From: Junio C Hamano Date: Wed, 20 Jan 2010 21:23:59 +0000 (-0800) Subject: Merge branch 'bg/maint-remote-update-default' into maint X-Git-Tag: v1.6.6.1~1 X-Git-Url: https://git.tokkee.org/?a=commitdiff_plain;h=66276d972c0682d49773c20ca67c33bdab9ab76d;hp=4f2e842dd0046a12bd82bb2b5aa066c686dc3f44;p=git.git Merge branch 'bg/maint-remote-update-default' into maint * bg/maint-remote-update-default: Fix "git remote update" with remotes.defalt set --- diff --git a/.gitignore b/.gitignore index 51a37b1af..ac02a580d 100644 --- a/.gitignore +++ b/.gitignore @@ -1,184 +1,187 @@ -GIT-BUILD-OPTIONS -GIT-CFLAGS -GIT-GUI-VARS -GIT-VERSION-FILE -git -git-add -git-add--interactive -git-am -git-annotate -git-apply -git-archimport -git-archive -git-bisect -git-bisect--helper -git-blame -git-branch -git-bundle -git-cat-file -git-check-attr -git-check-ref-format -git-checkout -git-checkout-index -git-cherry -git-cherry-pick -git-clean -git-clone -git-commit -git-commit-tree -git-config -git-count-objects -git-cvsexportcommit -git-cvsimport -git-cvsserver -git-daemon -git-diff -git-diff-files -git-diff-index -git-diff-tree -git-difftool -git-difftool--helper -git-describe -git-fast-export -git-fast-import -git-fetch -git-fetch--tool -git-fetch-pack -git-filter-branch -git-fmt-merge-msg -git-for-each-ref -git-format-patch -git-fsck -git-fsck-objects -git-gc -git-get-tar-commit-id -git-grep -git-hash-object -git-help -git-http-fetch -git-http-push -git-imap-send -git-index-pack -git-init -git-init-db -git-instaweb -git-log -git-lost-found -git-ls-files -git-ls-remote -git-ls-tree -git-mailinfo -git-mailsplit -git-merge -git-merge-base -git-merge-index -git-merge-file -git-merge-tree -git-merge-octopus -git-merge-one-file -git-merge-ours -git-merge-recursive -git-merge-resolve -git-merge-subtree -git-mergetool -git-mergetool--lib -git-mktag -git-mktree -git-name-rev -git-mv -git-pack-redundant -git-pack-objects -git-pack-refs -git-parse-remote -git-patch-id -git-peek-remote -git-prune -git-prune-packed -git-pull -git-push -git-quiltimport -git-read-tree -git-rebase -git-rebase--interactive -git-receive-pack -git-reflog -git-relink -git-remote -git-remote-curl -git-repack -git-replace -git-repo-config -git-request-pull -git-rerere -git-reset -git-rev-list -git-rev-parse -git-revert -git-rm -git-send-email -git-send-pack -git-sh-setup -git-shell -git-shortlog -git-show -git-show-branch -git-show-index -git-show-ref -git-stage -git-stash -git-status -git-stripspace -git-submodule -git-svn -git-symbolic-ref -git-tag -git-tar-tree -git-unpack-file -git-unpack-objects -git-update-index -git-update-ref -git-update-server-info -git-upload-archive -git-upload-pack -git-var -git-verify-pack -git-verify-tag -git-web--browse -git-whatchanged -git-write-tree -git-core-*/?* -gitk-wish -gitweb/gitweb.cgi -test-chmtime -test-ctype -test-date -test-delta -test-dump-cache-tree -test-genrandom -test-match-trees -test-parse-options -test-path-utils -test-sha1 -test-sigchain -common-cmds.h +/GIT-BUILD-OPTIONS +/GIT-CFLAGS +/GIT-GUI-VARS +/GIT-VERSION-FILE +/git +/git-add +/git-add--interactive +/git-am +/git-annotate +/git-apply +/git-archimport +/git-archive +/git-bisect +/git-bisect--helper +/git-blame +/git-branch +/git-bundle +/git-cat-file +/git-check-attr +/git-check-ref-format +/git-checkout +/git-checkout-index +/git-cherry +/git-cherry-pick +/git-clean +/git-clone +/git-commit +/git-commit-tree +/git-config +/git-count-objects +/git-cvsexportcommit +/git-cvsimport +/git-cvsserver +/git-daemon +/git-diff +/git-diff-files +/git-diff-index +/git-diff-tree +/git-difftool +/git-difftool--helper +/git-describe +/git-fast-export +/git-fast-import +/git-fetch +/git-fetch--tool +/git-fetch-pack +/git-filter-branch +/git-fmt-merge-msg +/git-for-each-ref +/git-format-patch +/git-fsck +/git-fsck-objects +/git-gc +/git-get-tar-commit-id +/git-grep +/git-hash-object +/git-help +/git-http-backend +/git-http-fetch +/git-http-push +/git-imap-send +/git-index-pack +/git-init +/git-init-db +/git-instaweb +/git-log +/git-lost-found +/git-ls-files +/git-ls-remote +/git-ls-tree +/git-mailinfo +/git-mailsplit +/git-merge +/git-merge-base +/git-merge-index +/git-merge-file +/git-merge-tree +/git-merge-octopus +/git-merge-one-file +/git-merge-ours +/git-merge-recursive +/git-merge-resolve +/git-merge-subtree +/git-mergetool +/git-mergetool--lib +/git-mktag +/git-mktree +/git-name-rev +/git-mv +/git-notes +/git-pack-redundant +/git-pack-objects +/git-pack-refs +/git-parse-remote +/git-patch-id +/git-peek-remote +/git-prune +/git-prune-packed +/git-pull +/git-push +/git-quiltimport +/git-read-tree +/git-rebase +/git-rebase--interactive +/git-receive-pack +/git-reflog +/git-relink +/git-remote +/git-remote-curl +/git-repack +/git-replace +/git-repo-config +/git-request-pull +/git-rerere +/git-reset +/git-rev-list +/git-rev-parse +/git-revert +/git-rm +/git-send-email +/git-send-pack +/git-sh-setup +/git-shell +/git-shortlog +/git-show +/git-show-branch +/git-show-index +/git-show-ref +/git-stage +/git-stash +/git-status +/git-stripspace +/git-submodule +/git-svn +/git-symbolic-ref +/git-tag +/git-tar-tree +/git-unpack-file +/git-unpack-objects +/git-update-index +/git-update-ref +/git-update-server-info +/git-upload-archive +/git-upload-pack +/git-var +/git-verify-pack +/git-verify-tag +/git-web--browse +/git-whatchanged +/git-write-tree +/git-core-*/?* +/gitk-git/gitk-wish +/gitweb/gitweb.cgi +/test-chmtime +/test-ctype +/test-date +/test-delta +/test-dump-cache-tree +/test-genrandom +/test-match-trees +/test-parse-options +/test-path-utils +/test-sha1 +/test-sigchain +/common-cmds.h *.tar.gz *.dsc *.deb -git.spec +/git.spec *.exe *.[aos] *.py[co] -config.mak -autom4te.cache -config.cache -config.log -config.status -config.mak.autogen -config.mak.append -configure -tags -TAGS -cscope* +*+ +/config.mak +/autom4te.cache +/config.cache +/config.log +/config.status +/config.mak.autogen +/config.mak.append +/configure +/tags +/TAGS +/cscope* *.obj *.lib *.sln @@ -188,5 +191,5 @@ cscope* *.user *.idb *.pdb -Debug/ -Release/ +/Debug/ +/Release/ diff --git a/Documentation/.gitignore b/Documentation/.gitignore index d8edd9040..1c3a9fead 100644 --- a/Documentation/.gitignore +++ b/Documentation/.gitignore @@ -8,3 +8,4 @@ gitman.info howto-index.txt doc.dep cmds-*.txt +manpage-base-url.xsl diff --git a/Documentation/Makefile b/Documentation/Makefile index 06b0c57b9..4797b2dc3 100644 --- a/Documentation/Makefile +++ b/Documentation/Makefile @@ -17,6 +17,7 @@ DOC_HTML=$(MAN_HTML) ARTICLES = howto-index ARTICLES += everyday ARTICLES += git-tools +ARTICLES += git-bisect-lk2009 # with their own formatting rules. SP_ARTICLES = howto/revert-branch-rebase howto/using-merge-subtree user-manual API_DOCS = $(patsubst %.txt,%,$(filter-out technical/api-index-skel.txt technical/api-index.txt, $(wildcard technical/api-*.txt))) @@ -103,6 +104,25 @@ ifdef DOCBOOK_SUPPRESS_SP XMLTO_EXTRA += -m manpage-suppress-sp.xsl endif +# Newer DocBook stylesheet emits warning cruft in the output when +# this is not set, and if set it shows an absolute link. Older +# stylesheets simply ignore this parameter. +# +# Distros may want to use MAN_BASE_URL=file:///path/to/git/docs/ +# or similar. +ifndef MAN_BASE_URL +MAN_BASE_URL = file://$(htmldir)/ +endif +XMLTO_EXTRA += -m manpage-base-url.xsl + +# If your target system uses GNU groff, it may try to render +# apostrophes as a "pretty" apostrophe using unicode. This breaks +# cut&paste, so you should set GNU_ROFF to force them to be ASCII +# apostrophes. Unfortunately does not work with non-GNU roff. +ifdef GNU_ROFF +XMLTO_EXTRA += -m manpage-quote-apos.xsl +endif + SHELL_PATH ?= $(SHELL) # Shell quote; SHELL_PATH_SQ = $(subst ','\'',$(SHELL_PATH)) @@ -222,6 +242,7 @@ clean: $(RM) howto-index.txt howto/*.html doc.dep $(RM) technical/api-*.html technical/api-index.txt $(RM) $(cmds_txt) *.made + $(RM) manpage-base-url.xsl $(MAN_HTML): %.html : %.txt $(QUIET_ASCIIDOC)$(RM) $@+ $@ && \ @@ -229,7 +250,10 @@ $(MAN_HTML): %.html : %.txt $(ASCIIDOC_EXTRA) -agit_version=$(GIT_VERSION) -o $@+ $< && \ mv $@+ $@ -%.1 %.5 %.7 : %.xml +manpage-base-url.xsl: manpage-base-url.xsl.in + sed "s|@@MAN_BASE_URL@@|$(MAN_BASE_URL)|" $< > $@ + +%.1 %.5 %.7 : %.xml manpage-base-url.xsl $(QUIET_XMLTO)$(RM) $@ && \ xmlto -m $(MANPAGE_XSL) $(XMLTO_EXTRA) man $< diff --git a/Documentation/RelNotes-1.6.5.1.txt b/Documentation/RelNotes-1.6.5.1.txt new file mode 100644 index 000000000..309ba181b --- /dev/null +++ b/Documentation/RelNotes-1.6.5.1.txt @@ -0,0 +1,20 @@ +GIT v1.6.5.1 Release Notes +========================== + +Fixes since v1.6.5 +------------------ + + * An corrupt pack could make codepath to read objects into an + infinite loop. + + * Download throughput display was always shown in KiB/s but on fast links + it is more appropriate to show it in MiB/s. + + * "git grep -f filename" used uninitialized variable and segfaulted. + + * "git clone -b branch" gave a wrong commit object name to post-checkout + hook. + + * "git pull" over http did not work on msys. + +Other minor documentation updates are included. diff --git a/Documentation/RelNotes-1.6.5.2.txt b/Documentation/RelNotes-1.6.5.2.txt new file mode 100644 index 000000000..aa7ccce3a --- /dev/null +++ b/Documentation/RelNotes-1.6.5.2.txt @@ -0,0 +1,19 @@ +GIT v1.6.5.2 Release Notes +========================== + +Fixes since v1.6.5.1 +-------------------- + + * Installation of templates triggered a bug in busybox when using tar + implementation from it. + + * "git add -i" incorrectly ignored paths that are already in the index + if they matched .gitignore patterns. + + * "git describe --always" should have produced some output even there + were no tags in the repository, but it didn't. + + * "git ls-files" when showing tracked files incorrectly paid attention + to the exclude patterns. + +Other minor documentation updates are included. diff --git a/Documentation/RelNotes-1.6.5.3.txt b/Documentation/RelNotes-1.6.5.3.txt new file mode 100644 index 000000000..b2fad1b22 --- /dev/null +++ b/Documentation/RelNotes-1.6.5.3.txt @@ -0,0 +1,63 @@ +Git v1.6.5.3 Release Notes +========================== + +Fixes since v1.6.5.2 +-------------------- + + * info/grafts file didn't ignore trailing CR at the end of lines. + + * Packages generated on newer FC were unreadable by older versions of + RPM as the new default is to use stronger hash. + + * output from "git blame" was unreadable when the file ended in an + incomplete line. + + * "git add -i/-p" didn't handle deletion of empty files correctly. + + * "git clone" takes up to two parameters, but did not complain when + given more arguments than necessary and silently ignored them. + + * "git cvsimport" did not read files given as command line arguments + correctly when it is run from a subdirectory. + + * "git diff --color-words -U0" didn't work correctly. + + * The handling of blank lines at the end of file by "git diff/apply + --whitespace" was inconsistent with the other kinds of errors. + They are now colored, warned against, and fixed the same way as others. + + * There was no way to allow blank lines at the end of file without + allowing extra blanks at the end of lines. You can use blank-at-eof + and blank-at-eol whitespace error class to specify them separately. + The old trailing-space error class is now a short-hand to set both. + + * "-p" option to "git format-patch" was supposed to suppress diffstat + generation, but it was broken since 1.6.1. + + * "git imap-send" did not compile cleanly with newer OpenSSL. + + * "git help -a" outside of a git repository was broken. + + * "git ls-files -i" was supposed to be inverse of "git ls-files" without -i + with respect to exclude patterns, but it was broken since 1.6.5.2. + + * "git ls-remote" outside of a git repository over http was broken. + + * "git rebase -i" gave bogus error message when the command word was + misspelled. + + * "git receive-pack" that is run in response to "git push" did not run + garbage collection nor update-server-info, but in larger hosting sites, + these almost always need to be run. To help site administrators, the + command now runs "gc --auto" and "u-s-i" by setting receive.autogc + and receive.updateserverinfo configuration variables, respectively. + + * Release notes spelled the package name with incorrect capitalization. + + * "gitweb" did not escape non-ascii characters correctly in the URL. + + * "gitweb" showed "patch" link even for merge commits. + + * "gitweb" showed incorrect links for blob line numbers in pathinfo mode. + +Other minor documentation updates are included. diff --git a/Documentation/RelNotes-1.6.5.4.txt b/Documentation/RelNotes-1.6.5.4.txt new file mode 100644 index 000000000..e42f8b239 --- /dev/null +++ b/Documentation/RelNotes-1.6.5.4.txt @@ -0,0 +1,32 @@ +Git v1.6.5.4 Release Notes +========================== + +Fixes since v1.6.5.3 +-------------------- + + * "git help" (without argument) used to check if you are in a directory + under git control. There was no breakage in behaviour per-se, but this + was unnecessary. + + * "git prune-packed" gave progress output even when its standard error is + not connected to a terminal; this caused cron jobs that run it to + produce crufts. + + * "git pack-objects --all-progress" is an option to ask progress output + from write-object phase _if_ progress output were to be produced, and + shouldn't have forced the progress output. + + * "git apply -p --directory=" did not work well for a + non-default value of n. + + * "git merge foo HEAD" was misparsed as an old-style invocation of the + command and produced a confusing error message. As it does not specify + any other branch to merge, it shouldn't be mistaken as such. We will + remove the old style "git merge HEAD ..." syntax in + future versions, but not in this release, + + * "git merge -m ..." added the standard merge message + on its own after user-supplied message, which should have overrided the + standard one. + +Other minor documentation updates are included. diff --git a/Documentation/RelNotes-1.6.5.5.txt b/Documentation/RelNotes-1.6.5.5.txt new file mode 100644 index 000000000..ecfc57d87 --- /dev/null +++ b/Documentation/RelNotes-1.6.5.5.txt @@ -0,0 +1,49 @@ +Git v1.6.5.5 Release Notes +========================== + +Fixes since v1.6.5.4 +-------------------- + + * Manual pages can be formatted with older xmlto again. + + * GREP_OPTIONS exported from user's environment could have broken + our scripted commands. + + * In configuration files, a few variables that name paths can begin with + ~/ and ~username/ and they are expanded as expected. This is not a + bugfix but 1.6.6 will have this and without backporting users cannot + easily use the same ~/.gitconfig across versions. + + * "git diff -B -M" did the same computation to hash lines of contents + twice, and held onto memory after it has used the data in it + unnecessarily before it freed. + + * "git diff -B" and "git diff --dirstat" was not counting newly added + contents correctly. + + * "git format-patch revisions... -- path" issued an incorrect error + message that suggested to use "--" on the command line when path + does not exist in the current work tree (it is a separate matter if + it makes sense to limit format-patch with pathspecs like that + without using the --full-diff option). + + * "git grep -F -i StRiNg" did not work as expected. + + * Enumeration of available merge strategies iterated over the list of + commands in a wrong way, sometimes producing an incorrect result. + + * "git shortlog" did not honor the "encoding" header embedded in the + commit object like "git log" did. + + * Reading progress messages that come from the remote side while running + "git pull" is given precedence over reading the actual pack data to + prevent garbled progress message on the user's terminal. + + * "git rebase" got confused when the log message began with certain + strings that looked like Subject:, Date: or From: header. + + * "git reset" accidentally run in .git/ directory checked out the + work tree contents in there. + + +Other minor documentation updates are included. diff --git a/Documentation/RelNotes-1.6.5.6.txt b/Documentation/RelNotes-1.6.5.6.txt new file mode 100644 index 000000000..a9eaf76f6 --- /dev/null +++ b/Documentation/RelNotes-1.6.5.6.txt @@ -0,0 +1,23 @@ +Git v1.6.5.6 Release Notes +========================== + +Fixes since v1.6.5.5 +-------------------- + + * "git add -p" had a regression since v1.6.5.3 that broke deletion of + non-empty files. + + * "git archive -o o.zip -- Makefile" produced an archive in o.zip + but in POSIX tar format. + + * Error message given to "git pull --rebase" when the user didn't give + enough clue as to what branch to integrate with still talked about + "merging with" the branch. + + * Error messages given by "git merge" when the merge resulted in a + fast-forward still were in plumbing lingo, even though in v1.6.5 + we reworded messages in other cases. + + * The post-upload-hook run by upload-pack in response to "git fetch" has + been removed, due to security concerns (the hook first appeared in + 1.6.5). diff --git a/Documentation/RelNotes-1.6.5.7.txt