{"thread":{"id":"36437","subject":"[ANNOUNCE] Git v2.0.0-rc0","startedAt":"2014-04-18T19:37:27Z","lastAt":"2014-04-23T16:40:35Z","messageCount":16,"participants":["Junio C Hamano","Johan Herland","Felipe Contreras","Kyle J. McKay","Jonathan Nieder","brian m. carlson"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"239077","messageId":"xmqqk3ambf9k.fsf@gitster.dls.corp.google.com","threadId":"36437","inReplyTo":null,"subject":"[ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-18T19:37:27Z","receivedAt":"2014-04-18T19:37:27Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"An early preview release Git v2.0.0-rc0 is now available for\ntesting at the usual places.\n\nA major version bump between v1.x.x series and the upcoming v2.0.0\nmeans there are a handful of backward incompatible UI improvements,\nbut for most people, all the tricky preparation for the transition\nwould have been already done for you and the upcoming release just\nflips the default. Unless you were living in a cave and have stayed\nwith an ancient version of Git (e.g. one before 1.8.2 that was\nreleased more than a year ago) for all these times, that is---those\nof you may want to double check the backward compatibility notes\nsection at the beginning of the draft release notes.\n\nThe tarballs are found at:\n\n https://www.kernel.org/pub/software/scm/git/testing/\n\nThe following public repositories all have a copy of the 'v2.0.0-rc0'\ntag and the 'master' branch that the tag points at:\n\n url = https://kernel.googlesource.com/pub/scm/git/git\n url = git://repo.or.cz/alt-git.git\n url = https://code.google.com/p/git-core/\n url = git://git.sourceforge.jp/gitroot/git-core/git.git\n url = git://git-core.git.sourceforge.net/gitroot/git-core/git-core\n url = https://github.com/gitster/git\n\nGit v2.0 Release Notes (draft)\n==============================\n\nBackward compatibility notes\n----------------------------\n\nWhen \"git push [$there]\" does not say what to push, we have used the\ntraditional \"matching\" semantics so far (all your branches were sent\nto the remote as long as there already are branches of the same name\nover there). In Git 2.0, the default is now the \"simple\" semantics,\nwhich pushes:\n\n - only the current branch to the branch with the same name, and only\n when the current branch is set to integrate with that remote\n branch, if you are pushing to the same remote as you fetch from; or\n\n - only the current branch to the branch with the same name, if you\n are pushing to a remote that is not where you usually fetch from.\n\nYou can use the configuration variable \"push.default\" to change\nthis. If you are an old-timer who wants to keep using the\n\"matching\" semantics, you can set the variable to \"matching\", for\nexample. Read the documentation for other possibilities.\n\nWhen \"git add -u\" and \"git add -A\" are run inside a subdirectory\nwithout specifying which paths to add on the command line, they\noperate on the entire tree for consistency with \"git commit -a\" and\nother commands (these commands used to operate only on the current\nsubdirectory). Say \"git add -u .\" or \"git add -A .\" if you want to\nlimit the operation to the current directory.\n\n\"git add \" is the same as \"git add -A \" now, so that\n\"git add dir/\" will notice paths you removed from the directory and\nrecord the removal. In older versions of Git, \"git add \" used\nto ignore removals. You can say \"git add --ignore-removal \" to\nadd only added or modified paths in , if you really want to.\n\nThe \"-q\" option to \"git diff-files\", which does *NOT* mean \"quiet\",\nhas been removed (it told Git to ignore deletion, which you can do\nwith \"git diff-files --diff-filter=d\").\n\n\"git request-pull\" lost a few \"heuristics\" that often led to mistakes.\n\n\nUpdates since v1.9 series\n-------------------------\n\nUI, Workflows & Features\n\n * The \"multi-mail\" post-receive hook (in contrib/) has been updated\n to a more recent version from the upstream.\n\n * \"git gc --aggressive\" learned \"--depth\" option and\n \"gc.aggressiveDepth\" configuration variable to allow use of a less\n insane depth than the built-in default value of 250.\n\n * \"git log\" learned the \"--show-linear-break\" option to show where a\n single strand-of-pearls is broken in its output.\n\n * The \"rev-parse --parseopt\" mechanism used by scripted Porcelains to\n parse command line options and to give help text learned to take\n the argv-help (the placeholder string for an option parameter,\n e.g. \"key-id\" in \"--gpg-sign=\").\n\n * The pattern to find where the function begins in C/C++ used in\n \"diff\" and \"grep -p\" have been updated to help C++ source better.\n\n * \"git rebase\" learned to interpret a lone \"-\" as \"@{-1}\", the\n branch that we were previously on.\n\n * \"git commit --cleanup=\" learned a new mode, scissors.\n\n * \"git tag --list\" output can be sorted using \"version sort\" with\n \"--sort=version:refname\".\n\n * Discard the accumulated \"heuristics\" to guess from which branch the\n result wants to be pulled from and make sure what the end user\n specified is not second-guessed by \"git request-pull\", to avoid\n mistakes. When you pushed out your 'master' branch to your public\n repository as 'for-linus', use the new \"master:for-linus\" syntax to\n denote the branch to be pulled.\n\n * \"git grep\" learned to behave in a way similar to native grep when\n \"-h\" (no header) and \"-c\" (count) options are given.\n\n * transport-helper, fast-import and fast-export have been updated to\n allow the ref mapping and ref deletion in a way similar to the\n natively supported transports.\n\n * The \"simple\" mode is the default for \"git push\".\n\n * \"git add -u\" and \"git add -A\", when run without any pathspec, is a\n tree-wide operation even when run inside a subdirectory of a\n working tree.\n\n * \"git add is the same as \"git add -A \" now.\n\n * \"core.statinfo\" configuration variable, which is a\n never-advertised synonym to \"core.checkstat\", has been removed.\n\n * The \"-q\" option to \"git diff-files\", which does *NOT* mean\n \"quiet\", has been removed (it told Git to ignore deletion, which\n you can do with \"git diff-files --diff-filter=d\").\n\n * Server operators can loosen the \"tips of refs only\" restriction for\n the remote archive service with the uploadarchive.allowUnreachable\n configuration option.\n\n * The progress indicators from various time-consuming commands have\n been marked for i18n/l10n.\n\n * \"git notes -C \" diagnoses an attempt to use an object that\n is not a blob as an error.\n\n * \"git config\" learned to read from the standard input when \"-\" is\n given as the value to its \"--file\" parameter (attempting an\n operation to update the configuration in the standard input of\n course is rejected).\n\n * Trailing whitespaces in .gitignore files, unless they are quoted\n for fnmatch(3), e.g. \"path\\ \", are warned and ignored. Strictly\n speaking, this is a backward incompatible change, but very unlikely\n to bite any sane user and adjusting should be obvious and easy.\n\n * Many commands that create commits, e.g. \"pull\", \"rebase\",\n learned to take the --gpg-sign option on the command line.\n\n * \"git commit\" can be told to always GPG sign the resulting commit\n by setting \"commit.gpgsign\" configuration variable to true (the\n command line option --no-gpg-sign should override it).\n\n * \"git pull\" can be told to only accept fast-forward by setting the\n new \"pull.ff\" configuration.\n\n * \"git reset\" learned \"-N\" option, which does not reset the index\n fully for paths the index knows about but the tree-ish the command\n resets to does not (these paths are kept as intend-to-add entries).\n\n\nPerformance, Internal Implementation, etc.\n\n * The compilation options to port to AIX and to MSVC have been\n updated.\n\n * We started using wildmatch() in place of fnmatch(3) a few releases\n ago; complete the process and stop using fnmatch(3).