checkout``. It warns you when you're going
511 to rebase with non-clean working tree. It refuses to pull new commits
512 over non-committed files.
514 But there are commands that do exactly that - overwrite files in the
515 working tree. Commands like ``git checkout $PATHs`` or ``git reset
516 --hard`` silently overwrite files including your uncommitted changes.
518 With that in mind you can understand the stance "commit early, commit
519 often". Commit as often as possible. Commit on every save in your
520 editor or IDE. You can edit your commits before pushing - change,
521 reorder, combine, remove. But save your changes in git database,
522 either commit changes or at least stash them with ``git stash``.
528 Internet is full of heated discussions on the topic: "merge or
529 rebase?" Most of them are meaningless. When a DVCS is being used in a
530 big team with a big and complex project with many branches there is
531 simply no way to avoid merges. So the question's diminished to
532 "whether to use rebase, and if yes - when to use rebase?" Considering
533 that it is very much recommended not to rebase published commits the
534 question's diminished even further: "whether to use rebase on
537 That small question is for the team to decide. The author of the PEP
538 recommends to use rebase when pulling, i.e. always do ``git pull
539 --rebase`` or even configure automatic setup of rebase for every new
542 $ git config branch.autosetuprebase always
544 and configure rebase for existing branches::
546 $ git config branch.$NAME.rebase true
550 $ git config branch.v1.rebase true
551 $ git config branch.v2.rebase true
553 After that ``git pull origin v2`` becomes equivalent to ``git pull
554 --rebase origin v2``.
556 In case when merge is preferred it is recommended to create new
557 commits in a separate feature or topic branch while using rebase to
558 update the mainline branch. When the topic branch is ready merge it
559 into mainline. To avoid a tedious task of resolving large number of
560 conflicts at once you can merge the topic branch to the mainline from
561 time to time and switch back to the topic branch to continue working
562 on it. The entire workflow would be something like::
564 $ git checkout -b issue-42 # create a new issue branch and switch to it
565 ...edit/test/commit...
567 $ git pull --rebase origin v2 # update v2 from the upstream
569 $ git branch -d issue-42 # delete the topic branch
572 When the topic branch is deleted only the label is removed, commits
573 are stayed in the database, they are now merged into v2::
575 o--o--o--o--o--M--< v2 - the mainline branch
577 --*--*--* - the topic branch, now unnamed
579 The topic branch is deleted to avoid cluttering branch namespace with
580 small topic branches. Information on what issue was fixed or what
581 feature was implemented should be in the commit messages.
587 Git has a builtin merge strategy for what Python core developers call
590 $ git merge -s ours v1 # null-merge v1 into v2
593 Advanced configuration
594 ======================
599 Git has builtin mechanisms to handle line endings between platforms
600 with different EOL styles. To allow git to do CRLF conversion assign
601 ``text`` attribute to files using `.gitattributes
602 <https://www.kernel.org/pub/software/scm/git/docs/gitattributes.html>`_.
603 For files that have to have specific line ending assign ``eol``
604 attribute. For binary files the attribute is, naturally, ``binary``.
614 To check what attributes git uses for files use ``git check-attr``
624 Staging area aka index is a distinguishing feature of git. See
626 <https://git.wiki.kernel.org/index.php/WhatIsTheIndex>`_ and
627 `IndexCommandQuickref
628 <https://git.wiki.kernel.org/index.php/IndexCommandQuickref>`_ in Git
635 https://git-scm.com/book/en/Git-Tools-Rerere
641 TODO: dangling objects, git gc, git repack.
643 https://gcc.gnu.org/ml/gcc/2007-12/msg00165.html
645 http://vcscompare.blogspot.ru/2008/06/git-repack-parameters.html
651 TODO: sticky options; example: git grep -O.
653 TODO: tricky options; example: git log -p3.
655 TODO: bash/zsh completion, bash/zsh prompt.
656 https://git.kernel.org/cgit/git/git.git/tree/contrib/completion
662 TODO: anonymous access; git over ssh; gitolite; gitweb; cgit; gitlab.
664 http://gitolite.com/gitolite/index.html
666 https://git.kernel.org/cgit/git/git.git/tree/gitweb
668 http://git.zx2c4.com/cgit/
670 From Mercurial to git
671 =====================
673 Mercurial for Git users https://mercurial.selenic.com/wiki/GitConcepts
675 https://github.com/felipec/git-remote-hg
677 https://hg-git.github.io/
689 This document has been placed in the public domain.
696 indent-tabs-mode: nil
697 sentence-end-double-space: t
701 vim: set fenc=us-ascii tw=70 :