b/Documentation/RelNotes-1.6.5.7.txt new file mode 100644 index 000000000..5b49ea53b --- /dev/null +++ b/Documentation/RelNotes-1.6.5.7.txt @@ -0,0 +1,19 @@ +Git v1.6.5.7 Release Notes +========================== + +Fixes since v1.6.5.6 +-------------------- + +* If a user specifies a color for a (i.e. a class of things to show + in a particular color) that is known only by newer versions of git + (e.g. "color.diff.func" was recently added for upcoming 1.6.6 release), + an older version of git should just ignore them. Instead we diagnosed + it as an error. + +* With help.autocorrect set to non-zero value, the logic to guess typoes + in the subcommand name misfired and ran a random nonsense command. + +* If a command is run with an absolute path as a pathspec inside a bare + repository, e.g. "rev-list HEAD -- /home", the code tried to run + strlen() on NULL, which is the result of get_git_work_tree(), and + segfaulted. diff --git a/Documentation/RelNotes-1.6.5.8.txt b/Documentation/RelNotes-1.6.5.8.txt new file mode 100644 index 000000000..8b24bebb9 --- /dev/null +++ b/Documentation/RelNotes-1.6.5.8.txt @@ -0,0 +1,28 @@ +Git v1.6.5.8 Release Notes +========================== + +Fixes since v1.6.5.7 +-------------------- + +* "git count-objects" did not handle packfiles that are bigger than 4G on + platforms with 32-bit off_t. + +* "git rebase -i" did not abort cleanly if it failed to launch the editor. + +* "git blame" did not work well when commit lacked the author name. + +* "git fast-import" choked when handling a tag that points at an object + that is not a commit. + +* "git reset --hard" did not work correctly when GIT_WORK_TREE environment + variable is used to point at the root of the true work tree. + +* "git grep" fed a buffer that is not NUL-terminated to underlying + regexec(). + +* "git checkout -m other" while on a branch that does not have any commit + segfaulted, instead of failing. + +* "git branch -a other" should have diagnosed the command as an error. + +Other minor documentation updates are also included. diff --git a/Documentation/RelNotes-1.6.6.1.txt b/Documentation/RelNotes-1.6.6.1.txt new file mode 100644 index 000000000..406fbc4e0 --- /dev/null +++ b/Documentation/RelNotes-1.6.6.1.txt @@ -0,0 +1,30 @@ +Git v1.6.6.1 Release Notes +========================== + +Fixes since v1.6.6 +------------------ + + * "git branch -a name" wasn't diagnosed as an error. + + * "git fast-import" choked when fed a tag that do not point at a + commit. + + * "git grep" finding from work tree files could have fed garbage to + the underlying regexec(3). + + * "git grep -L" didn't show empty files (they should never match, and + they should always appear in -L output as unmatching). + + * http-backend was not listed in the command list in the documentation. + + * Building on FreeBSD (both 7 and 8) needs OLD_ICONV set in the Makefile + + * "git checkout -m some-branch" while on an unborn branch crashed. + +Other minor documentation updates are included. + +-- +exec >/var/tmp/1 +O=v1.6.6-39-g6304c40 +echo O=$(git describe maint) +git shortlog --no-merges $O..maint diff --git a/Documentation/RelNotes-1.6.6.txt b/Documentation/RelNotes-1.6.6.txt new file mode 100644 index 000000000..04e205c45 --- /dev/null +++ b/Documentation/RelNotes-1.6.6.txt @@ -0,0 +1,224 @@ +Git v1.6.6 Release Notes +======================== + +Notes on behaviour change +------------------------- + + * In this release, "git fsck" defaults to "git fsck --full" and + checks packfiles, and because of this it will take much longer to + complete than before. If you prefer a quicker check only on loose + objects (the old default), you can say "git fsck --no-full". This + has been supported by 1.5.4 and newer versions of git, so it is + safe to write it in your script even if you use slightly older git + on some of your machines. + +Preparing yourselves for compatibility issues in 1.7.0 +------------------------------------------------------ + +In git 1.7.0, which is planned to be the release after 1.6.6, there will +be a handful of behaviour changes that will break backward compatibility. + +These changes were discussed long time ago and existing behaviours have +been identified as more problematic to the userbase than keeping them for +the sake of backward compatibility. + +When necessary, a transition strategy for existing users has been designed +not to force them running around setting configuration variables and +updating their scripts in order to either keep the traditional behaviour +or adjust to the new behaviour, on the day their sysadmin decides to install +the new version of git. When we switched from "git-foo" to "git foo" in +1.6.0, even though the change had been advertised and the transition +guide had been provided for a very long time, the users procrastinated +during the entire transtion period, and ended up panicking on the day +their sysadmins updated their git installation. We are trying to avoid +repeating that unpleasantness in the 1.7.0 release. + +For changes decided to be in 1.7.0, commands that will be affected +have been much louder to strongly discourage such procrastination, and +they continue to be in this release. If you have been using recent +versions of git, you would have seen warnings issued when you used +features whose behaviour will change, with a clear instruction on how +to keep the existing behaviour if you want to. You hopefully are +already well prepared. + +Of course, we have also been giving "this and that will change in +1.7.0; prepare yourselves" warnings in the release notes and +announcement messages for the past few releases. Let's see how well +users will fare this time. + + * "git push" into a branch that is currently checked out (i.e. pointed by + HEAD in a repository that is not bare) will be refused by default. + + Similarly, "git push $there :$killed" to delete the branch $killed + in a remote repository $there, when $killed branch is the current + branch pointed at by its HEAD, will be refused by default. + + Setting the configuration variables receive.denyCurrentBranch and + receive.denyDeleteCurrent to 'ignore' in the receiving repository + can be used to override these safety features. Versions of git + since 1.6.2 have issued a loud warning when you tried to do these + operations without setting the configuration, so repositories of + people who still need to be able to perform such a push should + already have been future proofed. + + Please refer to: + + http://git.or.cz/gitwiki/GitFaq#non-bare + http://thread.gmane.org/gmane.comp.version-control.git/107758/focus=108007 + + for more details on the reason why this change is needed and the + transition process that already took place so far. + + * "git send-email" will not make deep threads by default when sending a + patch series with more than two messages. All messages will be sent + as a reply to the first message, i.e. cover letter. Git 1.6.6 (this + release) will issue a warning about the upcoming default change, when + it uses the traditional "deep threading" behaviour as the built-in + default. To squelch the warning but still use the "deep threading" + behaviour, give --chain-reply-to option or set sendemail.chainreplyto + to true. + + It has been possible to configure send-email to send "shallow thread" + by setting sendemail.chainreplyto configuration variable to false. + The only thing 1.7.0 release will do is to change the default when + you haven't configured that variable. + + * "git status" will not be "git commit --dry-run". This change does not + affect you if you run the command without pathspec. + + Nobody sane found the current behaviour of "git status Makefile" useful + nor meaningful, and it confused users. "git commit --dry-run" has been + provided as a way to get the current behaviour of this command since + 1.6.5. + + * "git diff" traditionally treated various "ignore whitespace" options + only as a way to filter the patch output. "git diff --exit-code -b" + exited with non-zero status even if all changes were about changing the + ammount of whitespace and nothing else. and "git diff -b" showed the + "diff --git" header line for such a change without patch text. + + In 1.7.0, the "ignore whitespaces" will affect the semantics of the + diff operation itself. A change that does not affect anything but + whitespaces will be reported with zero exit status when run with + --exit-code, and there will not be "diff --git" header for such a + change. + + +Updates since v1.6.5 +-------------------- + +(subsystems) + + * various gitk updates including use of themed widgets under Tk 8.5, + Japanese translation, a fix to a bug when running "gui blame" from + a subdirectory, etc. + + * various git-gui updates including new translations, wm states fixes, + Tk bug workaround after quitting, improved heuristics to trigger gc, + etc. + + * various git-svn updates. + + * "git fetch" over http learned a new mode that is different from the + traditional "dumb commit walker". + +(portability) + + * imap-send can be built on mingw port. + +(performance) + + * "git diff -B" has smaller memory footprint. + +(usability, bells and whistles) + + * The object replace mechanism can be bypassed with --no-replace-objects + global option given to the "git" program. + + * In configuration files, a few variables that name paths can begin with ~/ + and ~username/ and they are expanded as expected. + + * "git subcmd -h" now shows short usage help for many more subcommands. + + * "git bisect reset" can reset to an arbitrary commit. + + * "git checkout frotz" when there is no local branch "frotz" but there + is only one remote tracking branch "frotz" is taken as a request to + start the named branch at the corresponding remote tracking branch. + + * "git commit -c/-C/--amend" can be told with a new "--reset-author" option + to ignore authorship information in the commit it is taking the message + from. + + * "git describe" can be told to add "-dirty" suffix with "--dirty" option. + + * "git diff" learned --submodule option to show a list of one-line logs + instead of differences between the commit object names. + + * "git diff" learned to honor diff.color.func configuration to paint + function name hint printed on the hunk header "@@ -j,k +l,m @@" line + in the specified color. + + * "git fetch" learned --all and --multiple options, to run fetch from + many repositories, and --prune option to remove remote tracking + branches that went stale. These make "git remote update" and "git + remote prune" less necessary (there is no plan to remove "remote + update" nor "remote prune", though). + + * "git fsck" by default checks the packfiles (i.e. "--full" is the + default); you can turn it off with "git fsck --no-full". + + * "git grep" can use -F (fixed strings) and -i (ignore case) together. + + * import-tars contributed fast-import frontend learned more types of + compressed tarballs. + + * "git instaweb" knows how to talk with mod_cgid to apache2. + + * "git log --decorate" shows the location of HEAD as well. + + * "git log" and "git rev-list" learned to take revs and pathspecs from + the standard input with the new "--stdin" option. + + * "--pretty=format" option to "log" family of commands learned: + + . to wrap text with the "%w()" specifier. + . to show reflog information with "%g[sdD]" specifier. + + * "git notes" command to annotate existing commits. + + * "git merge" (and "git pull") learned --ff-only option to make it fail + if the merge does not result in a fast-forward. + + * "git mergetool" learned to use p4merge. + + * "git rebase -i" learned "reword" that acts like "edit" but immediately + starts an editor to tweak the log message without returning control to + the shell, which is done by "edit" to give an opportunity to tweak the + contents. + + * "git send-email" can be told with "--envelope-sender=auto" to use the + same address as "From:" address as the envelope sender address. + + * "git send-email" will issue a warning when it defaults to the + --chain-reply-to behaviour without being told by the user and + instructs to prepare for the change of the default in 1.7.0 release. + + * In "git submodule add ", is now optional and + inferred from the same way "git clone " does. + + * "git svn" learned to read SVN 1.5+ and SVK merge tickets. + + * "git svn" learned to recreate empty directories tracked only by SVN. + + * "gitweb" can optionally render its "blame" output incrementally (this + requires JavaScript on the client side). + + * Author names shown in gitweb output are links to search commits by the + author. + +Fixes since v1.6.5 +------------------ + +All of the fixes in v1.6.5.X maintenance series are included in this +release, unless otherwise noted. diff --git a/Documentation/SubmittingPatches b/Documentation/SubmittingPatches index 76fc84d87..c686f8646 100644 --- a/Documentation/SubmittingPatches +++ b/Documentation/SubmittingPatches @@ -279,6 +279,20 @@ from the list and queue it to 'pu', in order to make it easier for people play with it without having to pick up and apply the patch to their trees themselves. +------------------------------------------------ +Know the status of your patch after submission + +* You can use Git itself to find out when your patch is merged in + master. 'git pull --rebase' will automatically skip already-applied + patches, and will let you know. This works only if you rebase on top + of the branch in which your patch has been merged (i.e. it will not + tell you if your patch is merged in pu if you rebase on top of + master). + +* Read the git mailing list, the maintainer regularly posts messages + entitled "What's cooking in git.git" and "What's in git.git" giving + the status of various proposed changes. + ------------------------------------------------ MUA specific hints diff --git a/Documentation/blame-options.txt b/Documentation/blame-options.txt index 1625ffce6..4833cac4b 100644 --- a/Documentation/blame-options.txt +++ b/Documentation/blame-options.txt @@ -98,8 +98,10 @@ commit. files that were modified in the same commit. This is useful when you reorganize your program and move code around across files. When this option is given twice, - the command additionally looks for copies from all other - files in the parent for the commit that creates the file. + the command additionally looks for copies from other + files in the commit that creates the file. When this + option is given three times, the command additionally + looks for copies from other files in any commit. + is optional but it is the lower bound on the number of alphanumeric characters that git must detect as moving diff --git a/Documentation/config.txt b/Documentation/config.txt index 50a65ab8d..f7728ec40 100644 --- a/Documentation/config.txt +++ b/Documentation/config.txt @@ -126,12 +126,20 @@ advice.