\n\n * Uses of curl's \"multi\" interface and \"easy\" interface do not mix\n well when we attempt to reuse outgoing connections. Teach the RPC\n over http code, used in the smart HTTP transport, not to use the\n \"easy\" interface.\n\n * The bitmap-index feature from JGit has been ported, which should\n significantly improve performance when serving objects form a\n repository that uses it.\n\n * The way \"git log --cc\" shows a combined diff against multiple\n parents have been optimized.\n\n * The prefixcmp() and suffixcmp() functions are gone. Use\n starts_with() and ends_with(), and also consider if skip_prefix()\n suits your needs better when using the former.\n\n\nAlso contains various documentation updates and code clean-ups. Many\nof them came from flurry of activities as GSoC candidate microproject\nexercises.\n\n\nFixes since v1.9 series\n-----------------------\n\nUnless otherwise noted, all the fixes since v1.9 in the maintenance\ntrack are contained in this release (see the maintenance releases'\nnotes for details).\n\n * zsh prompt (in contrib/) leaked unnecessary error messages.\n\n * bash completion (in contrib/) did not complete the refs and remotes\n correctly given \"git pu\" when \"pu\" is aliased to \"push\".\n\n * Some more Unicode codepoints defined in Unicode 6.3 as having zero\n width have been taught to our display column counting logic.\n (merge d813ab9 tb/unicode-6.3-zero-width later to maint).\n\n * Some tests used shell constructs that did not work well on FreeBSD\n (merge ff7a1c6 km/avoid-bs-in-shell-glob later to maint).\n (merge 00764ca km/avoid-cp-a later to maint).\n\n * \"git update-ref --stdin\" did not fail a request to create a ref\n when the ref already existed.\n (merge b9d56b5 mh/update-ref-batch-create-fix later to maint).\n\n * \"git diff --no-index -Mq a b\" fell into an infinite loop.\n (merge ad1c3fb jc/fix-diff-no-index-diff-opt-parse later to maint).\n\n * \"git fetch --prune\", when the right-hand-side of multiple fetch\n refspecs overlap (e.g. storing \"refs/heads/*\" to\n \"refs/remotes/origin/*\", while storing \"refs/frotz/*\" to\n \"refs/remotes/origin/fr/*\"), aggressively thought that lack of\n \"refs/heads/fr/otz\" on the origin site meant we should remove\n \"refs/remotes/origin/fr/otz\" from us, without checking their\n \"refs/frotz/otz\" first.\n\n Note that such a configuration is inherently unsafe (think what\n should happen when \"refs/heads/fr/otz\" does appear on the origin\n site), but that is not a reason not to be extra careful.\n (merge e6f6371 cn/fetch-prune-overlapping-destination later to maint).\n\n * \"git status --porcelain --branch\" showed its output with labels\n \"ahead/behind/gone\" translated to the user's locale.\n (merge 7a76c28 mm/status-porcelain-format-i18n-fix later to maint).\n\n * A stray environment variable $prefix could have leaked into and\n affected the behaviour of the \"subtree\" script (in contrib/).\n\n * When it is not necessary to edit a commit log message (e.g. \"git\n commit -m\" is given a message without specifying \"-e\"), we used to\n disable the spawning of the editor by overriding GIT_EDITOR, but\n this means all the uses of the editor, other than to edit the\n commit log message, are also affected.\n (merge b549be0 bp/commit-p-editor later to maint).\n\n * \"git mv\" that moves a submodule forgot to adjust the array that\n uses to keep track of which submodules were to be moved to update\n its configuration.\n (merge fb8a4e8 jk/mv-submodules-fix later to maint).\n\n * Length limit for the pathname used when removing a path in a deep\n subdirectory has been removed to avoid buffer overflows.\n (merge 2f29e0c mh/remove-subtree-long-pathname-fix later to maint).\n\n * The test helper lib-terminal always run an actual test_expect_*\n when included, which screwed up with the use of skil-all that may\n have to be done later.\n (merge 7e27173 jk/lib-terminal-lazy later to maint).\n\n * \"git index-pack\" used a wrong variable to name the keep-file in an\n error message when the file cannot be written or closed.\n (merge de983a0 nd/index-pack-error-message later to maint).\n\n * \"rebase -i\" produced a broken insn sheet when the title of a commit\n happened to contain '\\n' (or ended with '\\c') due to a careless use\n of 'echo'.\n (merge cb1aefd us/printf-not-echo later to maint).\n\n * There were a few instances of 'git-foo' remaining in the\n documentation that should have been spelled 'git foo'.\n (merge 3c3e6f5 rr/doc-merge-strategies later to maint).\n\n * Serving objects from a shallow repository needs to write a\n new file to hold the temporary shallow boundaries but it was not\n cleaned when we exit due to die() or a signal.\n (merge 7839632 jk/shallow-update-fix later to maint).\n\n * When \"git stash pop\" stops after failing to apply the stash\n (e.g. due to conflicting changes), the stash is not dropped. State\n that explicitly in the output to let the users know.\n (merge 2d4c993 jc/stash-pop-not-popped later to maint).\n\n * The labels in \"git status\" output that describe the nature of\n conflicts (e.g. \"both deleted\") were limited to 20 bytes, which was\n too short for some l10n (e.g. fr).\n (merge c7cb333 jn/wt-status later to maint).\n\n * \"git clean -d pathspec\" did not use the given pathspec correctly\n and ended up cleaning too much.\n (merge 1f2e108 jk/clean-d-pathspec later to maint).\n\n * \"git difftool\" misbehaved when the repository is bound to the\n working tree with the \".git file\" mechanism, where a textual file\n \".git\" tells us where it is.\n (merge fcfec8b da/difftool-git-files later to maint).\n\n * \"git push\" did not pay attention to branch.*.pushremote if it is\n defined earlier than remote.pushdefault; the order of these two\n variables in the configuration file should not matter, but it did\n by mistake.\n (merge 98b406f jk/remote-pushremote-config-reading later to maint).\n\n * Codepaths that parse timestamps in commit objects have been\n tightened.\n (merge f80d1f9 jk/commit-dates-parsing-fix later to maint).\n\n * \"git diff --external-diff\" incorrectly fed the submodule directory\n in the working tree to the external diff driver when it knew it is\n the same as one of the versions being compared.\n (merge aba4727 tr/diff-submodule-no-reuse-worktree later to maint).\n\n * \"git reset\" needs to refresh the index when working in a working\n tree (it can also be used to match the index to the HEAD in an\n otherwise bare repository), but it failed to set up the working\n tree properly, causing GIT_WORK_TREE to be ignored.\n (merge b7756d4 nd/reset-setup-worktree later to maint).\n\n * \"git check-attr\" when working on a repository with a working tree\n did not work well when the working tree was specified via the\n --work-tree (and obviously with --git-dir) option.\n (merge cdbf623 jc/check-attr-honor-working-tree later to maint).\n\n * \"merge-recursive\" was broken in 1.7.7 era and stopped working in\n an empty (temporary) working tree, when there are renames\n involved. This has been corrected.\n (merge 6e2068a bk/refresh-missing-ok-in-merge-recursive later to maint.)\n\n * \"git rev-parse\" was loose in rejecting command line arguments\n that do not make sense, e.g. \"--default\" without the required\n value for that option.\n (merge a43219f ds/rev-parse-required-args later to maint.)\n\n * include.path variable (or any variable that expects a path that\n can use ~username expansion) in the configuration file is not a\n boolean, but the code failed to check it.\n (merge 67beb60 jk/config-path-include-fix later to maint.)\n\n * Commands that take pathspecs on the command line misbehaved when\n the pathspec is given as an absolute pathname (which is a\n practice not particularly encouraged) that points at a symbolic\n link in the working tree.\n (merge later 655ee9e mw/symlinks to maint.)\n\n * \"git diff --quiet -- pathspec1 pathspec2\" sometimes did not return\n correct status value.