{"thread":{"id":"35776","subject":"[ANNOUNCE] Git v1.9-rc2","startedAt":"2014-01-31T23:03:23Z","lastAt":"2014-01-31T23:03:23Z","messageCount":1,"participants":["Junio C Hamano"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"234000","messageId":"xmqqlhxvhhz8.fsf@gitster.dls.corp.google.com","threadId":"35776","inReplyTo":null,"subject":"[ANNOUNCE] Git v1.9-rc2","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-01-31T23:03:23Z","receivedAt":"2014-01-31T23:03:23Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"A release candidate Git v1.9-rc2 is now available for testing at\nthe usual places.\n\nI've heard rumours that various third-party tools do not like the\ntwo-digit version numbers (e.g. \"Git 2.0\") and started barfing left\nand right when the users install v1.9-rc1.  While it is tempting to\nlaugh at them for their sloppy assumption, I am also practical and\ndo not mind calling the upcoming release v1.9.0 to help them.\n\nIf we go that route (and I am inclined to go that route at this\nmoment), the versioning scheme will be\n\n - The next release candidate will be v1.9.0-rc3, not v1.9-rc3;\n\n - The first maintenance release for v1.9.0 will be v1.9.1 (and Nth one\n   be v1.9.N); and\n\n - The feature release after v1.9.0 will be either v1.10.0 or\n   v2.0.0, depending on how big the feature jump we are looking at.\n\nThe release tarballs are found at:\n\n    http://code.google.com/p/git-core/downloads/list\n\nand their SHA-1 checksums are:\n\nc21bb5f8a138a5af9d330ce7289f5c533d49e2a2  git-1.9.rc2.tar.gz\ndec30e4b187c84c08c02d53aa5d5d4d3fae9bbda  git-htmldocs-1.9.rc2.tar.gz\neeee207e1e91357eed351b5b7cd4a7d2cc8985cd  git-manpages-1.9.rc2.tar.gz\n\nThe following public repositories all have a copy of the v1.9-rc2\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 v1.9 Release Notes (draft)\n======================\n\nBackward compatibility notes\n----------------------------\n\n\"git submodule foreach $cmd $args\" used to treat \"$cmd $args\" the same\nway \"ssh\" did, concatenating them into a single string and letting the\nshell unquote. Careless users who forget to sufficiently quote $args\ngets their argument split at $IFS whitespaces by the shell, and got\nunexpected results due to this. Starting from this release, the\ncommand line is passed directly to the shell, if it has an argument.\n\nRead-only support for experimental loose-object format, in which users\ncould optionally choose to write in their loose objects for a short\nwhile between v1.4.3 to v1.5.3 era, has been dropped.\n\nThe meanings of \"--tags\" option to \"git fetch\" has changed; the\ncommand fetches tags _in addition to_ what are fetched by the same\ncommand line without the option.\n\nThe way \"git push $there $what\" interprets $what part given on the\ncommand line, when it does not have a colon that explicitly tells us\nwhat ref at the $there repository is to be updated, has been enhanced.\n\nA handful of ancient commands that have long been deprecated are\nfinally gone (repo-config, tar-tree, lost-found, and peek-remote).\n\n\nBackward compatibility notes (for Git 2.0)\n------------------------------------------\n\nWhen \"git push [$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 will change to the \"simple\"\nsemantics, which 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\nUse the user preference configuration variable \"push.default\" to\nchange this.  If you are an old-timer who is used to the \"matching\"\nsemantics, you can set the variable to \"matching\" to keep the\ntraditional behaviour.  If you want to live in the future early, you\ncan set it to \"simple\" today without waiting for Git 2.0.\n\nWhen \"git add -u\" (and \"git add -A\") is run inside a subdirectory and\ndoes not specify which paths to add on the command line, it\nwill operate on the entire tree in Git 2.0 for consistency\nwith \"git commit -a\" and other commands.  There will be no\nmechanism to make plain \"git add -u\" behave like \"git add -u .\".\nCurrent users of \"git add -u\" (without a pathspec) should start\ntraining their fingers to explicitly say \"git add -u .\"\nbefore Git 2.0 comes.  A warning is issued when these commands are\nrun without a pathspec and when you have local changes outside the\ncurrent directory, because the behaviour in Git 2.0 will be different\nfrom today's version in such a situation.\n\nIn Git 2.0, \"git add <path>\" will behave as \"git add -A <path>\", so\nthat \"git add dir/\" will notice paths you removed from the directory\nand record the removal.  Versions before Git 2.0, including this\nrelease, will keep ignoring removals, but the users who rely on this\nbehaviour are encouraged to start using \"git add --ignore-removal <path>\"\nnow before 2.0 is released.