*:: Directions on how to stage/unstage/add shown in the output of linkgit:git-status[1] and the template shown when writing commit messages. Default: true. + commitBeforeMerge:: + Advice shown when linkgit:git-merge[1] refuses to + merge to avoid overwritting local changes. + Default: true. -- core.fileMode:: If false, the executable bit differences between the index and the working copy are ignored; useful on broken filesystems like FAT. - See linkgit:git-update-index[1]. True by default. + See linkgit:git-update-index[1]. ++ +The default is true, except linkgit:git-clone[1] or linkgit:git-init[1] +will probe and set core.fileMode false if appropriate when the +repository is created. core.ignoreCygwinFSTricks:: This option is only used by Cygwin implementation of Git. If false, @@ -144,6 +152,18 @@ core.ignoreCygwinFSTricks:: is true, in which case ignoreCygwinFSTricks is ignored as Cygwin's POSIX emulation is required to support core.filemode. +core.ignorecase:: + If true, this option enables various workarounds to enable + git to work better on filesystems that are not case sensitive, + like FAT. For example, if a directory listing finds + "makefile" when git expects "Makefile", git will assume + it is really the same file, and continue to remember it as + "Makefile". ++ +The default is false, except linkgit:git-clone[1] or linkgit:git-init[1] +will probe and set core.ignorecase true if appropriate when the repository +is created. + core.trustctime:: If false, the ctime differences between the index and the working copy are ignored; useful when the inode change time @@ -169,9 +189,10 @@ core.autocrlf:: writing to the filesystem. The variable can be set to 'input', in which case the conversion happens only while reading from the filesystem but files are written out with - `LF` at the end of lines. Currently, which paths to consider - "text" (i.e. be subjected to the autocrlf mechanism) is - decided purely based on the contents. + `LF` at the end of lines. A file is considered + "text" (i.e. be subjected to the autocrlf mechanism) based on + the file's `crlf` attribute, or if `crlf` is unspecified, + based on the file's contents. See linkgit:gitattributes[5]. core.safecrlf:: If true, makes git check if converting `CRLF` as controlled by @@ -223,7 +244,11 @@ core.symlinks:: contain the link text. linkgit:git-update-index[1] and linkgit:git-add[1] will not change the recorded type to regular file. Useful on filesystems like FAT that do not support - symbolic links. True by default. + symbolic links. ++ +The default is true, except linkgit:git-clone[1] or linkgit:git-init[1] +will probe and set core.symlinks false if appropriate when the repository +is created. core.gitProxy:: A "proxy command" to execute (as 'command host port') instead @@ -272,17 +297,24 @@ false), while all other repositories are assumed to be bare (bare = true). core.worktree:: - Set the path to the working tree. The value will not be - used in combination with repositories found automatically in - a .git directory (i.e. $GIT_DIR is not set). + Set the path to the root of the work tree. This can be overridden by the GIT_WORK_TREE environment variable and the '--work-tree' command line option. It can be - a absolute path or relative path to the directory specified by - --git-dir or GIT_DIR. - Note: If --git-dir or GIT_DIR are specified but none of + an absolute path or a relative path to the .git directory, + either specified by --git-dir or GIT_DIR, or automatically + discovered. + If --git-dir or GIT_DIR are specified but none of --work-tree, GIT_WORK_TREE and core.worktree is specified, - the current working directory is regarded as the top directory - of your working tree. + the current working directory is regarded as the root of the + work tree. ++ +Note that this variable is honored even when set in a configuration +file in a ".git" subdirectory of a directory, and its value differs +from the latter directory (e.g. "/path/to/.git/config" has +core.worktree set to "/different/path"), which is most likely a +misconfiguration. Running git commands in "/path/to" directory will +still use "/different/path" as the root of the work tree and can cause +great confusion to the users. core.logAllRefUpdates:: Enable the reflog. Updates to a ref is logged to the file @@ -380,16 +412,15 @@ Common unit suffixes of 'k', 'm', or 'g' are supported. core.excludesfile:: In addition to '.gitignore' (per-directory) and '.git/info/exclude', git looks into this file for patterns - of files which are not meant to be tracked. See - linkgit:gitignore[5]. + of files which are not meant to be tracked. "{tilde}/" is expanded + to the value of `$HOME` and "{tilde}user/" to the specified user's + home directory. See linkgit:gitignore[5]. core.editor:: Commands such as `commit` and `tag` that lets you edit messages by launching an editor uses the value of this variable when it is set, and the environment variable - `GIT_EDITOR` is not set. The order of preference is - `GIT_EDITOR` environment, `core.editor`, `VISUAL` and - `EDITOR` environment variables and then finally `vi`. + `GIT_EDITOR` is not set. See linkgit:git-var[1]. core.pager:: The command that git will use to paginate output. Can @@ -416,13 +447,17 @@ core.whitespace:: consider them as errors. You can prefix `-` to disable any of them (e.g. `-trailing-space`): + -* `trailing-space` treats trailing whitespaces at the end of the line +* `blank-at-eol` treats trailing whitespaces at the end of the line as an error (enabled by default). * `space-before-tab` treats a space character that appears immediately before a tab character in the initial indent part of the line as an error (enabled by default). * `indent-with-non-tab` treats a line that is indented with 8 or more space characters as an error (not enabled by default). +* `blank-at-eof` treats blank lines added at the end of file as an error + (enabled by default). +* `trailing-space` is a short-hand to cover both `blank-at-eol` and + `blank-at-eof`. * `cr-at-eol` treats a carriage-return at the end of line as part of the line terminator, i.e. with it, `trailing-space` does not trigger if the character before such a carriage-return @@ -454,6 +489,19 @@ On some file system/operating system combinations, this is unreliable. Set this config setting to 'rename' there; However, This will remove the check that makes sure that existing object files will not get overwritten. +core.notesRef:: + When showing commit messages, also show notes which are stored in + the given ref. This ref is expected to contain files named + after the full SHA-1 of the commit they annotate. ++ +If such a file exists in the given ref, the referenced blob is read, and +appended to the commit message, separated by a "Notes:" line. If the +given ref itself does not exist, it is not an error, but means that no +notes should be printed. ++ +This setting defaults to "refs/notes/commits", and can be overridden by +the `GIT_NOTES_REF` environment variable. + add.ignore-errors:: Tells 'git-add' to continue adding files when some files cannot be added due to indexing errors. Equivalent to the '--ignore-errors' @@ -489,7 +537,7 @@ apply.whitespace:: as the '--whitespace' option. See linkgit:git-apply[1]. branch.autosetupmerge:: - Tells 'git-branch' and 'git-checkout' to setup new branches + Tells 'git-branch' and 'git-checkout' to set up new branches so that linkgit:git-pull[1] will appropriately merge from the starting point branch. Note that even if this option is not set, this behavior can be chosen per-branch using the `--track` @@ -594,10 +642,10 @@ color.diff.:: Use customized color for diff colorization. `` specifies which part of the patch to use the specified color, and is one of `plain` (context text), `meta` (metainformation), `frag` - (hunk header), `old` (removed lines), `new` (added lines), - `commit` (commit headers), or `whitespace` (highlighting - whitespace errors). The values of these variables may be specified as - in color.branch.. + (hunk header), 'func' (function in hunk header), `old` (removed lines), + `new` (added lines), `commit` (commit headers), or `whitespace` + (highlighting whitespace errors). The values of these variables may be + specified as in color.branch.. color.grep:: When set to `always`, always highlight matches. When `false` (or @@ -666,6 +714,8 @@ color.ui:: commit.template:: Specify a file to use as the template for new commit messages. + "{tilde}/" is expanded to the value of `$HOME` and "{tilde}user/" to the + specified user's home directory. diff.autorefreshindex:: When using 'git-diff' to compare with work tree @@ -675,7 +725,7 @@ diff.autorefreshindex:: contents in the work tree match the contents in the index. This option defaults to true. Note that this affects only 'git-diff' Porcelain, and not lower level - 'diff' commands, such as 'git-diff-files'. + 'diff' commands such as 'git-diff-files'. diff.external:: If this config variable is set, diff generation is not @@ -791,8 +841,8 @@ format.pretty:: format.thread:: The default threading style for 'git-format-patch'. Can be - either a boolean value, `shallow` or `deep`. `shallow` - threading makes every mail a reply to the head of the series, + a boolean value, or `shallow` or `deep`. `shallow` threading + makes every mail a reply to the head of the series, where the head is chosen from the cover letter, the `\--in-reply-to`, and the first patch mail, in this order. `deep` threading makes every mail a reply to the previous one. @@ -825,15 +875,12 @@ gc.autopacklimit:: default value is 50. Setting this to 0 disables it. gc.packrefs:: - 'git-gc' does not run `git pack-refs` in a bare repository by - default so that older dumb-transport clients can still fetch - from the repository. Setting this to `true` lets 'git-gc' - to run `git pack-refs`. Setting this to `false` tells - 'git-gc' never to run `git pack-refs`. The default setting is - `notbare`. Enable it only when you know you do not have to - support such clients. The default setting will change to `true` - at some stage, and setting this to `false` will continue to - prevent `git pack-refs` from being run from 'git-gc'. + Running `git pack-refs` in a repository renders it + unclonable by Git versions prior to 1.5.1.2 over dumb + transports such as HTTP. This variable determines whether + 'git gc' runs `git pack-refs`. This can be set to "nobare" + to enable it within all non-bare repos or it can be set to a + boolean value. The default is `true`. gc.pruneexpire:: When 'git-gc' is run, it will call 'prune --expire 2.weeks.ago'. @@ -1089,6 +1136,14 @@ http.maxRequests:: How many HTTP requests to launch in parallel. Can be overridden by the 'GIT_HTTP_MAX_REQUESTS' environment variable. Default is 5. +http.postBuffer:: + Maximum size in bytes of the buffer used by smart HTTP + transports when POSTing data to the remote system. + For requests larger than this buffer size, HTTP/1.1 and + Transfer-Encoding: chunked is used to avoid creating a + massive pack file locally. Default is 1 MiB, which is + sufficient for most requests. + http.lowSpeedLimit, http.lowSpeedTime:: If the HTTP transfer speed is less than 'http.lowSpeedLimit' for longer than 'http.lowSpeedTime' seconds, the transfer is aborted. @@ -1320,6 +1375,11 @@ rebase.stat:: Whether to show a diffstat of what changed upstream since the last rebase. False by default. +receive.autogc:: + By default, git-receive-pack will run "git-gc --auto" after + receiving data from git-push and updating refs. You can stop + it by setting this variable to false. + receive.fsckObjects:: If it is set to true, git-receive-pack will check all received objects. It will abort in the case of a malformed object or a @@ -1351,10 +1411,14 @@ receive.denyCurrentBranch:: receive.denyNonFastForwards:: If set to true, git-receive-pack will deny a ref update which is - not a fast forward. Use this to prevent such an update via a push, + not a fast-forward. Use this to prevent such an update via a push, even if that push is forced. This configuration variable is set when initializing a shared repository. +receive.updateserverinfo:: + If set to true, git-receive-pack will run git-update-server-info + after receiving data from git-push and updating refs. + remote..url:: The URL of a remote repository. See linkgit:git-fetch[1] or linkgit:git-push[1]. diff --git a/Documentation/diff-options.txt b/Documentation/diff-options.txt index 9276faeb1..8707d0e74 100644 --- a/Documentation/diff-options.txt +++ b/Documentation/diff-options.txt @@ -14,7 +14,8 @@ endif::git-format-patch[] ifdef::git-format-patch[] -p:: - Generate patches without diffstat. +--no-stat:: + Generate plain patches without any diffstats. endif::git-format-patch[] ifndef::git-format-patch[] @@ -27,33 +28,40 @@ endif::git-format-patch[] -U:: --unified=:: Generate diffs with lines of context instead of - the usual three. Implies "-p". + the usual three. +ifndef::git-format-patch[] + Implies `-p`. +endif::git-format-patch[] +ifndef::git-format-patch[] --raw:: Generate the raw format. {git-diff-core? This is the default.} +endif::git-format-patch[] +ifndef::git-format-patch[] --patch-with-raw:: - Synonym for "-p --raw". + Synonym for `-p --raw`. +endif::git-format-patch[] --patience:: Generate a diff using the "patience diff" algorithm. --stat[=width[,name-width]]:: Generate a diffstat. You can override the default - output width for 80-column terminal by "--stat=width". + output width for 80-column terminal by `--stat=width`. The width of the filename part can be controlled by giving another width to it separated by a comma. --numstat:: - Similar to \--stat, but shows number of added and + Similar to `\--stat`, but shows number of added and deleted lines in decimal notation and pathname without abbreviation, to make it more machine friendly. For binary files, outputs two `-` instead of saying `0 0`. --shortstat:: - Output only the last line of the --stat format containing total + Output only the last line of the `--stat` format containing total number of modified files, as well as number of added and deleted lines. @@ -61,24 +69,39 @@ endif::git-format-patch[] Output the distribution of relative amount of changes (number of lines added or removed) for each sub-directory. Directories with changes below a cut-off percent (3% by default) are not shown. The cut-off percent - can be set with "--dirstat=limit". Changes in a child directory is not - counted for the parent directory, unless "--cumulative" is used. + can be set with `--dirstat=limit`. Changes in a child directory is not + counted for the parent directory, unless `--cumulative` is used. --dirstat-by-file[=limit]:: - Same as --dirstat, but counts changed files instead of lines. + Same as `--dirstat`, but counts changed files instead of lines. --summary:: Output a condensed summary of extended header information such as creations, renames and mode changes. +ifndef::git-format-patch[] --patch-with-stat:: - Synonym for "-p --stat". - {git-format-patch? This is the default.} + Synonym for `-p --stat`. +endif::git-format-patch[] + +ifndef::git-format-patch[] -z:: - NUL-line termination on output. This affects the --raw - output field terminator. Also output from commands such - as "git-log" will be delimited with NUL between commits. +ifdef::git-log[] + Separate the commits with NULs instead of with new newlines. ++ +Also, when `--raw` or `--numstat` has been given, do not munge +pathnames and use NULs as output field terminators. +endif::git-log[] +ifndef::git-log[] + When `--raw` or `--numstat` has been given, do not munge + pathnames and use NULs as output field terminators. +endif::git-log[] ++ +Without this option, each pathname output will have TAB, LF, double quotes, +and backslash characters replaced with `\t`, `\n`, `\"`, and `\\`, +respectively, and the pathname will be enclosed in double quotes if +any of those replacements occurred. --name-only:: Show only names of changed files. @@ -87,6 +110,13 @@ endif::git-format-patch[] Show only names and status of changed files. See the description of the `--diff-filter` option on what the status letters mean. +--submodule[=]:: + Chose the output format for submodule differences. can be one of + 'short' and 'log'. 'short' just shows pairs of commit names, this format + is used when this option is not given. 'log' is the default value for this + option and lists the commits in that commit range like the 'summary' + option of linkgit:git-submodule[1] does. + --color:: Show colored diff. @@ -110,16 +140,19 @@ The regex can also be set via a diff driver or configuration option, see linkgit:gitattributes[1] or linkgit:git-config[1]. Giving it explicitly overrides any diff driver or configuration setting. Diff drivers override configuration settings. +endif::git-format-patch[] --no-renames:: Turn off rename detection, even when the configuration file gives the default to do so. +ifndef::git-format-patch[] --check:: Warn if changes introduce trailing whitespace or an indent that uses a space before a tab. Exits with non-zero status if problems are found. Not compatible with --exit-code. +endif::git-format-patch[] --full-index:: Instead of the first handful of characters, show the full @@ -127,16 +160,16 @@ override configuration settings. line when generating patch format output. --binary:: - In addition to --full-index, output "binary diff" that - can be applied with "git apply". + In addition to `--full-index`, output a binary diff that + can be applied with `git-apply`. --abbrev[=]:: Instead of showing the full 40-byte hexadecimal object name in diff-raw format output and diff-tree header lines, show only a partial prefix. This is - independent of --full-index option above, which controls + independent of the `--full-index` option above, which controls the diff-patch output format. Non default number of - digits can be specified with --abbrev=. + digits can be specified with `--abbrev=`. -B:: Break complete rewrite changes into pairs of delete and create. @@ -147,6 +180,7 @@ override configuration settings. -C:: Detect copies as well as renames. See also `--find-copies-harder`. +ifndef::git-format-patch[] --diff-filter=[ACDMRTUXB*]:: Select only files that are Added (`A`), Copied (`C`), Deleted (`D`), Modified (`M`), Renamed (`R`), have their @@ -158,6 +192,7 @@ override configuration settings. paths are selected if there is any file that matches other criteria in the comparison; if there is no file that matches other criteria, nothing is selected. +endif::git-format-patch[] --find-copies-harder:: For performance reasons, by default, `-C` option finds copies only @@ -169,12 +204,13 @@ override configuration settings. `-C` option has the same effect. -l:: - -M and -C options require O(n^2) processing time where n + The `-M` and `-C` options require O(n^2) processing time where n is the number of potential rename/copy targets. This option prevents rename/copy detection from running if the number of rename/copy targets exceeds the specified number. +ifndef::git-format-patch[] -S:: Look for differences that introduce or remove an instance of . Note that this is different than the string simply @@ -182,18 +218,20 @@ override configuration settings. linkgit:gitdiffcore[7] for more details. --pickaxe-all:: - When -S finds a change, show all the changes in that + When `-S` finds a change, show all the changes in that changeset, not just the files that contain the change in . --pickaxe-regex:: Make the not a plain string but an extended POSIX regex to match. +endif::git-format-patch[] -O:: Output