\n (merge f34b205 nd/diff-quiet-stat-dirty later to maint.)\n\n * Attempting to deepen a shallow repository by fetching over smart\n HTTP transport failed in the protocol exchange, when no-done\n extension was used. The fetching side waited for the list of\n shallow boundary commits after the sending end stopped talking to\n it.\n (merge 0232852 nd/http-fetch-shallow-fix later to maint.)\n\n * Allow \"git cmd path/\", when the 'path' is where a submodule is\n bound to the top-level working tree, to match 'path', despite the\n extra and unnecessary trailing slash (such a slash is often\n given by command line completion).\n (merge 2e70c01 nd/submodule-pathspec-ending-with-slash later to maint.)\n\n * Documentation and in-code comments had many instances of mistaken\n use of \"nor\", which have been corrected.\n (merge 235e8d5 jl/nor-or-nand-and later to maint).\n\n----------------------------------------------------------------\n\nChanges since v1.9.2 are as follows:\n\nAdam (1):\n branch.c: install_branch_config: simplify if chain\n\nAlbert L. Lash, IV (4):\n docs/merge-strategies: remove hyphen from mis-merges\n docs/git-remote: capitalize first word of initial blurb\n docs/git-clone: clarify use of --no-hardlinks option\n docs/git-blame: explain more clearly the example pickaxe use\n\nAndrew Keller (1):\n gitweb: Avoid overflowing page body frame with large images\n\nAstril Hayato (1):\n Documentation/gitk: document the location of the configulation file\n\nBenoit Sigoure (1):\n git-compat-util.h: #undef (v)snprintf before #define them\n\nBrian Bourn (2):\n diff-no-index: rename read_directory()\n diff-no-index: replace manual \".\"/\"..\" check with is_dot_or_dotdot()\n\nBrian Gesiak (3):\n t3200-branch: test setting branch as own upstream\n branch: use skip_prefix() in install_branch_config()\n rebase: allow \"-\" short-hand for the previous branch\n\nCharles Bailey (2):\n dir.c: make git_fnmatch() not inline\n tests: don't rely on strerror text when testing rmdir failure\n\nChris Angelico (1):\n config.txt: third-party tools may and do use their own variables\n\nChris Packham (2):\n Documentation/git-am: Document supported --patch-format options\n Documentation/git-am: typofix\n\nChristian Couder (1):\n strbuf: remove prefixcmp() and suffixcmp()\n\nDavid Aguilar (2):\n pull: add pull.ff configuration\n pull: add --ff-only to the help text\n\nDavid Kastrup (6):\n builtin/blame.c: struct blame_entry does not need a prev link\n builtin/blame.c: eliminate same_suspect()\n builtin/blame.c::prepare_lines: fix allocation size of sb->lineno\n blame.c: prepare_lines should not call xrealloc for every line\n builtin/blame.c::find_copy_in_blob: no need to scan for region end\n skip_prefix(): scan prefix only once\n\nDavid Tran (1):\n tests: use \"env\" to run commands with temporary env-var settings\n\nDirk Wallenstein (1):\n doc: status, remove leftover statement about '#' prefix\n\nDmitry Marakasov (1):\n configure.ac: link with -liconv for locale_charset()\n\nDmitry S. Dolzhenko (15):\n commit.c: use the generic \"sha1_pos\" function for lookup\n builtin/pack-objects.c: use ALLOC_GROW() in check_pbase_path()\n bundle.c: use ALLOC_GROW() in add_to_ref_list()\n cache-tree.c: use ALLOC_GROW() in find_subtree()\n commit.c: use ALLOC_GROW() in register_commit_graft()\n diff.c: use ALLOC_GROW()\n diffcore-rename.c: use ALLOC_GROW()\n patch-ids.c: use ALLOC_GROW() in add_commit()\n replace_object.c: use ALLOC_GROW() in register_replace_object()\n reflog-walk.c: use ALLOC_GROW()\n dir.c: use ALLOC_GROW() in create_simplify()\n attr.c: use ALLOC_GROW() in handle_attr_line()\n builtin/mktree.c: use ALLOC_GROW() in append_to_tree()\n read-cache.c: use ALLOC_GROW() in add_index_entry()\n sha1_file.c: use ALLOC_GROW() in pretend_sha1_file()\n\nElia Pinto (9):\n bisect.c: reduce scope of variable\n builtin/apply.c: reduce scope of variables\n builtin/blame.c: reduce scope of variables\n builtin/clean.c: reduce scope of variable\n builtin/commit.c: reduce scope of variables\n builtin/fetch.c: reduce scope of variable\n builtin/gc.c: reduce scope of variables\n check-builtins.sh: use the $(...) construct for command substitution\n git-am.sh: use the $(...) construct for command substitution\n\nEric Sunshine (2):\n name-hash: retire unused index_name_exists()\n sh-i18n--envsubst: retire unused string_list_member()\n\nFabian Ruch (1):\n add: use struct argv_array in run_add_interactive()\n\nFelipe Contreras (10):\n transport-helper: mismerge fix\n transport-helper: don't update refs in dry-run\n transport-helper: add 'force' to 'export' helpers\n transport-helper: check for 'forced update' message\n remote-helpers: allow all tests running from any dir\n remote-hg: always normalize paths\n remote-bzr: add support for older versions\n completion: fix completing args of aliased \"push\", \"fetch\", etc.\n remote-bzr: trivial test fix\n prompt: fix missing file errors in zsh\n\nHiroyuki Sano (1):\n fsck: use bitwise-or assignment operator to set flag\n\nIlya Bobyr (1):\n rev-parse --parseopt: option argument name hints\n\nJacopo Notarstefano (2):\n git-bisect.sh: fix a few style issues\n branch.c: delete size check of newly tracked branch names\n\nJeff King (43):\n pack-objects: split add_object_entry\n repack: stop using magic number for ARRAY_SIZE(exts)\n repack: turn exts array into array-of-struct\n repack: handle optional files created by pack-objects\n t: add basic bitmap functionality tests\n t/perf: add tests for pack bitmaps\n cat-file: refactor error handling of batch_objects\n cat-file: fix a minor memory leak in batch_objects\n do not discard revindex when re-preparing packfiles\n block-sha1: factor out get_be and put_be wrappers\n read-cache: use get_be32 instead of hand-rolled ntoh_l\n tests: auto-set git-daemon port\n ewah: unconditionally ntohll ewah data\n tests: turn on network daemon tests by default\n http: never use curl_easy_perform\n config: disallow relative include paths from blobs\n docs: clarify remote restrictions for git-upload-archive\n CodingGuidelines: mention C whitespace rules\n repack: add `repack.packKeptObjects` config var\n docs: mark info/grafts as outdated\n match_explicit: hoist refspec lhs checks into their own function\n match_explicit_lhs: allow a \"verify only\" mode\n push: detect local refspec errors early\n cat-file: restore warn_on_object_refname_ambiguity flag\n rev-list: disable object/refname ambiguity check with --stdin\n pack-objects: show progress for reused packfiles\n pack-objects: show reused packfile objects in \"Counting objects\"\n pack-objects: turn off bitmaps when skipping objects\n subtree: initialize \"prefix\" variable\n t/Makefile: stop setting GIT_CONFIG\n t/test-lib: drop redundant unset of GIT_CONFIG\n t: drop useless sane_unset GIT_* calls\n t: stop using GIT_CONFIG to cross repo boundaries\n t: prefer \"git config --file\" to GIT_CONFIG with test_must_fail\n t: prefer \"git config --file\" to GIT_CONFIG\n t0001: make symlink reinit test more careful\n t0001: use test_path_is_*\n t0001: use test_config_global\n t0001: use test_must_fail\n t0001: drop useless subshells\n t0001: drop subshells just for \"cd\"\n pack-objects: do not reuse packfiles without --delta-base-offset\n config.c: mark die_bad_number as NORETURN\n\nJens Lehmann (2):\n submodule: don't access the .gitmodules cache entry after removing it\n submodule update: consistently document the '--checkout' option\n\nJohan Herland (1):\n notes: disallow reusing non-blob as a note object\n\nJohannes Sixt (11):\n t0008: skip trailing space test on Windows\n userdiff: support C++ ->* and .