the patch in the order specified in the , which has one shell glob pattern per line. +ifndef::git-format-patch[] -R:: Swap two inputs; that is, show differences from index or on-disk file to tree contents. @@ -205,6 +243,7 @@ override configuration settings. not in a subdirectory (e.g. in a bare repository), you can name which subdirectory to make the output relative to by giving a as an argument. +endif::git-format-patch[] -a:: --text:: @@ -229,13 +268,15 @@ override configuration settings. Show the context between diff hunks, up to the specified number of lines, thereby fusing hunks that are close to each other. +ifndef::git-format-patch[] --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. + Disable all output of the program. Implies `--exit-code`. +endif::git-format-patch[] --ext-diff:: Allow an external diff helper to be executed. If you set an diff --git a/Documentation/git-add.txt b/Documentation/git-add.txt index 45ebf87ca..1f1b19996 100644 --- a/Documentation/git-add.txt +++ b/Documentation/git-add.txt @@ -14,28 +14,32 @@ SYNOPSIS DESCRIPTION ----------- -This command adds the current content of new or modified files to the -index, thus staging that content for inclusion in the next commit. +This command updates the index using the current content found in +the working tree, to prepare the content staged for the next commit. +It typically adds the current content of existing paths as a whole, +but with some options it can also be used to add content with +only part of the changes made to the working tree files applied, or +remove paths that do not exist in the working tree anymore. The "index" holds a snapshot of the content of the working tree, and it is this snapshot that is taken as the contents of the next commit. Thus after making any changes to the working directory, and before running -the commit command, you must use the 'add' command to add any new or +the commit command, you must use the `add` command to add any new or modified files to the index. This command can be performed multiple times before a commit. It only adds the content of the specified file(s) at the time the add command is run; if you want subsequent changes included in the next commit, then -you must run 'git add' again to add the new content to the index. +you must run `git add` again to add the new content to the index. -The 'git status' command can be used to obtain a summary of which +The `git status` command can be used to obtain a summary of which files have changes that are staged for the next commit. -The 'git add' command will not add ignored files by default. If any -ignored files were explicitly specified on the command line, 'git add' +The `git add` command will not add ignored files by default. If any +ignored files were explicitly specified on the command line, `git add` will fail with a list of ignored files. Ignored files reached by directory recursion or filename globbing performed by Git (quote your -globs before the shell) will be silently ignored. The 'add' command can +globs before the shell) will be silently ignored. The `add` command can be used to add ignored files with the `-f` (force) option. Please see linkgit:git-commit[1] for alternative ways to add content to a @@ -76,10 +80,10 @@ OPTIONS work tree and add them to the index. This gives the user a chance to review the difference before adding modified contents to the index. - - This effectively runs ``add --interactive``, but bypasses the - initial command menu and directly jumps to `patch` subcommand. - See ``Interactive mode'' for details. ++ +This effectively runs `add --interactive`, but bypasses the +initial command menu and directly jumps to the `patch` subcommand. +See ``Interactive mode'' for details. -e, \--edit:: Open the diff vs. the index in an editor and let the user @@ -92,28 +96,31 @@ apply. -u:: --update:: - Update only files that git already knows about, staging modified - content for commit and marking deleted files for removal. This - is similar - to what "git commit -a" does in preparation for making a commit, - except that the update is limited to paths specified on the - command line. If no paths are specified, all tracked files in the - current directory and its subdirectories are updated. + Only match against already tracked files in + the index rather than the working tree. That means that it + will never stage new files, but that it will stage modified + new contents of tracked files and that it will remove files + from the index if the corresponding files in the working tree + have been removed. ++ +If no is given, default to "."; in other words, +update all tracked files in the current directory and its +subdirectories. -A:: --all:: - Update files that git already knows about (same as '\--update') - and add all untracked files that are not ignored by '.gitignore' - mechanism. - + Like `-u`, but match against files in the + working tree in addition to the index. That means that it + will find new files as well as staging modified content and + removing files that are no longer in the working tree. -N:: --intent-to-add:: Record only the fact that the path will be added later. An entry for the path is placed in the index with no content. This is useful for, among other things, showing the unstaged content of - such files with 'git diff' and committing them with 'git commit - -a'. + such files with `git diff` and committing them with `git commit + -a`. --refresh:: Don't add the file(s), but only refresh their stat() @@ -133,7 +140,7 @@ apply. Configuration ------------- -The optional configuration variable 'core.excludesfile' indicates a path to a +The optional configuration variable `core.excludesfile` indicates a path to a file containing patterns of file names to exclude from git-add, similar to $GIT_DIR/info/exclude. Patterns in the exclude file are used in addition to those in info/exclude. See linkgit:gitrepository-layout[5]. @@ -181,7 +188,7 @@ and type return, like this: What now> 1 ------------ -You also could say "s" or "sta" or "status" above as long as the +You also could say `s` or `sta` or `status` above as long as the choice is unique. The main command loop has 6 subcommands (plus help and quit). @@ -189,9 +196,9 @@ The main command loop has 6 subcommands (plus help and quit). status:: This shows the change between HEAD and index (i.e. what will be - committed if you say "git commit"), and between index and + committed if you say `git commit`), and between index and working tree files (i.e. what you could stage further before - "git commit" using "git-add") for each path. A sample output + `git commit` using `git add`) for each path. A sample output looks like this: + ------------ diff --git a/Documentation/git-apply.txt b/Documentation/git-apply.txt index 5ee8c91f2..c2528a765 100644 --- a/Documentation/git-apply.txt +++ b/Documentation/git-apply.txt @@ -3,7 +3,7 @@ git-apply(1) NAME ---- -git-apply - Apply a patch on a git index file and/or a working tree +git-apply - Apply a patch to files and/or to the index SYNOPSIS @@ -20,8 +20,11 @@ SYNOPSIS DESCRIPTION ----------- -Reads supplied 'diff' output and applies it on a git index file -and a work tree. +Reads the supplied diff output (i.e. "a patch") and applies it to files. +With the `--index` option the patch is also applied to the index, and +with the `--cache` option the patch is only applied to the index. +Without these options, the command applies the patch only to files, +and does not require them to be in a git repository. OPTIONS ------- @@ -34,7 +37,7 @@ OPTIONS input. Turns off "apply". --numstat:: - Similar to \--stat, but shows the number of added and + Similar to `--stat`, but shows the number of added and deleted lines in decimal notation and the pathname without abbreviation, to make it more machine friendly. For binary files, outputs two `-` instead of saying @@ -48,22 +51,22 @@ OPTIONS --check:: Instead of applying the patch, see if the patch is - applicable to the current work tree and/or the index + applicable to the current working tree and/or the index file and detects errors. Turns off "apply". --index:: - When --check is in effect, or when applying the patch + When `--check` is in effect, or when applying the patch (which is the default when none of the options that disables it is in effect), make sure the patch is applicable to what the current index file records. If - the file to be patched in the work tree is not + the file to be patched in the working tree is not up-to-date, it is flagged as an error. This flag also causes the index file to be updated. --cached:: Apply a patch without touching the working tree. Instead take the cached data, apply the patch, and store the result in the index - without using the working tree. This implies '--index'. + without using the working tree. This implies `--index`. --build-fake-ancestor=:: Newer 'git-diff' output has embedded 'index information' @@ -87,11 +90,13 @@ the information is read from the current index instead. rejected hunks in corresponding *.rej files. -z:: - When showing the index information, do not munge paths, - but use NUL terminated machine readable format. Without - this flag, the pathnames output will have TAB, LF, and - backslash characters replaced with `\t`, `\n`, and `\\`, - respectively. + When `--numstat` has been given, do not munge pathnames, + but use a NUL-terminated machine-readable format. ++ +Without this option, each pathname output will have TAB, LF, double quotes, +and backslash characters replaced with `\t`, `\n`, `\"`, and `\\`, +respectively, and the pathname will be enclosed in double quotes if +any of those replacements occurred. -p:: Remove leading slashes from traditional diff paths. The @@ -107,8 +112,8 @@ the information is read from the current index instead. By default, 'git-apply' expects that the patch being applied is a unified diff with at least one line of context. This provides good safety measures, but breaks down when - applying a diff generated with --unified=0. To bypass these - checks use '--unidiff-zero'. + applying a diff generated with `--unified=0`. To bypass these + checks use `--unidiff-zero`. + Note, for the reasons stated above usage of context-free patches is discouraged. @@ -144,7 +149,7 @@ discouraged. be useful when importing patchsets, where you want to include certain files or directories. + -When --exclude and --include patterns are used, they are examined in the +When `--exclude` and `--include` patterns are used, they are examined in the order they appear on the command line, and the first match determines if a patch to each path is used. A patch to a path that does not match any include/exclude pattern is used by default if there is no include pattern @@ -227,13 +232,13 @@ Submodules If the patch contains any changes to submodules then 'git-apply' treats these changes as follows. -If --index is specified (explicitly or implicitly), then the submodule +If `--index` is specified (explicitly or implicitly), then the submodule commits must match the index exactly for the patch to apply. If any of the submodules are checked-out, then these check-outs are completely ignored, i.e., they are not required to be up-to-date or clean and they are not updated. -If --index is not specified, then the submodule commits in the patch +If `--index` is not specified, then the submodule commits in the patch are ignored and only the absence or presence of the corresponding subdirectory is checked and (if possible) updated. diff --git a/Documentation/git-archive.txt b/Documentation/git-archive.txt index 3d1c1e75b..e57979198 100644 --- a/Documentation/git-archive.txt +++ b/Documentation/git-archive.txt @@ -74,8 +74,9 @@ OPTIONS The tree or commit to produce an archive for. path:: - If one or more paths are specified, include only these in the - archive, otherwise include all files and subdirectories. + Without an optional path parameter, all files and subdirectories + of the current working directory are included in the archive. + If one or more paths are specified, only these are included. BACKEND EXTRA OPTIONS --------------------- diff --git a/Documentation/git-bisect-lk2009.txt b/Documentation/git-bisect-lk2009.txt new file mode 100644 index 000000000..6b7b2e549 --- /dev/null +++ b/Documentation/git-bisect-lk2009.txt @@ -0,0 +1,1358 @@ +Fighting regressions with git bisect +==================================== +:Author: Christian Couder +:Email: chriscool@tuxfamily.org +:Date: 2009/11/08 + +Abstract +-------- + +"git bisect" enables software users and developers to easily find the +commit that introduced a regression. We show why it is important to +have good tools to fight regressions. We describe how "git bisect" +works from the outside and the algorithms it uses inside. Then we +explain how to take advantage of "git bisect" to improve current +practices. And we discuss how "git bisect" could improve in the +future. + + +Introduction to "git bisect" +---------------------------- + +Git is a Distributed Version Control system (DVCS) created by Linus +Torvalds and maintained by Junio Hamano. + +In Git like in many other Version Control Systems (VCS), the different +states of the data that is managed by the system are called +commits. And, as VCS are mostly used to manage software source code, +sometimes "interesting" changes of behavior in the software are +introduced in some commits. + +In fact people are specially interested in commits that introduce a +"bad" behavior, called a bug or a regression. They are interested in +these commits because a commit (hopefully) contains a very small set +of source code changes. And it's much easier to understand and +properly fix a problem when you only need to check a very small set of +changes, than when you don't know where look in the first place. + +So to help people find commits that introduce a "bad" behavior, the +"git bisect" set of commands was invented. And it follows of course +that in "git bisect" parlance, commits where the "interesting +behavior" is present are called "bad" commits, while other commits are +called "good" commits. And a commit that introduce the behavior we are +interested in is called a "first bad commit". Note that there could be +more than one "first bad commit" in the commit space we are searching. + +So "git bisect" is designed to help find a "first bad commit". And to +be as efficient as possible, it tries to perform a binary search. + + +Fighting regressions overview +----------------------------- + +Regressions: a big problem +~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Regressions are a big problem in the software industry. But it's +difficult to put some real numbers behind that claim. + +There are some numbers about bugs in general, like a NIST study in +2002 <<1>> that said: + +_____________ +Software bugs, or errors, are so prevalent and so detrimental that +they cost the U.S. economy an estimated $59.5 billion annually, or +about 0.6 percent of the gross domestic product, according to a newly +released study commissioned by the Department of Commerce's National +Institute of Standards and Technology (NIST). At the national level, +over half of the costs are borne by software users and the remainder +by software developers/vendors. The study also found that, although +all errors cannot be removed, more than a third of these costs, or an +estimated $22.2 billion, could be eliminated by an improved testing +infrastructure that enables earlier and more effective identification +and removal of software defects. These are the savings associated with +finding an increased percentage (but not 100 percent) of errors closer +to the development stages in which they are introduced. Currently, +over half of all errors are not found until "downstream" in the +development process or during post-sale software use. +_____________ + +And then: + +_____________ +Software developers already spend approximately 80 percent of +development costs on identifying and correcting defects, and yet few +products of any type other than software are shipped with such high +levels of errors. +_____________ + +Eventually the conclusion started with: + +_____________ +The path to higher software quality is significantly improved software +testing. +_____________ + +There are other estimates saying that 80% of the cost related to +software is about maintenance <<2>>. + +Though, according to Wikipedia <<3>>: + +_____________ +A common perception of maintenance is that it is merely fixing +bugs. However, studies and surveys over the years have indicated that +the majority, over 80%, of the maintenance effort is used for +non-corrective actions (Pigosky 1997). This perception is perpetuated +by users submitting problem reports that in reality are functionality +enhancements to the system. +_____________ + +But we can guess that improving on existing software is very costly +because you have to watch out for regressions. At least this would +make the above studies consistent among themselves. + +Of course some kind of software is developed, then used during some +time without being improved on much, and then finally thrown away. In +this