* operators in the word regexp\n userdiff: support unsigned and long long suffixes of integer constants\n t4018: an infrastructure to test hunk headers\n t4018: convert perl pattern tests to the new infrastructure\n t4018: convert java pattern test to the new infrastructure\n t4018: convert custom pattern test to the new infrastructure\n t4018: reduce test files for pattern compilation tests\n t4018: test cases for the built-in cpp pattern\n t4018: test cases showing that the cpp pattern misses many anchor points\n userdiff: have 'cpp' hunk header pattern catch more C++ anchor points\n\nJohn Keeping (4):\n notes-utils: handle boolean notes.rewritemode correctly\n utf8: fix iconv error detection\n utf8: use correct type for values in interval table\n builtin/mv: don't use memory after free\n\nJohn Marshall (1):\n stash doc: mention short form -k in save description\n\nJonathan Nieder (3):\n am doc: add a pointer to relevant hooks\n .gitignore: test-hashmap is a generated file\n test-hashmap.c: drop unnecessary #includes\n\nJunio C Hamano (35):\n git add ... defaults to \"-A\"\n git add: -u/-A now affects the entire working tree\n core.statinfo: remove as promised in Git 2.0\n push: switch default from \"matching\" to \"simple\"\n diff: remove \"diff-files -q\" in a version of Git in a distant future\n push: switch default from \"matching\" to \"simple\"\n t3004: add test for ls-files on symlinks via absolute paths\n open_istream(): do not dereference NULL in the error case\n combine-diff: simplify intersect_paths() further\n commit-tree: add and document --no-gpg-sign\n request-pull: pick up tag message as before\n request-pull: test updates\n request-pull: resurrect \"pretty refname\" feature\n *.sh: drop useless use of \"env\"\n tag: grok \"--with\" as synonym to \"--contains\"\n Start preparing for Git 2.0\n request-pull: documentation updates\n Update draft release notes to Git 2.0\n Update draft release notes to Git 2.0\n Update draft release notes to 2.0\n t1502: protect runs of SPs used in the indentation\n parse-options: multi-word argh should use dash to separate words\n update-index: teach --cacheinfo a new syntax \"mode,sha1,path\"\n parse-options: make sure argh string does not have SP or _\n Update draft release notes to 2.0\n Update draft release notes to 2.0\n parse-options: add cast to correct pointer type to OPT_SET_PTR\n Update draft release notes to 2.0\n Revert \"Merge branch 'wt/doc-submodule-name-path-confusion-2'\"\n Revert \"submodule: explicit local branch creation in module_clone\"\n Revert part of 384364b (Start preparing for Git 2.0, 2014-03-07)\n Update draft release notes to 2.0\n Update draft release notes to 2.0\n Update draft release notes for 2.0\n Git 2.0-rc0\n\nKarsten Blees (14):\n add a hashtable implementation that supports O(1) removal\n buitin/describe.c: use new hash map implementation\n diffcore-rename.c: move code around to prepare for the next patch\n diffcore-rename.c: simplify finding exact renames\n diffcore-rename.c: use new hash map implementation\n name-hash.c: use new hash map implementation for directories\n name-hash.c: remove unreferenced directory entries\n name-hash.c: use new hash map implementation for cache entries\n name-hash.c: remove cache entries instead of marking them CE_UNHASHED\n remove old hash.[ch] implementation\n fix 'git update-index --verbose --again' output\n builtin/update-index.c: cleanup update_one\n read-cache.c: fix memory leaks caused by removed cache entries\n hashmap.h: use 'unsigned int' for hash-codes everywhere\n\nKirill A. Shutemov (3):\n builtin/config.c: rename check_blob_write() -> check_write()\n config: change git_config_with_options() interface\n config: teach \"git config --file -\" to read from the standard input\n\nKirill Smelkov (10):\n tree-diff: allow diff_tree_sha1 to accept NULL sha1\n tree-diff: convert diff_root_tree_sha1() to just call diff_tree_sha1 with old=NULL\n line-log: convert to using diff_tree_sha1()\n revision: convert to using diff_tree_sha1()\n tree-walk: finally switch over tree descriptors to contain a pre-parsed entry\n diffcore-order: export generic ordering interface\n diff test: add tests for combine-diff with orderfile\n combine-diff: optimize combine_diff_path sets intersection\n combine-diff: combine_diff_path.len is not needed anymore\n tests: add checking that combine-diff emits only correct paths\n\nKyle J. McKay (2):\n test: fix t7001 cp to use POSIX options\n test: fix t5560 on FreeBSD\n\nLars Gullik Bjønnes (1):\n git-contacts: do not fail parsing of good diffs\n\nLinus Torvalds (2):\n request-pull: more strictly match local/remote branches\n request-pull: allow \"local:remote\" to specify names on both ends\n\nMarat Radchenko (5):\n MSVC: allow linking with the cURL library\n MSVC: link in invalidcontinue.obj for better POSIX compatibility\n MSVC: fix t0040-parse-options crash\n parse-options: remove unused OPT_SET_PTR\n MSVC: allow using ExtUtils::MakeMaker\n\nMartin Erik Werner (5):\n t0060: add test for prefix_path on symlinks via absolute paths\n t0060: add test for prefix_path when path == work tree\n t0060: add tests for prefix_path when path begins with work tree\n setup: add abspath_part_inside_repo() function\n setup: don't dereference in-tree symlinks for absolute paths\n\nMax Horn (2):\n transport-helper.c: do not overwrite forced bit\n remote-hg: do not fail on invalid bookmarks\n\nMichael Haggerty (14):\n rename read_replace_refs to check_replace_refs\n replace_object: use struct members instead of an array\n find_pack_entry(): document last_found_pack\n sha1_file_name(): declare to return a const string\n sha1_file.c: document a bunch of functions defined in the file\n Add docstrings for lookup_replace_object() and do_lookup_replace_object()\n Document some functions defined in object.c\n cache_tree_find(): remove redundant checks\n cache_tree_find(): find the end of path component using strchrnul()\n cache_tree_find(): fix comment formatting\n cache_tree_find(): remove redundant check\n cache_tree_find(): remove early return\n cache_tree_find(): use path variable when passing over slashes\n git-multimail: update to version 1.0.0\n\nNguyễn Thái Ngọc Duy (24):\n count-objects: recognize .bitmap in garbage-checking\n t7101, t7014: rename test files to indicate what that file is for\n reset: support \"--mixed --intent-to-add\" mode\n daemon: move daemonize() to libgit.a\n gc: config option for running --auto in background\n dir: warn about trailing spaces in exclude patterns\n dir: ignore trailing spaces in exclude patterns\n wt-status.c: make cut_line[] const to shrink .data section a bit\n wt-status.c: move cut-line print code out to wt_status_add_cut_line\n use wildmatch() directly without fnmatch() wrapper\n Revert \"test-wildmatch: add \"perf\" command to compare wildmatch and fnmatch\"\n stop using fnmatch (either native or compat)\n actually remove compat fnmatch source code\n sha1_file: fix delta_stack memory leak in unpack_entry\n i18n: mark all progress lines for translation\n commit: add --cleanup=scissors\n tag: support --sort=\n strbuf: style fix -- top opening bracket on a separate line\n upload-pack: send shallow info over stdin to pack-objects\n connect.c: SP after \"}\", not TAB\n object.h: centralize object flag allocation\n log: add --show-linear-break to help see non-linear history\n gc --aggressive: make --depth configurable\n environment.c: fix constness for odb_pack_keep()\n\nNicolas Vigier (10):\n cherry-pick, revert: add the --gpg-sign option\n git-sh-setup.sh: add variable to use the stuck-long mode\n am: parse options in stuck-long mode\n am: add the --gpg-sign option\n rebase: remove useless arguments check\n rebase: don't try to match -M option\n rebase: parse options in stuck-long mode\n rebase: add the --gpg-sign option\n commit-tree: add the commit.gpgsign option to sign all commits\n test the commit.gpgsign config option\n\nRalf Thielow (1):\n help.c: rename function \"pretty_print_string_list\"\n\nRené Scharfe (13):\n t7810: add missing variables to tests in loop\n grep: support -h (no header) with --count\n t4209: set up expectations up front\n t4209: factor out helper function test_log()\n t4209: factor out helper function test_log_icase()\n t4209: use helper functions to test --grep\n t4209: use helper functions to test --author\n pickaxe: honor -i when used with -S and --pickaxe-regex\n pickaxe: merge diffcore_pickaxe_grep() and diffcore_pickaxe_count() into diffcore_pickaxe()\n pickaxe: move pickaxe() after pickaxe_match()\n pickaxe: call strlen only when necessary in diffcore_pickaxe_count()\n pickaxe: simplify kwset loop in contains()\n rev-parse: fix typo in example on manpage\n\nRichard Hansen (2):\n test-hg.sh: tests are now expected to pass\n remote-bzr: support the new 'force' option\n\nRichard Lowe (1):\n diffcore.h: be explicit about the signedness of is_binary\n\nRoberto Tyley (1):\n Documentation: fix documentation AsciiDoc links for external urls\n\nRohit Mani (1):\n use strchrnul() in place of strchr() and strlen()\n\nScott J. Goldman (1):\n add uploadarchive.allowUnreachable option\n\nSebastian Schuberth (1):\n t5510: Do not use $(pwd) when fetching / pushing / pulling via rsync\n\nSimon Ruderich (2):\n git-config: document interactive.singlekey requires Term::ReadKey\n git-add--interactive: warn if module for interactive.singlekey is missing\n\nSun He (3):\n write_pack_file: use correct variable in diagnostic\n finish_tmp_packfile():use strbuf for pathname construction\n Use hashcpy() when copying object names\n\nSup Yut Sum (1):\n completion: teach --recurse-submodules to fetch, pull and push\n\nTanay Abhra (1):\n commit.c: use skip_prefix() instead of starts_with()\n\nTay Ray Chuan (1):\n demonstrate git-commit --dry-run exit code behaviour\n\nThomas Gummerer (3):\n introduce GIT_INDEX_VERSION environment variable\n test-lib: allow setting the index format version\n read-cache: add index.version config variable\n\nTorsten Bögershausen (1):\n utf8.c: partially update to version 6.3\n\nVicent Marti (16):\n revindex: export new APIs\n pack-objects: refactor the packing list\n pack-objects: factor out name_hash\n revision: allow setting custom limiter function\n sha1_file: export `git_open_noatime`\n compat: add endianness helpers\n ewah: compressed bitmap implementation\n documentation: add documentation for the bitmap format\n pack-bitmap: add support for bitmap indexes\n pack-objects: use bitmaps when packing objects\n rev-list: add bitmap mode to speed up object lists\n pack-objects: implement bitmap writing\n repack: consider bitmaps when performing repacks\n pack-bitmap: implement optional name_hash cache\n ewah: support platforms that require aligned reads\n add `ignore_missing_links` mode to revwalk\n\nVlad Dogaru (1):\n git-p4: explicitly specify that HEAD is a revision\n\nW. Trevor King (6):\n submodule: make 'checkout' update_module mode more explicit\n submodule: document module_clone arguments in comments\n submodule: explicit local branch creation in module_clone\n Documentation: describe 'submodule update --remote' use case\n doc: submodule.* config are keyed by submodule names\n doc: submodule.*.branch config is keyed by name\n\nYuxuan Shui (2):\n fsck.c:fsck_ident(): ident points at a const string\n fsck.c:fsck_commit(): use skip_prefix() to verify and skip constant\n\nbrian m. carlson (1):\n pull: add the --gpg-sign option.\n\ndequis (1):\n remote-bzr: include authors field in pushed commits\n\nДилян Палаузов (1):\n Makefile: describe CHARSET_LIB better\n"},{"id":"239095","messageId":"CALKQrgdv3xN78dvFNzR7K7dWMP-brG7r-OoHHMAus4pskmizmw@mail.gmail.com","threadId":"36437","inReplyTo":"xmqqk3ambf9k.fsf@gitster.dls.corp.google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Johan Herland","fromEmail":"johan@herland.net","sentAt":"2014-04-19T01:13:18Z","receivedAt":"2014-04-19T01:13:18Z","isPatch":false,"sender":{"key":"johan@herland.net","avatar":"https://avatars.githubusercontent.com/u/547031?v=4"},"body":"On Fri, Apr 18, 2014 at 9:37 PM, Junio C Hamano wrote:\n> An early preview release Git v2.0.0-rc0 is now available for\n> testing at the usual places.\n\nThis is supposed to have _all_ the v2.0 topics, correct?\n\nI'm unable to find the commit that actually _changes_ the default\nprefix for \"git svn\" (as announced in Documentation/git-svn.txt and\nthe release notes for v1.8.5 and v1.9.0).\n\nFor reference, it was posted as patch 3/3 back in October:\nhttp://thread.gmane.org/gmane.comp.version-control.git/232761/focus=235900\n\nVery sorry for not discovering this earlier.\n\n...Johan\n\n-- \nJohan Herland, \nwww.herland.net\n"},{"id":"239104","messageId":"7vwqellr99.fsf@alter.siamese.dyndns.org","threadId":"36437","inReplyTo":"CALKQrgdv3xN78dvFNzR7K7dWMP-brG7r-OoHHMAus4pskmizmw@mail.gmail.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-19T07:20:34Z","receivedAt":"2014-04-19T07:20:34Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Johan Herland writes:\n\n> On Fri, Apr 18, 2014 at 9:37 PM, Junio C Hamano wrote:\n>> An early preview release Git v2.0.0-rc0 is now available for\n>> testing at the usual places.\n>\n> This is supposed to have _all_ the v2.0 topics, correct?\n>\n> I'm unable to find the commit that actually _changes_ the default\n> prefix for \"git svn\" (as announced in Documentation/git-svn.txt and\n> the release notes for v1.8.5 and v1.9.0).\n\nYes, I noticed that the topic has been in the release notes for a\nfew cycles but the changes never came to my tree (perhaps review of\nthe patch series never concluded?) at the beginning of this cycle,\nso dropped it from the release notes.\n\n> For reference, it was posted as patch 3/3 back in October:\n> http://thread.gmane.org/gmane.comp.version-control.git/232761/focus=235900\n>\n> Very sorry for not discovering this earlier.\n\nWell, things happen X-<.\n\nI see the other two contained in the merge from Eric at 1668b7d78f81\n(Merge git://git.bogomips.org/git-svn, 2013-10-16). Is the last one\nstill viable? From my point of view, it would be best if Eric can\ntake another look on it and throw me a pull request (and I have to\nremember to resurrect the entry in the release notes).\n\nAlternatively I could revert the \"Warn about changing the default\"\nfor now and defer the topic to v2.1.\n\nEric, what do you think? My preference is not to revert but at the\nsame time I am hesitant to take a patch that was posted as RFC this\nlate in the cycle without input from the area expert, so...\n"},{"id":"239164","messageId":"53541b1214393_1f08dd12ece5@nysa.notmuch","threadId":"36437","inReplyTo":"xmqqk3ambf9k.fsf@gitster.dls.corp.google.com","subject":"RE: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Felipe Contreras","fromEmail":"felipe.contreras@gmail.com","sentAt":"2014-04-20T19:08:02Z","receivedAt":"2014-04-20T19:08:02Z","isPatch":false,"sender":{"key":"felipe.contreras@gmail.com","avatar":"https://avatars.githubusercontent.com/u/8358?v=4"},"body":"Junio C Hamano wrote:\n> * transport-helper, fast-import and fast-export have been updated to\n> allow the ref mapping and ref deletion in a way similar to the\n> natively supported transports.\n\nHave they?\n\n% git --version\ngit version 2.0.0.rc0\n% git push origin origin master:foo /tmp/test[master] nysa\nsearching for changes\nno changes found\nfatal: remote-helpers do not support old:new syntax\nERROR: unhandled export command:\n\nI think you are missing the patches which I just resent[1].\n\n[1] http://article.gmane.org/gmane.comp.version-control.git/246558\n\n-- \nFelipe Contreras\n"},{"id":"239182","messageId":"7vk3ajln3p.fsf@alter.siamese.dyndns.org","threadId":"36437","inReplyTo":"53541b1214393_1f08dd12ece5@nysa.notmuch","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-20T21:14:50Z","receivedAt":"2014-04-20T21:14:50Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Felipe Contreras writes:\n\n> Junio C Hamano wrote:\n>> * transport-helper, fast-import and fast-export have been updated to\n>> allow the ref mapping and ref deletion in a way similar to the\n>> natively supported transports.\n>\n> Have they?\n\nYikes, no. This was me mischaracterizingg the merge at 90e6255a6d01\n(Merge branch 'fc/transport-helper-fixes', 2014-03-18). It was\nabout letting the transport respond to 'force', and not about ref\nmapping, I think.\n\nThanks for spotting.