case, of course, regressions may not be a big problem. But on the +other hand, there is a lot of big software that is continually +developed and maintained during years or even tens of years by a lot +of people. And as there are often many people who depend (sometimes +critically) on such software, regressions are a really big problem. + +One such software is the linux kernel. And if we look at the linux +kernel, we can see that a lot of time and effort is spent to fight +regressions. The release cycle start with a 2 weeks long merge +window. Then the first release candidate (rc) version is tagged. And +after that about 7 or 8 more rc versions will appear with around one +week between each of them, before the final release. + +The time between the first rc release and the final release is +supposed to be used to test rc versions and fight bugs and especially +regressions. And this time is more than 80% of the release cycle +time. But this is not the end of the fight yet, as of course it +continues after the release. + +And then this is what Ingo Molnar (a well known linux kernel +developer) says about his use of git bisect: + +_____________ +I most actively use it during the merge window (when a lot of trees +get merged upstream and when the influx of bugs is the highest) - and +yes, there have been cases that i used it multiple times a day. My +average is roughly once a day. +_____________ + +So regressions are fought all the time by developers, and indeed it is +well known that bugs should be fixed as soon as possible, so as soon +as they are found. That's why it is interesting to have good tools for +this purpose. + +Other tools to fight regressions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +So what are the tools used to fight regressions? They are nearly the +same as those used to fight regular bugs. The only specific tools are +test suites and tools similar as "git bisect". + +Test suites are very nice. But when they are used alone, they are +supposed to be used so that all the tests are checked after each +commit. This means that they are not very efficient, because many +tests are run for no interesting result, and they suffer from +combinational explosion. + +In fact the problem is that big software often has many different +configuration options and that each test case should pass for each +configuration after each commit. So if you have for each release: N +configurations, M commits and T test cases, you should perform: + +------------- +N * M * T tests +------------- + +where N, M and T are all growing with the size your software. + +So very soon it will not be possible to completely test everything. + +And if some bugs slip through your test suite, then you can add a test +to your test suite. But if you want to use your new improved test +suite to find where the bug slipped in, then you will either have to +emulate a bisection process or you will perhaps bluntly test each +commit backward starting from the "bad" commit you have which may be +very wasteful. + +"git bisect" overview +--------------------- + +Starting a bisection +~~~~~~~~~~~~~~~~~~~~ + +The first "git bisect" subcommand to use is "git bisect start" to +start the search. Then bounds must be set to limit the commit +space. This is done usually by giving one "bad" and at least one +"good" commit. They can be passed in the initial call to "git bisect +start" like this: + +------------- +$ git bisect start [BAD [GOOD...]] +------------- + +or they can be set using: + +------------- +$ git bisect bad [COMMIT] +------------- + +and: + +------------- +$ git bisect good [COMMIT...] +------------- + +where BAD, GOOD and COMMIT are all names that can be resolved to a +commit. + +Then "git bisect" will checkout a commit of its choosing and ask the +user to test it, like this: + +------------- +$ git bisect start v2.6.27 v2.6.25 +Bisecting: 10928 revisions left to test after this (roughly 14 steps) +[2ec65f8b89ea003c27ff7723525a2ee335a2b393] x86: clean up using max_low_pfn on 32-bit +------------- + +Note that the example that we will use is really a toy example, we +will be looking for the first commit that has a version like +"2.6.26-something", that is the commit that has a "SUBLEVEL = 26" line +in the top level Makefile. This is a toy example because there are +better ways to find this commit with git than using "git bisect" (for +example "git blame" or "git log -S"). + +Driving a bisection manually +~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +At this point there are basically 2 ways to drive the search. It can +be driven manually by the user or it can be driven automatically by a +script or a command. + +If the user is driving it, then at each step of the search, the user +will have to test the current commit and say if it is "good" or "bad" +using the "git bisect good" or "git bisect bad" commands respectively +that have been described above. For example: + +------------- +$ git bisect bad +Bisecting: 5480 revisions left to test after this (roughly 13 steps) +[66c0b394f08fd89236515c1c84485ea712a157be] KVM: kill file->f_count abuse in kvm +------------- + +And after a few more steps like that, "git bisect" will eventually +find a first bad commit: + +------------- +$ git bisect bad +2ddcca36c8bcfa251724fe342c8327451988be0d is the first bad commit +commit 2ddcca36c8bcfa251724fe342c8327451988be0d +Author: Linus Torvalds +Date: Sat May 3 11:59:44 2008 -0700 + + Linux 2.6.26-rc1 + +:100644 100644 5cf8258195331a4dbdddff08b8d68642638eea57 4492984efc09ab72ff6219a7bc21fb6a957c4cd5 M Makefile +------------- + +At this point we can see what the commit does, check it out (if it's +not already checked out) or tinker with it, for example: + +------------- +$ git show HEAD +commit 2ddcca36c8bcfa251724fe342c8327451988be0d +Author: Linus Torvalds +Date: Sat May 3 11:59:44 2008 -0700 + + Linux 2.6.26-rc1 + +diff --git a/Makefile b/Makefile +index 5cf8258..4492984 100644 +--- a/Makefile ++++ b/Makefile +@@ -1,7 +1,7 @@ + VERSION = 2 + PATCHLEVEL = 6 +-SUBLEVEL = 25 +-EXTRAVERSION = ++SUBLEVEL = 26 ++EXTRAVERSION = -rc1 + NAME = Funky Weasel is Jiggy wit it + + # *DOCUMENTATION* +------------- + +And when we are finished we can use "git bisect reset" to go back to +the branch we were in before we started bisecting: + +------------- +$ git bisect reset +Checking out files: 100% (21549/21549), done. +Previous HEAD position was 2ddcca3... Linux 2.6.26-rc1 +Switched to branch 'master' +------------- + +Driving a bisection automatically +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +The other way to drive the bisection process is to tell "git bisect" +to launch a script or command at each bisection step to know if the +current commit is "good" or "bad". To do that, we use the "git bisect +run" command. For example: + +------------- +$ git bisect start v2.6.27 v2.6.25 +Bisecting: 10928 revisions left to test after this (roughly 14 steps) +[2ec65f8b89ea003c27ff7723525a2ee335a2b393] x86: clean up using max_low_pfn on 32-bit +$ +$ git bisect run grep '^SUBLEVEL = 25' Makefile +running grep ^SUBLEVEL = 25 Makefile +Bisecting: 5480 revisions left to test after this (roughly 13 steps) +[66c0b394f08fd89236515c1c84485ea712a157be] KVM: kill file->f_count abuse in kvm +running grep ^SUBLEVEL = 25 Makefile +SUBLEVEL = 25 +Bisecting: 2740 revisions left to test after this (roughly 12 steps) +[671294719628f1671faefd4882764886f8ad08cb] V4L/DVB(7879): Adding cx18 Support for mxl5005s +... +... +running grep ^SUBLEVEL = 25 Makefile +Bisecting: 0 revisions left to test after this (roughly 0 steps) +[2ddcca36c8bcfa251724fe342c8327451988be0d] Linux 2.6.26-rc1 +running grep ^SUBLEVEL = 25 Makefile +2ddcca36c8bcfa251724fe342c8327451988be0d is the first bad commit +commit 2ddcca36c8bcfa251724fe342c8327451988be0d +Author: Linus Torvalds +Date: Sat May 3 11:59:44 2008 -0700 + + Linux 2.6.26-rc1 + +:100644 100644 5cf8258195331a4dbdddff08b8d68642638eea57 4492984efc09ab72ff6219a7bc21fb6a957c4cd5 M Makefile +bisect run success +------------- + +In this example, we passed "grep '^SUBLEVEL = 25' Makefile" as +parameter to "git bisect run". This means that at each step, the grep +command we passed will be launched. And if it exits with code 0 (that +means success) then git bisect will mark the current state as +"good". If it exits with code 1 (or any code between 1 and 127 +included, except the special code 125), then the current state will be +marked as "bad". + +Exit code between 128 and 255 are special to "git bisect run". They +make it stop immediately the bisection process. This is useful for +example if the command passed takes too long to complete, because you +can kill it with a signal and it will stop the bisection process. + +It can also be useful in scripts passed to "git bisect run" to "exit +255" if some very abnormal situation is detected. + +Avoiding untestable commits +~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Sometimes it happens that the current state cannot be tested, for +example if it does not compile because there was a bug preventing it +at that time. This is what the special exit code 125 is for. It tells +"git bisect run" that the current commit should be marked as +untestable and that another one should be chosen and checked out. + +If the bisection process is driven manually, you can use "git bisect +skip" to do the same thing. (In fact the special exit code 125 makes +"git bisect run" use "git bisect skip" in the background.) + +Or if you want more control, you can inspect the current state using +for example "git bisect visualize". It will launch gitk (or "git log" +if the DISPLAY environment variable is not set) to help you find a +better bisection point. + +Either way, if you have a string of untestable commits, it might +happen that the regression you are looking for has been introduced by +one of these untestable commits. In this case it's not possible to +tell for sure which commit introduced the regression. + +So if you used "git bisect skip" (or the run script exited with +special code 125) you could get a result like this: + +------------- +There are only 'skip'ped commits left to test. +The first bad commit could be any of: +15722f2fa328eaba97022898a305ffc8172db6b1 +78e86cf3e850bd755bb71831f42e200626fbd1e0 +e15b73ad3db9b48d7d1ade32f8cd23a751fe0ace +070eab2303024706f2924822bfec8b9847e4ac1b +We cannot bisect more! +------------- + +Saving a log and replaying it +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +If you want to show other people your bisection process, you can get a +log using for example: + +------------- +$ git bisect log > bisect_log.txt +------------- + +And it is possible to replay it using: + +------------- +$ git bisect replay bisect_log.txt +------------- + + +"git bisect" details +-------------------- + +Bisection algorithm +~~~~~~~~~~~~~~~~~~~ + +As the Git commits form a directed acyclic graph (DAG), finding the +best bisection commit to test at each step is not so simple. Anyway +Linus found and implemented a "truly stupid" algorithm, later improved +by Junio Hamano, that works quite well. + +So the algorithm used by "git bisect" to find the best bisection +commit when there are no skipped commits is the following: + +1) keep only the commits that: + +a) are ancestor of the "bad" commit (including the "bad" commit itself), +b) are not ancestor of a "good" commit (excluding the "good" commits). + +This means that we get rid of the uninteresting commits in the DAG. + +For example if we start with a graph like this: + +------------- +G-Y-G-W-W-W-X-X-X-X + \ / + W-W-B + / +Y---G-W---W + \ / \ +Y-Y X-X-X-X + +-> time goes this way -> +------------- + +where B is the "bad" commit, "G" are "good" commits and W, X, and Y +are other commits, we will get the following graph after this first +step: + +------------- +W-W-W + \ + W-W-B + / +W---W +------------- + +So only the W and B commits will be kept. Because commits X and Y will +have been removed by rules a) and b) respectively, and because commits +G are removed by rule b) too. + +Note for git users, that it is equivalent as keeping only the commit +given by: + +------------- +git rev-list BAD --not GOOD1 GOOD2... +------------- + +Also note that we don't require the commits that are kept to be +descendants of a "good" commit. So in the following example, commits W +and Z will be kept: + +------------- +G-W-W-W-B + / +Z-Z +------------- + +2) starting from the "good" ends of the graph, associate to each +commit the number of ancestors it has plus one + +For example with the following graph where H is the "bad" commit and A +and D are some parents of some "good" commits: + +------------- +A-B-C + \ + F-G-H + / +D---E +------------- + +this will give: + +------------- +1 2 3 +A-B-C + \6 7 8 + F-G-H +1 2/ +D---E +------------- + +3) associate to each commit: min(X, N - X) + +where X is the value associated to the commit in step 2) and N is the +total number of commits in the graph. + +In the above example we have N = 8, so this will give: + +------------- +1 2 3 +A-B-C + \2 1 0 + F-G-H +1 2/ +D---E +------------- + +4) the best bisection point is the commit with the highest associated +number + +So in the above example the best bisection point is commit C. + +5) note that some shortcuts are implemented to speed up the algorithm + +As we know N from the beginning, we know that min(X, N - X) can't be +greater than N/2. So during steps 2) and 3), if we would associate N/2 +to a commit, then we know this is the best bisection point. So in this +case we can just stop processing any other commit and return the +current commit. + +Bisection algorithm debugging +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +For any commit graph, you can see the number associated with each +commit using "git rev-list --bisect-all". + +For example, for the above graph, a command like: + +------------- +$ git rev-list --bisect-all BAD --not GOOD1 GOOD2 +------------- + +would output something like: + +------------- +e15b73ad3db9b48d7d1ade32f8cd23a751fe0ace (dist=3) +15722f2fa328eaba97022898a305ffc8172db6b1 (dist=2) +78e86cf3e850bd755bb71831f42e200626fbd1e0 (dist=2) +a1939d9a142de972094af4dde9a544e577ddef0e (dist=2) +070eab2303024706f2924822bfec8b9847e4ac1b (dist=1) +a3864d4f32a3bf5ed177ddef598490a08760b70d (dist=1) +a41baa717dd74f1180abf55e9341bc7a0bb9d556 (dist=1) +9e622a6dad403b71c40979743bb9d5be17b16bd6 (dist=0) +------------- + +Bisection algorithm discussed +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +First let's define "best bisection point". We will say that a commit X +is a best bisection point or a best bisection commit if knowing its +state ("good" or "bad") gives as much information as possible whether +the state of the commit happens to be "good" or "bad". + +This means that the best bisection commits are the commits where the +following function is maximum: + +------------- +f(X) = min(information_if_good(X), information_if_bad(X)) +------------- + +where information_if_good(X) is the information we get if X is good +and information_if_bad(X) is the information we get if X is bad. + +Now we will suppose that there is only one "first bad commit". This +means that all its descendants are "bad" and all the other commits are +"good". And we will suppose that all commits have an equal probability +of being good or bad, or of being the first bad commit, so knowing the +state of c commits gives always the same amount of information +wherever these c commits are on the graph and whatever c is. (So we +suppose that these commits being for example on a branch or near a +good or a bad commit does not give more or less information). + +Let's also suppose that we have a cleaned up graph like one after step +1) in the bisection algorithm above. This means that we can measure +the information we get in terms of number of commit we can remove from +the graph.. + +And let's take a commit X in the graph. + +If X is found to be "good", then we know that its ancestors are all +"good", so we want to say that: + +------------- +information_if_good(X) = number_of_ancestors(X) (TRUE) +------------- + +And this is true because at step 1) b) we remove the ancestors of the +"good" commits. + +If X is found to be "bad", then we know that its descendants are all +"bad", so we want to say that: + +------------- +information_if_bad(X) = number_of_descendants(X) (WRONG) +------------- + +But this is wrong because at step 1) a) we keep only the ancestors of +the bad commit. So we get more information when a commit is marked as +"bad", because we also know that the ancestors of the previous "bad" +commit that are not ancestors of the new "bad" commit are not the +first bad commit. We don't know if they are good or bad, but we know +that they are not the first bad commit because they are not ancestor +of the new "bad" commit. + +So when a commit is marked as "bad" we know we can remove all the +commits in the graph except those that are ancestors of the new "bad" +commit. This means that: + +------------- +information_if_bad(X) = N - number_of_ancestors(X) (TRUE) +------------- + +where N is the number of commits in the (cleaned up) graph. + +So in the end this