\n\nWill remove that section from the release notes in the meantime. As\nto the resurrection of the series, that will have to cook outside\n'master' at least for the remainder of this cycle.\n"},{"id":"239327","messageId":"475e137b5095e45c92a87a9969f58f0@74d39fa044aa309eaea14b9f57fe79c","threadId":"36437","inReplyTo":"xmqqk3ambf9k.fsf@gitster.dls.corp.google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Kyle J. McKay","fromEmail":"mackyle@gmail.com","sentAt":"2014-04-22T11:16:22Z","receivedAt":"2014-04-22T11:16:22Z","isPatch":false,"sender":{"key":"mackyle@gmail.com","avatar":"https://avatars.githubusercontent.com/u/813346?v=4"},"body":"On Apr 18, 2014, at 12:37, Junio C Hamano wrote:\n> An early preview release Git v2.0.0-rc0 is now available for\n> testing at the usual places.\n\nI have run into the following test failures with v2.0.0-rc0:\n\nTest Summary Report\n-------------------\nt9117-git-svn-init-clone.sh (Wstat: 256 Tests: 11 Failed: 2)\n Failed tests: 10-11\n Non-zero exit status: 1\nFiles=658, Tests=11878, 3684 wallclock secs\nResult: FAIL\n\nThe cause of the failure in both tests is this:\n\n $ git svn init -s \"$svnrepo\"/project project --prefix=\"\"\n Option prefix requires an argument\n\nThe problem with --prefix=\"\" is this (from the Getopt::Long CHANGES file):\n\nChanges in version 2.37\n-----------------------\n\n * Bugfix: With gnu_compat, --foo= will no longer trigger \"Option\n requires an argument\" but return the empty string.\n\nThe system I ran the tests on happens to have Getopt::Long version 2.35.\n\nThe --prefix=\"\" option can be rewritten --prefix \"\" in both tests and then \nthey pass.\n\nAlternatively this change can be made in git-svn.perl:\n\n|diff --git a/git-svn.perl b/git-svn.perl\n|index 7349ffea..284f458a 100755\n|--- a/git-svn.perl\n|+++ b/git-svn.perl\n|@@ -149,7 +149,7 @@ my ($_trunk, @_tags, @_branches, $_stdlayout);\n my %icv;\n my %init_opts = ( 'template=s' => \\$_template, 'shared:s' => \\$_shared,\n 'trunk|T=s' => \\$_trunk, 'tags|t=s@' => \\@_tags,\n- 'branches|b=s@' => \\@_branches, 'prefix=s' => \\$_prefix,\n+ 'branches|b=s@' => \\@_branches, 'prefix:s' => \\$_prefix,\n 'stdlayout|s' => \\$_stdlayout,\n 'minimize-url|m!' => \\$Git::SVN::_minimize_url,\n 'no-metadata' => sub { $icv{noMetadata} = 1 },\n\nWhich makes the argument to --prefix optional (in which case it is set \nto \"\").\n\nI'm not really sure what the best thing to do here is. I thought the \ndocumentation talked somewhere about using --prefix=\"\" (or just --prefix=)\nto get the old behavior, but now I can't find that. In that \ncase I think probably just changing the tests to use --prefix \"\" \ninstead of --prefix=\"\" should take care of it especially since the log \nmessage for fe191fca (which actually changes the default) talks about \nusing --prefix \"\" and not --prefix=\"\". I have attached a patch below \n(against master) to make that change to the two affected subtests of t9117.\n\n--Kyle\n\n---- >8 ----\nSubject: [PATCH] t9117: use --prefix \"\" instead of --prefix=\"\"\n\nVersions of Perl's Getopt::Long module before 2.37 do not contain\nthis fix that first appeared in Getopt::Long version 2.37:\n\n* Bugfix: With gnu_compat, --foo= will no longer trigger \"Option\n requires an argument\" but return the empty string.\n\nInstead of using --prefix=\"\" use --prefix \"\" when testing an\nexplictly empty prefix string in order to work with older versions\nof Perl's Getopt::Long module.\n\nSigned-off-by: Kyle J. McKay \n---\n t/t9117-git-svn-init-clone.sh | 8 ++++----\n 1 file changed, 4 insertions(+), 4 deletions(-)\n\ndiff --git a/t/t9117-git-svn-init-clone.sh b/t/t9117-git-svn-init-clone.sh\nindex dfed76fe..a66f43c6 100755\n--- a/t/t9117-git-svn-init-clone.sh\n+++ b/t/t9117-git-svn-init-clone.sh\n@@ -101,18 +101,18 @@ test_expect_success 'clone with -s/-T/-b/-t assumes --prefix=origin/' '\n \trm -f warning\n \t'\n \n-test_expect_success 'init with -s/-T/-b/-t and --prefix=\"\" still works' '\n+test_expect_success 'init with -s/-T/-b/-t and --prefix \"\" still works' '\n \ttest ! -d project &&\n-\tgit svn init -s \"$svnrepo\"/project project --prefix=\"\" 2>warning &&\n+\tgit svn init -s \"$svnrepo\"/project project --prefix \"\" 2>warning &&\n \ttest_must_fail grep -q prefix warning &&\n \ttest_svn_configured_prefix \"\" &&\n \trm -rf project &&\n \trm -f warning\n \t'\n \n-test_expect_success 'clone with -s/-T/-b/-t and --prefix=\"\" still works' '\n+test_expect_success 'clone with -s/-T/-b/-t and --prefix \"\" still works' '\n \ttest ! -d project &&\n-\tgit svn clone -s \"$svnrepo\"/project --prefix=\"\" 2>warning &&\n+\tgit svn clone -s \"$svnrepo\"/project --prefix \"\" 2>warning &&\n \ttest_must_fail grep -q prefix warning &&\n \ttest_svn_configured_prefix \"\" &&\n \trm -rf project &&\n-- \n1.8.5\n"},{"id":"239367","messageId":"20140422185829.GB15516@google.com","threadId":"36437","inReplyTo":"475e137b5095e45c92a87a9969f58f0@74d39fa044aa309eaea14b9f57fe79c","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2014-04-22T18:58:29Z","receivedAt":"2014-04-22T18:58:29Z","isPatch":false,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Kyle J. McKay wrote:\n\n> The problem with --prefix=\"\" is this (from the Getopt::Long CHANGES file):\n>\n> Changes in version 2.37\n> -----------------------\n>\n> * Bugfix: With gnu_compat, --foo= will no longer trigger \"Option\n> requires an argument\" but return the empty string.\n>\n> The system I ran the tests on happens to have Getopt::Long version 2.35.\n\nThanks for catching it.\n\nGetopt::Long was updated to 2.37 in perl 5.8.9 (in 5.8.8 it was\nupdated to 2.35). INSTALL still only recommends Perl 5.8 so that's an\nissue.\n\n> The --prefix=\"\" option can be rewritten --prefix \"\" in both tests and then \n> they pass.\n\nHm, perhaps we should introduce a 'no-prefix' option to work around\nthis.\n\n> |diff --git a/git-svn.perl b/git-svn.perl\n> |index 7349ffea..284f458a 100755\n> |--- a/git-svn.perl\n> |+++ b/git-svn.perl\n> |@@ -149,7 +149,7 @@ my ($_trunk, @_tags, @_branches, $_stdlayout);\n> my %icv;\n> my %init_opts = ( 'template=s' => \\$_template, 'shared:s' => \\$_shared,\n> 'trunk|T=s' => \\$_trunk, 'tags|t=s@' => \\@_tags,\n> 'branches|b=s@' => \\@_branches, 'prefix=s' => \\$_prefix,\n +\t\t 'no-prefix' => sub { $_prefix = \"\" },\n\n> 'stdlayout|s' => \\$_stdlayout,\n> 'minimize-url|m!' => \\$Git::SVN::_minimize_url,\n> 'no-metadata' => sub { $icv{noMetadata} = 1 },\n\nThat way, normal usage of --prefix would still be consistent with\nother git commands that prefer the form with argument attached\n(--prefix=foo, not --prefix foo; see gitcli(7)).\n\nThoughts?\nJonathan\n"},{"id":"239377","messageId":"xmqqbnvt3zp8.fsf@gitster.dls.corp.google.com","threadId":"36437","inReplyTo":"475e137b5095e45c92a87a9969f58f0@74d39fa044aa309eaea14b9f57fe79c","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-22T19:56:51Z","receivedAt":"2014-04-22T19:56:51Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"\"Kyle J. McKay\" writes:\n\n> Alternatively this change can be made in git-svn.perl:\n> |diff --git a/git-svn.perl b/git-svn.perl\n> |index 7349ffea..284f458a 100755\n> |--- a/git-svn.perl\n> |+++ b/git-svn.perl\n> |@@ -149,7 +149,7 @@ my ($_trunk, @_tags, @_branches, $_stdlayout);\n> my %icv;\n> my %init_opts = ( 'template=s' => \\$_template, 'shared:s' => \\$_shared,\n> 'trunk|T=s' => \\$_trunk, 'tags|t=s@' => \\@_tags,\n> - 'branches|b=s@' => \\@_branches, 'prefix=s' => \\$_prefix,\n> + 'branches|b=s@' => \\@_branches, 'prefix:s' => \\$_prefix,\n> 'stdlayout|s' => \\$_stdlayout,\n> 'minimize-url|m!' => \\$Git::SVN::_minimize_url,\n> 'no-metadata' => sub { $icv{noMetadata} = 1 },\n>\n> Which makes the argument to --prefix optional (in which case it is set \n> to \"\").