means that to find the best bisection commits we +should maximize the function: + +------------- +f(X) = min(number_of_ancestors(X), N - number_of_ancestors(X)) +------------- + +And this is nice because at step 2) we compute number_of_ancestors(X) +and so at step 3) we compute f(X). + +Let's take the following graph as an example: + +------------- + G-H-I-J + / \ +A-B-C-D-E-F O + \ / + K-L-M-N +------------- + +If we compute the following non optimal function on it: + +------------- +g(X) = min(number_of_ancestors(X), number_of_descendants(X)) +------------- + +we get: + +------------- + 4 3 2 1 + G-H-I-J +1 2 3 4 5 6/ \0 +A-B-C-D-E-F O + \ / + K-L-M-N + 4 3 2 1 +------------- + +but with the algorithm used by git bisect we get: + +------------- + 7 7 6 5 + G-H-I-J +1 2 3 4 5 6/ \0 +A-B-C-D-E-F O + \ / + K-L-M-N + 7 7 6 5 +------------- + +So we chose G, H, K or L as the best bisection point, which is better +than F. Because if for example L is bad, then we will know not only +that L, M and N are bad but also that G, H, I and J are not the first +bad commit (since we suppose that there is only one first bad commit +and it must be an ancestor of L). + +So the current algorithm seems to be the best possible given what we +initially supposed. + +Skip algorithm +~~~~~~~~~~~~~~ + +When some commits have been skipped (using "git bisect skip"), then +the bisection algorithm is the same for step 1) to 3). But then we use +roughly the following steps: + +6) sort the commit by decreasing associated value + +7) if the first commit has not been skipped, we can return it and stop +here + +8) otherwise filter out all the skipped commits in the sorted list + +9) use a pseudo random number generator (PRNG) to generate a random +number between 0 and 1 + +10) multiply this random number with its square root to bias it toward +0 + +11) multiply the result by the number of commits in the filtered list +to get an index into this list + +12) return the commit at the computed index + +Skip algorithm discussed +~~~~~~~~~~~~~~~~~~~~~~~~ + +After step 7) (in the skip algorithm), we could check if the second +commit has been skipped and return it if it is not the case. And in +fact that was the algorithm we used from when "git bisect skip" was +developed in git version 1.5.4 (released on February 1st 2008) until +git version 1.6.4 (released July 29th 2009). + +But Ingo Molnar and H. Peter Anvin (another well known linux kernel +developer) both complained that sometimes the best bisection points +all happened to be in an area where all the commits are +untestable. And in this case the user was asked to test many +untestable commits, which could be very inefficient. + +Indeed untestable commits are often untestable because a breakage was +introduced at one time, and that breakage was fixed only after many +other commits were introduced. + +This breakage is of course most of the time unrelated to the breakage +we are trying to locate in the commit graph. But it prevents us to +know if the interesting "bad behavior" is present or not. + +So it is a fact that commits near an untestable commit have a high +probability of being untestable themselves. And the best bisection +commits are often found together too (due to the bisection algorithm). + +This is why it is a bad idea to just chose the next best unskipped +bisection commit when the first one has been skipped. + +We found that most commits on the graph may give quite a lot of +information when they are tested. And the commits that will not on +average give a lot of information are the one near the good and bad +commits. + +So using a PRNG with a bias to favor commits away from the good and +bad commits looked like a good choice. + +One obvious improvement to this algorithm would be to look for a +commit that has an associated value near the one of the best bisection +commit, and that is on another branch, before using the PRNG. Because +if such a commit exists, then it is not very likely to be untestable +too, so it will probably give more information than a nearly randomly +chosen one. + +Checking merge bases +~~~~~~~~~~~~~~~~~~~~ + +There is another tweak in the bisection algorithm that has not been +described in the "bisection algorithm" above. + +We supposed in the previous examples that the "good" commits were +ancestors of the "bad" commit. But this is not a requirement of "git +bisect". + +Of course the "bad" commit cannot be an ancestor of a "good" commit, +because the ancestors of the good commits are supposed to be +"good". And all the "good" commits must be related to the bad commit. +They cannot be on a branch that has no link with the branch of the +"bad" commit. But it is possible for a good commit to be related to a +bad commit and yet not be neither one of its ancestor nor one of its +descendants. + +For example, there can be a "main" branch, and a "dev" branch that was +forked of the main branch at a commit named "D" like this: + +------------- +A-B-C-D-E-F-G <--main + \ + H-I-J <--dev +------------- + +The commit "D" is called a "merge base" for branch "main" and "dev" +because it's the best common ancestor for these branches for a merge. + +Now let's suppose that commit J is bad and commit G is good and that +we apply the bisection algorithm like it has been previously +described. + +As described in step 1) b) of the bisection algorithm, we remove all +the ancestors of the good commits because they are supposed to be good +too. + +So we would be left with only: + +------------- +H-I-J +------------- + +But what happens if the first bad commit is "B" and if it has been +fixed in the "main" branch by commit "F"? + +The result of such a bisection would be that we would find that H is +the first bad commit, when in fact it's B. So that would be wrong! + +And yes it's can happen in practice that people working on one branch +are not aware that people working on another branch fixed a bug! It +could also happen that F fixed more than one bug or that it is a +revert of some big development effort that was not ready to be +released. + +In fact development teams often maintain both a development branch and +a maintenance branch, and it would be quite easy for them if "git +bisect" just worked when they want to bisect a regression on the +development branch that is not on the maintenance branch. They should +be able to start bisecting using: + +------------- +$ git bisect start dev main +------------- + +To enable that additional nice feature, when a bisection is started +and when some good commits are not ancestors of the bad commit, we +first compute the merge bases between the bad and the good commits and +we chose these merge bases as the first commits that will be checked +out and tested. + +If it happens that one merge base is bad, then the bisection process +is stopped with a message like: + +------------- +The merge base BBBBBB is bad. +This means the bug has been fixed between BBBBBB and [GGGGGG,...]. +------------- + +where BBBBBB is the sha1 hash of the bad merge base and [GGGGGG,...] +is a comma separated list of the sha1 of the good commits. + +If some of the merge bases are skipped, then the bisection process +continues, but the following message is printed for each skipped merge +base: + +------------- +Warning: the merge base between BBBBBB and [GGGGGG,...] must be skipped. +So we cannot be sure the first bad commit is between MMMMMM and BBBBBB. +We continue anyway. +------------- + +where BBBBBB is the sha1 hash of the bad commit, MMMMMM is the sha1 +hash of the merge base that is skipped and [GGGGGG,...] is a comma +separated list of the sha1 of the good commits. + +So if there is no bad merge base, the bisection process continues as +usual after this step. + +Best bisecting practices +------------------------ + +Using test suites and git bisect together +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +If you both have a test suite and use git bisect, then it becomes less +important to check that all tests pass after each commit. Though of +course it is probably a good idea to have some checks to avoid +breaking too many things because it could make bisecting other bugs +more difficult. + +You can focus your efforts to check at a few points (for example rc +and beta releases) that all the T test cases pass for all the N +configurations. And when some tests don't pass you can use "git +bisect" (or better "git bisect run"). So you should perform roughly: + +------------- +c * N * T + b * M * log2(M) tests +------------- + +where c is the number of rounds of test (so a small constant) and b is +the ratio of bug per commit (hopefully a small constant too). + +So of course it's much better as it's O(N \* T) vs O(N \* T \* M) if +you would test everything after each commit. + +This means that test suites are good to prevent some bugs from being +committed and they are also quite good to tell you that you have some +bugs. But they are not so good to tell you where some bugs have been +introduced. To tell you that efficiently, git bisect is needed. + +The other nice thing with test suites, is that when you have one, you +already know how to test for bad behavior. So you can use this +knowledge to create a new test case for "git bisect" when it appears +that there is a regression. So it will be easier to bisect the bug and +fix it. And then you can add the test case you just created to your +test suite. + +So if you know how to create test cases and how to bisect, you will be +subject to a virtuous circle: + +more tests => easier to create tests => easier to bisect => more tests + +So test suites and "git bisect" are complementary tools that are very +powerful and efficient when used together. + +Bisecting build failures +~~~~~~~~~~~~~~~~~~~~~~~~ + +You can very easily automatically bisect broken builds using something +like: + +------------- +$ git bisect start BAD GOOD +$ git bisect run make +------------- + +Passing sh -c "some commands" to "git bisect run" +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +For example: + +------------- +$ git bisect run sh -c "make || exit 125; ./my_app | grep 'good output'" +------------- + +On the other hand if you do this often, then it can be worth having +scripts to avoid too much typing. + +Finding performance regressions +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Here is an example script that comes slightly modified from a real +world script used by Junio Hamano <<4>>. + +This script can be passed to "git bisect run" to find the commit that +introduced a performance regression: + +------------- +#!/bin/sh + +# Build errors are not what I am interested in. +make my_app || exit 255 + +# We are checking if it stops in a reasonable amount of time, so +# let it run in the background... + +./my_app >log 2>&1 & + +# ... and grab its process ID. +pid=$! + +# ... and then wait for sufficiently long. +sleep $NORMAL_TIME + +# ... and then see if the process is still there. +if kill -0 $pid +then + # It is still running -- that is bad. + kill $pid; sleep 1; kill $pid; + exit 1 +else + # It has already finished (the $pid process was no more), + # and we are happy. + exit 0 +fi +------------- + +Following general best practices +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +It is obviously a good idea not to have commits with changes that +knowingly break things, even if some other commits later fix the +breakage. + +It is also a good idea when using any VCS to have only one small +logical change in each commit. + +The smaller the changes in your commit, the most effective "git +bisect" will be. And you will probably need "git bisect" less in the +first place, as small changes are easier to review even if they are +only reviewed by the commiter. + +Another good idea is to have good commit messages. They can be very +helpful to understand why some changes were made. + +These general best practices are very helpful if you bisect often. + +Avoiding bug prone merges +~~~~~~~~~~~~~~~~~~~~~~~~~ + +First merges by themselves can introduce some regressions even when +the merge needs no source code conflict resolution. This is because a +semantic change can happen in one branch while the other branch is not +aware of it. + +For example one branch can change the semantic of a function while the +other branch add more calls to the same function. + +This is made much worse if many files have to be fixed to resolve +conflicts. That's why such merges are called "evil merges". They can +make regressions very difficult to track down. It can even be +misleading to know the first bad commit if it happens to be such a +merge, because people might think that the bug comes from bad conflict +resolution when it comes from a semantic change in one branch. + +Anyway "git rebase" can be used to linearize history. This can be used +either to avoid merging in the first place. Or it can be used to +bisect on a linear history instead of the non linear one, as this +should give more information in case of a semantic change in one +branch. + +Merges can be also made simpler by using smaller branches or by using +many topic branches instead of only long version related branches. + +And testing can be done more often in special integration branches +like linux-next for the linux kernel. + +Adapting your work-flow +~~~~~~~~~~~~~~~~~~~~~~~ + +A special work-flow to process regressions can give great results. + +Here is an example of a work-flow used by Andreas Ericsson: + +* write, in the test suite, a test script that exposes the regression +* use "git bisect run" to find the commit that introduced it +* fix the bug that is often made obvious by the previous step +* commit both the fix and the test script (and if needed more tests) + +And here is what Andreas said about this work-flow <<5>>: + +_____________ +To give some hard figures, we used to have an average report-to-fix +cycle of 142.6 hours (according to our somewhat weird bug-tracker +which just measures wall-clock time). Since we moved to git, we've +lowered that to 16.2 hours. Primarily because we can stay on top of +the bug fixing now, and because everyone's jockeying to get to fix +bugs (we're quite proud of how lazy we are to let git find the bugs +for us). Each new release results in ~40% fewer bugs (almost certainly +due to how we now feel about writing tests). +_____________ + +Clearly this work-flow uses the virtuous circle between test suites +and "git bisect". In fact it makes it the standard procedure to deal +with regression. + +In other messages Andreas says that they also use the "best practices" +described above: small logical commits, topic branches, no evil +merge,... These practices all improve the bisectability of the commit +graph, by making it easier and more useful to bisect. + +So a good work-flow should be designed around the above points. That +is making bisecting easier, more useful and standard. + +Involving QA people and if possible end users +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +One nice about "git bisect" is that it is not only a developer +tool. It can effectively be used by QA people or even end users (if +they have access to the source code or if they can get access to all +the builds). + +There was a discussion at one point on the linux kernel mailing list +of whether it was ok to always ask end user to bisect, and very good +points were made to support the point of view that it is ok. + +For example David Miller wrote <<6>>: + +_____________ +What people don't get is that this is a situation where the "end node +principle" applies. When you have limited resources (here: developers) +you don't push the bulk of the burden upon them. Instead you push +things out to the resource you have a lot of, the end nodes (here: +users), so that the situation actually scales. +_____________ + +This means that it is often "cheaper" if QA people or end users can do +it. + +What is interesting too is that end users that are reporting bugs (or +QA people that reproduced a bug) have access to the environment where +the bug happens. So they can often more easily reproduce a +regression. And if they can bisect, then more information will be +extracted from the environment where the bug happens, which means that +it will be easier to understand and then fix the bug. + +For open source projects it can be a good way to get more useful +contributions from end users, and to introduce them to QA and +development activities. + +Using complex scripts +~~~~~~~~~~~~~~~~~~~~~ + +In some cases like for kernel development it can be worth developing +complex scripts to be able to fully automate bisecting. + +Here is what Ingo Molnar says about that <<7>>: + +_____________ +i have a fully automated bootup-hang bisection script. It is based on +"git-bisect run". I run the script, it builds and boots kernels fully +automatically, and when the bootup fails (the script notices that via +the serial log, which it continuously watches - or via a timeout, if +the system does not come up within 10 minutes it's a "bad" kernel), +the script raises my attention via a beep and i power cycle the test +box. (yeah, i should make use of a managed power outlet to 100% +automate it) +_____________ + +Combining test suites, git bisect and other systems together +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +We have seen that test suites an git bisect are very powerful when +used together. It can be even more powerful if you can combine them +with other systems. + +For example some test suites could be run automatically at night with +some unusual (or even random) configurations. And if a regression is +found by a test suite, then "git bisect" can be automatically +launched, and its result can be emailed to the author of the first bad +commit found by "git bisect", and perhaps other people too. And a new +entry in the bug tracking system could be automatically created too. + + +The future of bisecting +----------------------- + +"git replace" +~~~~~~~~~~~~~ + +We saw earlier that "git bisect skip" is now using a PRNG to try to +avoid areas in the commit graph where commits are untestable. The +problem is that sometimes the first bad commit will be in an +untestable area. + +To simplify the discussion we will suppose that the untestable area is +a simple string of commits and that it was created by a breakage +introduced by one commit (let's call it BBC for bisect breaking +commit) and later fixed by another one (let's call it BFC for bisect +fixing commit). + +For example: + +------------- +...