\n\nWe do want to learn what prefix the user wants to use when they\nstart their command line with \"git svn --prefix\". Stopping the\ncommand line right there and expecting it to default to whatever\nbuilt-in prefix is simply strange; the user could have omitted that\n\"--prefix\" that lacks the argument. If the user did want us to use\nthe default by passing an empty string as its argument, both forms,\ni.e. \"--prefix ''\" and \"--prefix=''\", ought to be usable, but if\nolder Getopt::Long() does not allow the latter form, at least the\nformer is the right way for the user to tell us that.\n\nSo I think your fix (workaround for a bug in older Getopt::Long) in\nthe patch is the right one. Johan may want to help me by pointing\nout if I am missing something.\n\nThanks.\n\n\n> I'm not really sure what the best thing to do here is. I thought the \n> documentation talked somewhere about using --prefix=\"\" (or just --prefix=)\n> to get the old behavior, but now I can't find that. In that \n> case I think probably just changing the tests to use --prefix \"\" \n> instead of --prefix=\"\" should take care of it especially since the log \n> message for fe191fca (which actually changes the default) talks about \n> using --prefix \"\" and not --prefix=\"\". I have attached a patch below \n> (against master) to make that change to the two affected subtests of t9117.\n>\n> --Kyle\n>\n> ---- >8 ----\n> Subject: [PATCH] t9117: use --prefix \"\" instead of --prefix=\"\"\n>\n> Versions of Perl's Getopt::Long module before 2.37 do not contain\n> this fix that first appeared in Getopt::Long version 2.37:\n>\n> * Bugfix: With gnu_compat, --foo= will no longer trigger \"Option\n> requires an argument\" but return the empty string.\n>\n> Instead of using --prefix=\"\" use --prefix \"\" when testing an\n> explictly empty prefix string in order to work with older versions\n> of Perl's Getopt::Long module.\n>\n> Signed-off-by: Kyle J. McKay \n> ---\n> t/t9117-git-svn-init-clone.sh | 8 ++++----\n> 1 file changed, 4 insertions(+), 4 deletions(-)\n>\n> diff --git a/t/t9117-git-svn-init-clone.sh b/t/t9117-git-svn-init-clone.sh\n> index dfed76fe..a66f43c6 100755\n> --- a/t/t9117-git-svn-init-clone.sh\n> +++ b/t/t9117-git-svn-init-clone.sh\n> @@ -101,18 +101,18 @@ test_expect_success 'clone with -s/-T/-b/-t assumes --prefix=origin/' '\n> \trm -f warning\n> \t'\n> \n> -test_expect_success 'init with -s/-T/-b/-t and --prefix=\"\" still works' '\n> +test_expect_success 'init with -s/-T/-b/-t and --prefix \"\" still works' '\n> \ttest ! -d project &&\n> -\tgit svn init -s \"$svnrepo\"/project project --prefix=\"\" 2>warning &&\n> +\tgit svn init -s \"$svnrepo\"/project project --prefix \"\" 2>warning &&\n> \ttest_must_fail grep -q prefix warning &&\n> \ttest_svn_configured_prefix \"\" &&\n> \trm -rf project &&\n> \trm -f warning\n> \t'\n> \n> -test_expect_success 'clone with -s/-T/-b/-t and --prefix=\"\" still works' '\n> +test_expect_success 'clone with -s/-T/-b/-t and --prefix \"\" still works' '\n> \ttest ! -d project &&\n> -\tgit svn clone -s \"$svnrepo\"/project --prefix=\"\" 2>warning &&\n> +\tgit svn clone -s \"$svnrepo\"/project --prefix \"\" 2>warning &&\n> \ttest_must_fail grep -q prefix warning &&\n> \ttest_svn_configured_prefix \"\" &&\n> \trm -rf project &&\n"},{"id":"239378","messageId":"xmqq4n1l3yyf.fsf@gitster.dls.corp.google.com","threadId":"36437","inReplyTo":"20140422185829.GB15516@google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-22T20:12:56Z","receivedAt":"2014-04-22T20:12:56Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"The following message is a courtesy copy of an article\nthat has been posted to gmane.comp.version-control.git as well.\n\nJonathan Nieder writes:\n\n> Hm, perhaps we should introduce a 'no-prefix' option to work around\n> this.\n> ...\n>> |diff --git a/git-svn.perl b/git-svn.perl\n>> |index 7349ffea..284f458a 100755\n>> |--- a/git-svn.perl\n>> |+++ b/git-svn.perl\n>> |@@ -149,7 +149,7 @@ my ($_trunk, @_tags, @_branches, $_stdlayout);\n>> my %icv;\n>> my %init_opts = ( 'template=s' => \\$_template, 'shared:s' => \\$_shared,\n>> 'trunk|T=s' => \\$_trunk, 'tags|t=s@' => \\@_tags,\n>> 'branches|b=s@' => \\@_branches, 'prefix=s' => \\$_prefix,\n> +\t\t 'no-prefix' => sub { $_prefix = \"\" },\n>\n>> 'stdlayout|s' => \\$_stdlayout,\n>> 'minimize-url|m!' => \\$Git::SVN::_minimize_url,\n>> 'no-metadata' => sub { $icv{noMetadata} = 1 },\n>\n> That way, normal usage of --prefix would still be consistent with\n> other git commands that prefer the form with argument attached\n> (--prefix=foo, not --prefix foo; see gitcli(7)).\n>\n> Thoughts?\n\nI do not think that it is a good idea to use \"--no-anything\" for\nsomething that is not a boolean.\n\nI can buy \"--old-default-prefix\", or \"--empty-prefix\", but running\n\"git svn --prefix ''\" (or \"--prefix=''\") would be OK and logically\nconsistent anyway (i.e. the option tells us what string to add after\n\"refs/remotes/\", and the old default that everybody hated were to\nuse an empty string), so...\n"},{"id":"239381","messageId":"20140422210002.GC15516@google.com","threadId":"36437","inReplyTo":"xmqq4n1l3yyf.fsf@gitster.dls.corp.google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2014-04-22T21:00:02Z","receivedAt":"2014-04-22T21:00:02Z","isPatch":false,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Junio C Hamano wrote:\n> Jonathan Nieder writes:\n\n>> Hm, perhaps we should introduce a 'no-prefix' option to work around\n>> this.\n[...]\n>> That way, normal usage of --prefix would still be consistent with\n>> other git commands that prefer the form with argument attached\n>> (--prefix=foo, not --prefix foo; see gitcli(7)).\n>>\n>> Thoughts?\n>\n> I do not think that it is a good idea to use \"--no-anything\" for\n> something that is not a boolean.\n\nDo you mean it is a bad idea to support or a bad idea to make use of\nsuch support?\n\nI suggested --no- for consistency with current git commands that use\nparseopt. But on second thought, I agree that it be confusing for\n\n\t--prefix=foo --no-prefix\n\nto mean something different from no --prefix parameter at all.\n\nThe documentation says\n\n\t--prefix=\n\n\t\t...\n\n\t\tBefore Git 2.0, the default prefix was \"\" (no prefix).\n\t\tThis meant that ...\n\nwhich suggests that I can use --prefix=\"\" to mean no prefix. Perhaps\nit needs a note to suggest using '--prefix \"\"' instead?\n\nThanks,\nJonathan\n"},{"id":"239385","messageId":"xmqqy4yx2gz9.fsf@gitster.dls.corp.google.com","threadId":"36437","inReplyTo":"20140422210002.GC15516@google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-22T21:26:34Z","receivedAt":"2014-04-22T21:26:34Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Jonathan Nieder writes:\n\n> Junio C Hamano wrote:\n>> Jonathan Nieder writes:\n>\n>>> Hm, perhaps we should introduce a 'no-prefix' option to work around\n>>> this.\n> [...]\n>>> That way, normal usage of --prefix would still be consistent with\n>>> other git commands that prefer the form with argument attached\n>>> (--prefix=foo, not --prefix foo; see gitcli(7)).\n>>>\n>>> Thoughts?\n>>\n>> I do not think that it is a good idea to use \"--no-anything\" for\n>> something that is not a boolean.\n>\n> Do you mean it is a bad idea to support or a bad idea to make use of\n> such support?\n>\n> I suggested --no- for consistency with current git commands that use\n> parseopt. But on second thought, I agree that it be confusing for\n>\n> \t--prefix=foo --no-prefix\n>\n> to mean something different from no --prefix parameter at all.\n>\n> The documentation says\n>\n> \t--prefix=\n>\n> \t\t...\n>\n> \t\tBefore Git 2.0, the default prefix was \"\" (no prefix).\n> \t\tThis meant that ...\n>\n> which suggests that I can use --prefix=\"\" to mean no prefix. Perhaps\n> it needs a note to suggest using '--prefix \"\"' instead?\n\nIs there another --option that takes an arbitrary user string that\ncould be an empty string (or will there be one in the future)? If\nthat is the case, a better alternative might be to add an comment to\nsay that those with older Getopt::Long may have to use --option \"\"\ninstead of the --option=\"\" form for any option whose value happens\nto be an empty string to work around the command parser bug.\n"},{"id":"239390","messageId":"20140422221148.GD15516@google.com","threadId":"36437","inReplyTo":"xmqqy4yx2gz9.fsf@gitster.dls.corp.google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2014-04-22T22:11:48Z","receivedAt":"2014-04-22T22:11:48Z","isPatch":false,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Junio C Hamano wrote:\n> Jonathan Nieder writes:\n\n>> The documentation says\n>>\n>> \t--prefix=\n>>\n>> \t\t...\n>>\n>> \t\tBefore Git 2.0, the default prefix was \"\" (no prefix).\n>> \t\tThis meant that ...\n>>\n>> which suggests that I can use --prefix=\"\" to mean no prefix. Perhaps\n>> it needs a note to suggest using '--prefix \"\"' instead?\n>\n> Is there another --option that takes an arbitrary user string that\n> could be an empty string (or will there be one in the future)?\n\nIn git in general, yes --- for example, 'git diff --src-prefix=\"\"\nHEAD^' tells \"git diff\" to leave off the usual c/ prefix in the\nsrc filename it prints.\n\nIn git-svn, --trunk=\"\" or --message=\"\" might be meaningful, but not\nnearly as much as --prefix=\"\".\n\n> If\n> that is the case, a better alternative might be to add an comment to\n> say that those with older Getopt::Long may have to use --option \"\"\n> instead of the --option=\"\" form for any option whose value happens\n> to be an empty string to work around the command parser bug.\n\nAnother possibility would be to require Perl 5.8.9 or newer. It was\nreleased in 2008.\n\ndiff --git i/git-svn.perl w/git-svn.perl\nindex 0a32372..ec7910d 100755\n--- i/git-svn.perl\n+++ w/git-svn.perl\n@@ -1,7 +1,7 @@\n #!/usr/bin/perl\n # Copyright (C) 2006, Eric Wong \n # License: GPL v2 or later\n-use 5.008;\n+use 5.008_009;\n use warnings;\n use strict;\n use vars qw/\t$AUTHOR $VERSION\n"},{"id":"239394","messageId":"20140422222519.GC144079@vauxhall.crustytoothpaste.net","threadId":"36437","inReplyTo":"20140422221148.GD15516@google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"brian m. carlson","fromEmail":"sandals@crustytoothpaste.net","sentAt":"2014-04-22T22:25:19Z","receivedAt":"2014-04-22T22:25:19Z","isPatch":false,"sender":{"key":"sandals@crustytoothpaste.net","avatar":"https://avatars.githubusercontent.com/u/497054?v=4"},"body":"On Tue, Apr 22, 2014 at 03:11:48PM -0700, Jonathan Nieder wrote:\n> Another possibility would be to require Perl 5.8.9 or newer. It was\n> released in 2008.\n\nRHEL 5 and CentOS 5 are still shipping with 5.8.8. They are still\nsecurity-supported until 2017, and believe it or not people still\ndevelop on them. I am personally fine with this change, though.\n\nWhat we could do instead is simply require a newer version of\nGetopt::Long, which would let people continue using their ancient OSes\nand install a newer version from CPAN if necessary. It's also the\nproper way to specify the dependency.\n\n-- \nbrian m. carlson / brian with sandals: Houston, Texas, US\n+1 832 623 2791 | http://www.crustytoothpaste.net/~bmc | My opinion only\nOpenPGP: RSA v4 4096b: 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187\n"},{"id":"239398","messageId":"xmqqbnvt2d7v.fsf@gitster.dls.corp.google.com","threadId":"36437","inReplyTo":"20140422222519.GC144079@vauxhall.crustytoothpaste.net","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-22T22:47:48Z","receivedAt":"2014-04-22T22:47:48Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"\"brian m. carlson\" writes:\n\n> On Tue, Apr 22, 2014 at 03:11:48PM -0700, Jonathan Nieder wrote:\n>> Another possibility would be to require Perl 5.8.9 or newer. It was\n>> released in 2008.\n>\n> RHEL 5 and CentOS 5 are still shipping with 5.8.8. They are still\n> security-supported until 2017, and believe it or not people still\n> develop on them. I am personally fine with this change, though.\n>\n> What we could do instead is simply require a newer version of\n> Getopt::Long, which would let people continue using their ancient OSes\n> and install a newer version from CPAN if necessary. It's also the\n> proper way to specify the dependency.\n\nYes, but if its inability to properly grok --option=\"\" is the only\nreason we want to add a dependency, wouldn't it suffice to simply\nstate in the documentation (1) how to recognise the symptom to see\nif the version the user has is too old, e.g. \"if you see this error\nmessage\", \"run 'perl -v' to see if your perl is older than X\",\netc. and (2) how to work it around, i.e. \"instead of giving an empty\nvalue with --option='', say --option ''\"?\n"},{"id":"239409","messageId":"CALKQrgd0Psp0nWUsmW16aR4io4xu_gumJmRdLGU+MOSzJvBobQ@mail.gmail.com","threadId":"36437","inReplyTo":"xmqqbnvt2d7v.fsf@gitster.dls.corp.google.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Johan Herland","fromEmail":"johan@herland.net","sentAt":"2014-04-23T07:33:40Z","receivedAt":"2014-04-23T07:33:40Z","isPatch":false,"sender":{"key":"johan@herland.net","avatar":"https://avatars.githubusercontent.com/u/547031?v=4"},"body":"On Wed, Apr 23, 2014 at 12:47 AM, Junio C Hamano wrote:\n> \"brian m. carlson\" writes:\n>> What we could do instead is simply require a newer version of\n>> Getopt::Long, which would let people continue using their ancient OSes\n>> and install a newer version from CPAN if necessary. It's also the\n>> proper way to specify the dependency.\n>\n> Yes, but if its inability to properly grok --option=\"\" is the only\n> reason we want to add a dependency, wouldn't it suffice to simply\n> state in the documentation (1) how to recognise the symptom to see\n> if the version the user has is too old, e.g. \"if you see this error\n> message\", \"run 'perl -v' to see if your perl is older than X\",\n> etc. and (2) how to work it around, i.e. \"instead of giving an empty\n> value with --option='', say --option ''\"?\n\nFWIW, the least intrusive approach is what I find most agreeable:\n\n - Fix the tests to use --prefix \"\" instead of --prefix=\"\"\n - Update the documentation like Junio suggests above.\n - Reformat an example using --prefix \"\"\n\nI.e. use Kyle's patch to t9117, plus something like this:\n\ndiff --git a/Documentation/git-svn.txt b/Documentation/git-svn.txt\nindex 5b3c38d..9f579e0 100644\n--- a/Documentation/git-svn.txt\n+++ b/Documentation/git-svn.txt\n@@ -91,6 +91,9 @@ COMMANDS\n NOTE: Before Git v2.0, the default prefix was \"\" (no prefix). This\n meant that SVN-tracking refs were put at \"refs/remotes/*\", which is\n incompatible with how Git's own remote-tracking refs are organized.\n+If you still want the old default, you can get it by passing\n+'--prefix \"\"' on the command line ('--prefix=\"\"' may not work if\n+your Perl's Getopt::Long is < v2.37).\n\n --ignore-paths=;;\n When passed to 'init' or 'clone' this regular expression will\n\n\n\n...Johan\n\n-- \nJohan Herland, \nwww.herland.net\n"},{"id":"239446","messageId":"xmqqtx9k0zjw.fsf@gitster.dls.corp.google.com","threadId":"36437","inReplyTo":"CALKQrgd0Psp0nWUsmW16aR4io4xu_gumJmRdLGU+MOSzJvBobQ@mail.gmail.com","subject":"Re: [ANNOUNCE] Git v2.0.0-rc0","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-04-23T16:40:35Z","receivedAt":"2014-04-23T16:40:35Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Johan Herland writes:\n\n> I.e. use Kyle's patch to t9117, plus something like this:\n>\n> diff --git a/Documentation/git-svn.txt b/Documentation/git-svn.txt\n> index 5b3c38d..9f579e0 100644\n> --- a/Documentation/git-svn.txt\n> +++ b/Documentation/git-svn.txt\n> @@ -91,6 +91,9 @@ COMMANDS\n> NOTE: Before Git v2.0, the default prefix was \"\" (no prefix). This\n> meant that SVN-tracking refs were put at \"refs/remotes/*\", which is\n> incompatible with how Git's own remote-tracking refs are organized.\n> +If you still want the old default, you can get it by passing\n> +'--prefix \"\"' on the command line ('--prefix=\"\"' may not work if\n> +your Perl's Getopt::Long is < v2.37).\n>\n> --ignore-paths=;;\n> When passed to 'init' or 'clone' this regular expression will\n\nThanks, will squash in.\n"}]}