-Y-BBC-X1-X2-X3-X4-X5-X6-BFC-Z-... +------------- + +where we know that Y is good and BFC is bad, and where BBC and X1 to +X6 are untestable. + +In this case if you are bisecting manually, what you can do is create +a special branch that starts just before the BBC. The first commit in +this branch should be the BBC with the BFC squashed into it. And the +other commits in the branch should be the commits between BBC and BFC +rebased on the first commit of the branch and then the commit after +BFC also rebased on. + +For example: + +------------- + (BBC+BFC)-X1'-X2'-X3'-X4'-X5'-X6'-Z' + / +...-Y-BBC-X1-X2-X3-X4-X5-X6-BFC-Z-... +------------- + +where commits quoted with ' have been rebased. + +You can easily create such a branch with Git using interactive rebase. + +For example using: + +------------- +$ git rebase -i Y Z +------------- + +and then moving BFC after BBC and squashing it. + +After that you can start bisecting as usual in the new branch and you +should eventually find the first bad commit. + +For example: + +------------- +$ git bisect start Z' Y +------------- + +If you are using "git bisect run", you can use the same manual fix up +as above, and then start another "git bisect run" in the special +branch. Or as the "git bisect" man page says, the script passed to +"git bisect run" can apply a patch before it compiles and test the +software <<8>>. The patch should turn a current untestable commits +into a testable one. So the testing will result in "good" or "bad" and +"git bisect" will be able to find the first bad commit. And the script +should not forget to remove the patch once the testing is done before +exiting from the script. + +(Note that instead of a patch you can use "git cherry-pick BFC" to +apply the fix, and in this case you should use "git reset --hard +HEAD^" to revert the cherry-pick after testing and before returning +from the script.) + +But the above ways to work around untestable areas are a little bit +clunky. Using special branches is nice because these branches can be +shared by developers like usual branches, but the risk is that people +will get many such branches. And it disrupts the normal "git bisect" +work-flow. So, if you want to use "git bisect run" completely +automatically, you have to add special code in your script to restart +bisection in the special branches. + +Anyway one can notice in the above special branch example that the Z' +and Z commits should point to the same source code state (the same +"tree" in git parlance). That's because Z' result from applying the +same changes as Z just in a slightly different order. + +So if we could just "replace" Z by Z' when we bisect, then we would +not need to add anything to a script. It would just work for anyone in +the project sharing the special branches and the replacements. + +With the example above that would give: + +------------- + (BBC+BFC)-X1'-X2'-X3'-X4'-X5'-X6'-Z'-... + / +...-Y-BBC-X1-X2-X3-X4-X5-X6-BFC-Z +------------- + +That's why the "git replace" command was created. Technically it +stores replacements "refs" in the "refs/replace/" hierarchy. These +"refs" are like branches (that are stored in "refs/heads/") or tags +(that are stored in "refs/tags"), and that means that they can +automatically be shared like branches or tags among developers. + +"git replace" is a very powerful mechanism. It can be used to fix +commits in already released history, for example to change the commit +message or the author. And it can also be used instead of git "grafts" +to link a repository with another old repository. + +In fact it's this last feature that "sold" it to the git community, so +it is now in the "master" branch of git's git repository and it should +be released in git 1.6.5 in October or November 2009. + +One problem with "git replace" is that currently it stores all the +replacements refs in "refs/replace/", but it would be perhaps better +if the replacement refs that are useful only for bisecting would be in +"refs/replace/bisect/". This way the replacement refs could be used +only for bisecting, while other refs directly in "refs/replace/" would +be used nearly all the time. + +Bisecting sporadic bugs +~~~~~~~~~~~~~~~~~~~~~~~ + +Another possible improvement to "git bisect" would be to optionally +add some redundancy to the tests performed so that it would be more +reliable when tracking sporadic bugs. + +This has been requested by some kernel developers because some bugs +called sporadic bugs do not appear in all the kernel builds because +they are very dependent on the compiler output. + +The idea is that every 3 test for example, "git bisect" could ask the +user to test a commit that has already been found to be "good" or +"bad" (because one of its descendants or one of its ancestors has been +found to be "good" or "bad" respectively). If it happens that a commit +has been previously incorrectly classified then the bisection can be +aborted early, hopefully before too many mistakes have been made. Then +the user will have to look at what happened and then restart the +bisection using a fixed bisect log. + +There is already a project called BBChop created by Ealdwulf Wuffinga +on Github that does something like that using Bayesian Search Theory +<<9>>: + +_____________ +BBChop is like 'git bisect' (or equivalent), but works when your bug +is intermittent. That is, it works in the presence of false negatives +(when a version happens to work this time even though it contains the +bug). It assumes that there are no false positives (in principle, the +same approach would work, but adding it may be non-trivial). +_____________ + +But BBChop is independent of any VCS and it would be easier for Git +users to have something integrated in Git. + +Conclusion +---------- + +We have seen that regressions are an important problem, and that "git +bisect" has nice features that complement very well practices and +other tools, especially test suites, that are generally used to fight +regressions. But it might be needed to change some work-flows and +(bad) habits to get the most out of it. + +Some improvements to the algorithms inside "git bisect" are possible +and some new features could help in some cases, but overall "git +bisect" works already very well, is used a lot, and is already very +useful. To back up that last claim, let's give the final word to Ingo +Molnar when he was asked by the author how much time does he think +"git bisect" saves him when he uses it: + +_____________ +a _lot_. + +About ten years ago did i do my first 'bisection' of a Linux patch +queue. That was prior the Git (and even prior the BitKeeper) days. I +literally days spent sorting out patches, creating what in essence +were standalone commits that i guessed to be related to that bug. + +It was a tool of absolute last resort. I'd rather spend days looking +at printk output than do a manual 'patch bisection'. + +With Git bisect it's a breeze: in the best case i can get a ~15 step +kernel bisection done in 20-30 minutes, in an automated way. Even with +manual help or when bisecting multiple, overlapping bugs, it's rarely +more than an hour. + +In fact it's invaluable because there are bugs i would never even +_try_ to debug if it wasn't for git bisect. In the past there were bug +patterns that were immediately hopeless for me to debug - at best i +could send the crash/bug signature to lkml and hope that someone else +can think of something. + +And even if a bisection fails today it tells us something valuable +about the bug: that it's non-deterministic - timing or kernel image +layout dependent. + +So git bisect is unconditional goodness - and feel free to quote that +;-) +_____________ + +Acknowledgements +---------------- + +Many thanks to Junio Hamano for his help in reviewing this paper, for +reviewing the patches I sent to the git mailing list, for discussing +some ideas and helping me improve them, for improving "git bisect" a +lot and for his awesome work in maintaining and developing Git. + +Many thanks to Ingo Molnar for giving me very useful information that +appears in this paper, for commenting on this paper, for his +suggestions to improve "git bisect" and for evangelizing "git bisect" +on the linux kernel mailing lists. + +Many thanks to Linus Torvalds for inventing, developing and +evangelizing "git bisect", Git and Linux. + +Many thanks to the many other great people who helped one way or +another when I worked on git, especially to Andreas Ericsson, Johannes +Schindelin, H. Peter Anvin, Daniel Barkalow, Bill Lear, John Hawley, +Shawn O. Pierce, Jeff King, Sam Vilain, Jon Seymour. + +Many thanks to the Linux-Kongress program committee for choosing the +author to given a talk and for publishing this paper. + +References +---------- + +- [[[1]]] http://www.nist.gov/public_affairs/releases/n02-10.htm['Software Errors Cost U.S. Economy $59.5 Billion Annually'. Nist News Release.] +- [[[2]]] http://java.sun.com/docs/codeconv/html/CodeConventions.doc.html#16712['Code Conventions for the Java Programming Language'. Sun Microsystems.] +- [[[3]]] http://en.wikipedia.org/wiki/Software_maintenance['Software maintenance'. Wikipedia.] +- [[[4]]] http://article.gmane.org/gmane.comp.version-control.git/45195/[Junio C Hamano. 'Automated bisect success story'. Gmane.] +- [[[5]]] http://lwn.net/Articles/317154/[Christian Couder. 'Fully automated bisecting with "git bisect run"'. LWN.net.] +- [[[6]]] http://lwn.net/Articles/277872/[Jonathan Corbet. 'Bisection divides users and developers'. LWN.net.] +- [[[7]]] http://article.gmane.org/gmane.linux.scsi/36652/[Ingo Molnar. 'Re: BUG 2.6.23-rc3 can't see sd partitions on Alpha'. Gmane.] +- [[[8]]] http://www.kernel.org/pub/software/scm/git/docs/git-bisect.html[Junio C Hamano and the git-list. 'git-bisect(1) Manual Page'. Linux Kernel Archives.] +- [[[9]]] http://github.com/Ealdwulf/bbchop[Ealdwulf. 'bbchop'. GitHub.] diff --git a/Documentation/git-bisect.txt b/Documentation/git-bisect.txt index 63e7a42cb..c39d957c3 100644 --- a/Documentation/git-bisect.txt +++ b/Documentation/git-bisect.txt @@ -20,7 +20,7 @@ on the subcommand: git bisect bad [] git bisect good [...] git bisect skip [(|)...] - git bisect reset [] + git bisect reset [] git bisect visualize git bisect replay git bisect log @@ -81,16 +81,27 @@ will have been left with the first bad kernel revision in "refs/bisect/bad". Bisect reset ~~~~~~~~~~~~ -To return to the original head after a bisect session, issue the -following command: +After a bisect session, to clean up the bisection state and return to +the original HEAD, issue the following command: ------------------------------------------------ $ git bisect reset ------------------------------------------------ -This resets the tree to the original branch instead of being on the -bisection commit ("git bisect start" will also do that, as it resets -the bisection state). +By default, this will return your tree to the commit that was checked +out before `git bisect start`. (A new `git bisect start` will also do +that, as it cleans up the old bisection state.) + +With an optional argument, you can return to a different commit +instead: + +------------------------------------------------ +$ git bisect reset +------------------------------------------------ + +For example, `git bisect reset HEAD` will leave you on the current +bisection commit and avoid switching commits at all, while `git bisect +reset bisect/bad` will check out the first bad revision. Bisect visualize ~~~~~~~~~~~~~~~~ @@ -319,6 +330,11 @@ Documentation ------------- Documentation by Junio C Hamano and the git-list . +SEE ALSO +-------- +link:git-bisect-lk2009.html[Fighting regressions with git bisect], +linkgit:git-blame[1]. + GIT --- Part of the linkgit:git[1] suite diff --git a/Documentation/git-blame.txt b/Documentation/git-blame.txt index 8c7b7b083..b786471dd 100644 --- a/Documentation/git-blame.txt +++ b/Documentation/git-blame.txt @@ -9,7 +9,7 @@ SYNOPSIS -------- [verse] 'git blame' [-c] [-b] [-l] [--root] [-t] [-f] [-n] [-s] [-p] [-w] [--incremental] [-L n,m] - [-S ] [-M] [-C] [-C] [--since=] + [-S ] [-M] [-C] [-C] [-C] [--since=] [ | --contents | --reverse ] [--] DESCRIPTION diff --git a/Documentation/git-bundle.txt b/Documentation/git-bundle.txt index aee7e4a8c..c3a066e60 100644 --- a/Documentation/git-bundle.txt +++ b/Documentation/git-bundle.txt @@ -24,7 +24,7 @@ ssh, rsync, http) cannot be used. This command provides support for 'git-fetch' and 'git-pull' to operate by packaging objects and references in an archive at the originating machine, then importing those into another repository using 'git-fetch' and 'git-pull' -after moving the archive by some means (i.e., by sneakernet). As no +after moving the archive by some means (e.g., by sneakernet). As no direct connection between the repositories exists, the user must specify a basis for the bundle that is held by the destination repository: the bundle assumes that all objects in the basis are already in the diff --git a/Documentation/git-check-ref-format.txt b/Documentation/git-check-ref-format.txt index 0b7982ea7..0aeef2478 100644 --- a/Documentation/git-check-ref-format.txt +++ b/Documentation/git-check-ref-format.txt @@ -9,7 +9,8 @@ SYNOPSIS -------- [verse] 'git check-ref-format' -'git check-ref-format' [--branch] +'git check-ref-format' --print +'git check-ref-format' --branch DESCRIPTION ----------- @@ -63,16 +64,31 @@ reference name expressions (see linkgit:git-rev-parse[1]): . at-open-brace `@{` is used as a notation to access a reflog entry. -With the `--branch` option, it expands a branch name shorthand and -prints the name of the branch the shorthand refers to. +With the `--print` option, if 'refname' is acceptable, it prints the +canonicalized name of a hypothetical reference with that name. That is, +it prints 'refname' with any extra `/` characters removed. -EXAMPLE -------- +With the `--branch` option, it expands the ``previous branch syntax'' +`@{-n}`. For example, `@{-1}` is a way to refer the last branch you +were on. This option should be used by porcelains to accept this +syntax anywhere a branch name is expected, so they can act as if you +typed the branch name. -git check-ref-format --branch @{-1}:: - -Print the name of the previous branch. +EXAMPLES +-------- +* Print the name of the previous branch: ++ +------------ +$ git check-ref-format --branch @{-1} +------------ + +* Determine the reference name to use for a new branch: ++ +------------ +$ ref=$(git check-ref-format --print "refs/heads/$newbranch") || +die "we do not like '$newbranch' as a branch name." +------------ GIT --- diff --git a/Documentation/git-clone.txt b/Documentation/git-clone.txt index 5ebcba1c7..7ccd742a8 100644 --- a/Documentation/git-clone.txt +++ b/Documentation/git-clone.txt @@ -11,7 +11,7 @@ SYNOPSIS [verse] 'git clone' [--template=] [-l] [-s] [--no-hardlinks] [-q] [-n] [--bare] [--mirror] - [-o ] [-u ] [--reference ] + [-o ] [-b ] [-u ] [--reference ] [--depth ] [--recursive] [--] [] DESCRIPTION @@ -39,7 +39,7 @@ OPTIONS --local:: -l:: When the repository to clone from is on a local machine, - this flag bypasses normal "git aware" transport + this flag bypasses the normal "git aware" transport mechanism and clones the repository by making a copy of HEAD and everything under objects and refs directories. The files under `.git/objects/` directory are hardlinked @@ -60,7 +60,7 @@ OPTIONS -s:: When the repository to clone is on the local machine, instead of using hard links, automatically setup - .git/objects/info/alternates to share the objects + `.git/objects/info/alternates` to share the objects with the source repository. The resulting repository starts out without any object of its own. + @@ -69,7 +69,7 @@ it unless you understand what it does. If you clone your repository using this option and then delete branches (or use any other git command that makes any existing commit unreferenced) in the source repository, some objects may become unreferenced (or dangling). -These objects may be removed by normal git operations (such as 'git-commit') +These objects may be removed by normal git operations (such as `git commit`) which automatically call `git gc --auto`. (See linkgit:git-gc[1].) If these objects are removed and were referenced by the cloned repository, then the cloned repository will become corrupt. @@ -86,13 +86,13 @@ objects from the source repository into a pack in the cloned repository. --reference :: If the reference repository is on the local machine, - automatically setup .git/objects/info/alternates to + automatically setup `.git/objects/info/alternates` to obtain objects from the reference repository. Using an already existing repository as an alternate will require fewer objects to be copied from the repository being cloned, reducing network and local storage costs. + -*NOTE*: see NOTE to --shared option. +*NOTE*: see the NOTE for the `--shared` option. --quiet:: -q:: @@ -101,7 +101,7 @@ objects from the source repository into a pack in the cloned repository. --verbose:: -v:: - Display the progressbar, even in case the standard output is not + Display the progress bar, even in case the standard output is not a terminal. --no-checkout:: @@ -121,17 +121,17 @@ objects from the source repository into a pack in the cloned repository. configuration variables are created. --mirror:: - Set up a mirror of the remote repository. This implies --bare. + Set up a mirror of the remote repository. This implies `--bare`. --origin :: -o :: - Instead of using the remote name 'origin' to keep track - of the upstream repository, use . + Instead of using the remote name `origin` to keep track + of the upstream repository, use ``. --branch :: -b :: Instead of pointing the newly created HEAD to the branch pointed - to by the cloned repository's HEAD, point to branch + to by the cloned repository's HEAD, point to `` branch instead. In a non-bare repository, this is the branch that will be checked out. @@ -158,7 +158,7 @@ objects from the source repository into a pack in the cloned repository. --recursive:: After the clone is created, initialize all submodules within, using their default settings. This is equivalent to running - 'git submodule update --init --recursive' immediately after + `git submodule update --init --recursive` immediately after the clone is finished. This option is ignored if the cloned repository does not have a worktree/checkout (i.e. if any of `--no-checkout`/`-n`, `--bare`, or `--mirror` is given) @@ -171,8 +171,8 @@ objects from the source repository into a pack in the cloned repository. :: The name of a new directory to clone into. The "humanish" part of the source repository is used if no directory is - explicitly given ("repo" for "/path/to/repo.git" and "foo" - for "host.xz:foo/.git"). Cloning into an existing directory + explicitly given (`repo` for `/path/to/repo.git` and `foo` + for `host.xz:foo/.git`). Cloning into an existing directory is only allowed if the directory is empty. :git-clone: 1 diff --git a/Documentation/git-commit.txt b/Documentation/git-commit.txt index 0578a40d8..d227cec9b 100644 --- a/Documentation/git-commit.txt +++ b/Documentation/git-commit.txt @@ -9,7 +9,7 @@ SYNOPSIS -------- [verse] 'git commit' [-a | --interactive] [-s] [-v] [-u] [--amend] [--dry-run] - [(-c | -C) ] [-F | -m ] + [(-c | -C) ] [-F | -m ] [--reset-author] [--allow-empty] [--no-verify] [-e] [--author=] [--cleanup=] [--] [[-i | -o ]...] @@ -69,6 +69,11 @@ OPTIONS Like '-C', but with '-c' the editor is invoked, so that the user can further edit the commit message. +--reset-author:: + When used with -C/-c/--amend options, declare that the + authorship of the resulting commit now belongs of the committer. + This also renews the author timestamp. + -F :: --file=:: Take the commit message from the given file. Use '-' to @@ -323,7 +328,7 @@ ENVIRONMENT AND CONFIGURATION VARIABLES The editor used to edit the commit log message will be chosen from the GIT_EDITOR environment variable, the core.editor configuration variable, the VISUAL environment variable, or the EDITOR environment variable (in that -order). +order). See linkgit:git-var[1] for details. HOOKS ----- diff --git a/Documentation/git-cvsserver.txt b/Documentation/git-cvsserver.txt index 785779e22..99a7c1470 100644 --- a/Documentation/git-cvsserver.txt +++ b/Documentation/git-cvsserver.txt @@ -182,10 +182,9 @@ Database Backend ---------------- 'git-cvsserver' uses one database per git head (i.e. CVS module) to -store information about the repository for faster access. The -database doesn't contain any persistent data and can be completely -regenerated from the git repository at any time. The database -needs to be updated (i.e. written to) after every commit. +store information about the repository to maintain consistent +CVS revision numbers. The database needs to be +updated (i.e. written to) after every commit. If the commit is done directly by using `git` (as opposed to using 'git-cvsserver') the update will need to happen on the @@ -204,6 +203,18 @@ write so it might not be enough to grant the users using 'git-cvsserver' write access to the database file without granting them write access to the directory, too. +The database can not be reliably regenerated in a +consistent form after the branch it is tracking has changed. +Example: For merged branches, 'git-cvsserver' only tracks +one branch of development, and after a 'git-merge' an +incrementally updated database may track a different branch +than a database regenerated from scratch, causing inconsistent +CVS revision numbers. `git-cvsserver` has no way of knowing which +branch it would have picked if it had been run incrementally +pre-merge. So if you have to fully or partially (from old +backup) regenerate the database, you should be suspicious +of pre-existing CVS sandboxes. + You can configure the database backend with the following configuration variables: diff --git a/Documentation/git-describe.txt b/Documentation/git-describe.txt index b231dbb94..78b9808aa 100644 --- a/Documentation/git-describe.txt +++ b/Documentation/git-describe.txt @@ -8,7 +8,9 @@ git-describe - Show the most recent tag that is reachable from a commit SYNOPSIS -------- +[verse] 'git describe' [--all] [--tags] [--contains] [--abbrev=] ... +'git describe' [--all] [--tags] [--contains] [--abbrev=] --dirty[=] DESCRIPTION ----------- @@ -27,6 +29,11 @@ OPTIONS ...:: Committish object names to describe. +--dirty[=]:: + Describe the working tree. + It means describe HEAD and appends (`-dirty` by + default) if the working tree is dirty. + --all:: Instead of using only the annotated tags, use any ref found in `.git/refs/`. This option enables matching @@ -44,7 +51,9 @@ OPTIONS --abbrev=:: Instead of using the default 7 hexadecimal digits as the - abbreviated object name, use digits. + abbreviated object name, use digits, or as many digits + as needed to form a unique object name. An of 0 + will suppress long format, only showing the closest tag. --candidates=:: Instead of considering only the 10 most recent tags as @@ -68,8 +77,8 @@ OPTIONS This is useful when you want to see parts of the commit object name in "describe" output, even when the commit in question happens to be a tagged version. Instead of just emitting the tag name, it will - describe such a commit as v1.2-0-deadbeef (0th commit since tag v1.2 - that points at object deadbeef....). + describe such a commit as v1.2-0-gdeadbee (0th commit since tag v1.2 + that points at object deadbee....). --match :: Only consider tags matching the given pattern (can be used to avoid @@ -108,7 +117,7 @@ the output shows the reference path as well: [torvalds@g5 git]$ git describe --all --abbrev=4 v1.0.5^2 tags/v1.0.0-21-g975b - [torvalds@g5 git]$ git describe --all HEAD^ + [torvalds@g5 git]$ git describe --all --abbrev=4 HEAD^ heads/lt/describe-7-g975b With --abbrev set to 0, the command can be used to find the @@ -117,6 +126,13 @@ closest tagname without any suffix: [torvalds@g5 git]$ git describe --abbrev=0 v1.0.5^2 tags/v1.0.0 +Note that the suffix you get if you type these commands today may be +longer than what Linus saw above when he ran these commands, as your +git repository may have new commits whose object names begin with +975b that did not exist back then, and "-g975b" suffix alone may not +be sufficient to disambiguate these commits. + + SEARCH STRATEGY --------------- diff --git a/Documentation/git-difftool.txt b/Documentation/git-difftool.txt index 96a6c51a4..8e9aed67d 100644 --- a/Documentation/git-difftool.txt +++ b/Documentation/git-difftool.txt @@ -31,7 +31,7 @@ OPTIONS Use the diff tool specified by . Valid merge tools are: kdiff3, kompare, tkdiff, meld, xxdiff, emerge, vimdiff, gvimdiff, - ecmerge, diffuse, opendiff and araxis. + ecmerge, diffuse, opendiff, p4merge and araxis. + If a diff tool is not specified, 'git-difftool' will use the configuration variable `diff.tool`. If the diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt index c2f483a8d..e6d364f53 100644 --- a/Documentation/git-fast-import.txt +++ b/Documentation/git-fast-import.txt @@ -311,12 +311,12 @@ change to the project. .... 'commit' SP LF mark? - ('author' SP SP LT GT SP LF)? - 'committer' SP SP LT GT SP LF + ('author' (SP )? SP LT GT SP LF)? + 'committer' (SP )? SP LT GT SP LF data ('from' SP LF)? ('merge' SP LF)? - (filemodify | filedelete | filecopy | filerename | filedeleteall)* + (filemodify | filedelete | filecopy | filerename | filedeleteall | notemodify)* LF? .... @@ -339,14 +339,13 @@ commit message use a 0 length data. Commit messages are free-form and are not interpreted by Git. Currently they must be encoded in UTF-8, as fast-import does not permit other encodings to be specified. -Zero or more `filemodify`, `filedelete`, `filecopy`, `filerename` -and `filedeleteall` commands +Zero or more `filemodify`, `filedelete`, `filecopy`, `filerename`, +`filedeleteall` and `notemodify` commands may be included to update the contents of the branch prior to creating the commit. These commands may be supplied in any order. However it is recommended that a `filedeleteall` command precede -all `filemodify`, `filecopy` and `filerename` commands in the same -commit, as `filedeleteall` -wipes the branch clean (see below). +all `filemodify`, `filecopy`, `filerename` and `notemodify` commands in +the same commit, as `filedeleteall` wipes the branch clean (see below). The `LF` after the command is optional (it used to be required). @@ -595,6 +594,40 @@ more memory per active branch (less than 1 MiB for even most large projects); so frontends that can easily obtain only the affected paths for a commit are encouraged to do so. +`notemodify` +^^^^^^^^^^^^ +Included in a `commit` command to add a new note (annotating a given +commit) or change the content of an existing note. This command has +two different means of specifying the content of the note. + +External data format:: + The data content for the note was already supplied by a prior + `blob` command. The frontend just needs to connect it to the + commit that is to be annotated. ++ +.... + 'N' SP SP LF +.... ++ +Here `` can be either a mark reference (`:`) +set by a prior `blob` command, or a full 40-byte SHA-1 of an +existing Git blob object. + +Inline data format:: + The data content for the note has not been supplied yet. + The frontend wants to supply it as part of this modify + command. ++ +.... + 'N' SP 'inline' SP LF + data +.... ++ +See below for a detailed description of the `data` command. + +In both formats `` is any of the commit specification +expressions also accepted by `from` (see above). + `mark` ~~~~~~ Arranges for fast-import to save a reference to the current object, allowing @@ -624,7 +657,7 @@ lightweight (non-annotated) tags see the `reset` command below. .... 'tag' SP LF 'from' SP LF - 'tagger' SP SP LT GT SP LF + 'tagger' (SP )? SP LT GT SP LF data .... diff --git a/Documentation/git-fetch.txt b/Documentation/git-fetch.txt index edb77dc54..9b9e5686e 100644 --- a/Documentation/git-fetch.txt +++ b/Documentation/git-fetch.txt @@ -47,6 +47,35 @@ include::pull-fetch-param.txt[] include::urls-remotes.txt[] + +EXAMPLES +-------- + +* Update the remote-tracking branches: ++ +------------------------------------------------ +$ git fetch origin +------------------------------------------------ ++ +The above command copies all branches from the remote refs/heads/ +namespace and stores them to the local refs/remotes/origin/ namespace, +unless the branch..fetch option is used to specify a non-default +refspec. + +* Using refspecs explicitly: ++ +------------------------------------------------ +$ git fetch origin +pu:pu maint:tmp +------------------------------------------------ ++ +This updates (or creates, as necessary) branches `pu` and `tmp` in +the local repository by fetching from the branches (respectively) +`pu` and `maint` from the remote repository. ++ +The `pu` branch will be updated even if it is does not fast-forward, +because it is prefixed with a plus sign; `tmp` will not be. + + SEE ALSO -------- linkgit:git-pull[1] diff --git a/Documentation/git-filter-branch.txt b/Documentation/git-filter-branch.txt index 2b40babb6..394a77a35 100644 --- a/Documentation/git-filter-branch.txt +++ b/Documentation/git-filter-branch.txt @@ -159,7 +159,18 @@ to other tags will be rewritten to point to the underlying commit. --subdirectory-filter :: Only look at the history which touches the given subdirectory. The result will contain that directory (and only that) as its - project root. + project root. Implies --remap-to-ancestor. + +--remap-to-ancestor:: + Rewrite refs to the nearest rewritten ancestor instead of + ignoring them. ++ +Normally, positive refs on the command line are only changed if the +commit they point to was rewritten. However, you can limit the extent +of this rewriting by using linkgit:rev-list[1] arguments, e.g., path +limiters. Refs pointing to such excluded commits would then normally +be ignored. With this option, they are instead rewritten to point at +the nearest ancestor that was not excluded. --prune-empty:: Some kind of filters will generate empty commits, that left the tree diff --git a/Documentation/git-format-patch.txt b/Documentation/git-format-patch.txt index 687e66759..f1fd0df08 100644 --- a/Documentation/git-format-patch.txt +++ b/Documentation/git-format-patch.txt @@ -43,28 +43,28 @@ There are two ways to specify which commits to operate on. The first rule takes precedence in the case of a single . To apply the second rule, i.e., format everything since the beginning of -history up until , use the '\--root' option: "git format-patch -\--root ". If you want to format only itself, you -can do this with "git format-patch -1 ". +history up until , use the '\--root' option: `git format-patch +\--root `. If you want to format only itself, you +can do this with `git format-patch -1 `. By default, each output file is numbered sequentially from 1, and uses the first line of the commit message (massaged for pathname safety) as -the filename. With the --numbered-files option, the output file names +the filename. With the `--numbered-files` option, the output file names will only be numbers, without the first line of the commit appended. The names of the output files are printed to standard -output, unless the --stdout option is specified. +output, unless the `--stdout` option is specified. -If -o is specified, output files are created in . Otherwise +If `-o` is specified, output files are created in . Otherwise they are created in the current working directory. By default, the subject of a single patch is "[PATCH] First Line" and the subject when multiple patches are output is "[PATCH n/m] First -Line". To force 1/1 to be added for a single patch, use -n. To omit -patch numbers from the subject, use -N +Line". To force 1/1 to be added for a single patch, use `-n`. To omit +patch numbers from the subject, use `-N`. -If given --thread, 'git-format-patch' will generate In-Reply-To and -References headers to make the second and subsequent patch mails appear -as replies to the first mail; this also generates a Message-Id header to +If given `--thread`, `git-format-patch` will generate `In-Reply-To` and +`References` headers to make the second and subsequent patch mails appear +as replies to the first mail; this also generates a `Message-Id` header to reference. OPTIONS @@ -112,7 +112,7 @@ include::diff-options.txt[] --attach[=]:: Create multipart/mixed attachment, the first part of which is the commit message and the patch itself in the - second part, with "Content-Disposition: attachment". + second part, with `Content-Disposition: attachment`. --no-attach:: Disable the creation of an attachment, overriding the @@ -121,13 +121,13 @@ include::diff-options.txt[] --inline[=]:: Create multipart/mixed attachment, the first part of which is the commit message and the patch itself in the - second part, with "Content-Disposition: inline". + second part, with `Content-Disposition: inline`. --thread[=