\n\nThe default prefix for \"git svn\" will change in Git 2.0.  For a long\ntime, \"git svn\" created its remote-tracking branches directly under\nrefs/remotes, but it will place them under refs/remotes/origin/ unless\nit is told otherwise with its --prefix option.\n\n\nUpdates since v1.8.5\n--------------------\n\nForeign interfaces, subsystems and ports.\n\n * The HTTP transport, when talking GSS-Negotiate, uses \"100\n   Continue\" response to avoid having to rewind and resend a large\n   payload, which may not be always doable.\n\n * Various bugfixes to remote-bzr and remote-hg (in contrib/).\n\n * The build procedure is aware of MirBSD now.\n\n * Various \"git p4\", \"git svn\" and \"gitk\" updates.\n\n\nUI, Workflows & Features\n\n * Fetching from a shallowly-cloned repository used to be forbidden,\n   primarily because the codepaths involved were not carefully vetted\n   and we did not bother supporting such usage. This release attempts\n   to allow object transfer out of a shallowly-cloned repository in a\n   more controlled way (i.e. the receiver become a shallow repository\n   with a truncated history).\n\n * Just like we give a reasonable default for \"less\" via the LESS\n   environment variable, we now specify a reasonable default for \"lv\"\n   via the \"LV\" environment variable when spawning the pager.\n\n * Two-level configuration variable names in \"branch.*\" and \"remote.*\"\n   hierarchies, whose variables are predominantly three-level, were\n   not completed by hitting a <TAB> in bash and zsh completions.\n\n * Fetching 'frotz' branch with \"git fetch\", while 'frotz/nitfol'\n   remote-tracking branch from an earlier fetch was still there, would\n   error out, primarily because the command was not told that it is\n   allowed to lose any information on our side.  \"git fetch --prune\"\n   now can be used to remove 'frotz/nitfol' to make room to fetch and\n   store 'frotz' remote-tracking branch.\n\n * \"diff.orderfile=<file>\" configuration variable can be used to\n   pretend as if the \"-O<file>\" option were given from the command\n   line of \"git diff\", etc.\n\n * The negative pathspec syntax allows \"git log -- . ':!dir'\" to tell\n   us \"I am interested in everything but 'dir' directory\".\n\n * \"git difftool\" shows how many different paths there are in total,\n   and how many of them have been shown so far, to indicate progress.\n\n * \"git push origin master\" used to push our 'master' branch to update\n   the 'master' branch at the 'origin' repository.  This has been\n   enhanced to use the same ref mapping \"git push origin\" would use to\n   determine what ref at the 'origin' to be updated with our 'master'.\n   For example, with this configuration\n\n   [remote \"origin\"]\n      push = refs/heads/*:refs/review/*\n\n   that would cause \"git push origin\" to push out our local branches\n   to corresponding refs under refs/review/ hierarchy at 'origin',\n   \"git push origin master\" would update 'refs/review/master' over\n   there.  Alternatively, if push.default is set to 'upstream' and our\n   'master' is set to integrate with 'topic' from the 'origin' branch,\n   running \"git push origin\" while on our 'master' would update their\n   'topic' branch, and running \"git push origin master\" while on any\n   of our branches does the same.\n\n * \"gitweb\" learned to treat ref hierarchies other than refs/heads as\n   if they are additional branch namespaces (e.g. refs/changes/ in\n   Gerrit).\n\n * \"git for-each-ref --format=...\" learned a few formatting directives;\n   e.g. \"%(color:red)%(HEAD)%(color:reset) %(refname:short) %(subject)\".\n\n * The command string given to \"git submodule foreach\" is passed\n   directly to the shell, without being eval'ed.  This is a backward\n   incompatible change that may break existing users.\n\n * \"git log\" and friends learned the \"--exclude=<glob>\" option, to\n   allow people to say \"list history of all branches except those that\n   match this pattern\" with \"git log --exclude='*/*' --branches\".\n\n * \"git rev-parse --parseopt\" learned a new \"--stuck-long\" option to\n   help scripts parse options with an optional parameter.\n\n * The \"--tags\" option to \"git fetch\" no longer tells the command to\n   fetch _only_ the tags. It instead fetches tags _in addition to_\n   what are fetched by the same command line without the option.\n\n\nPerformance, Internal Implementation, etc.\n\n * When parsing a 40-hex string into the object name, the string is\n   checked to see if it can be interpreted as a ref so that a warning\n   can be given for ambiguity. The code kicked in even when the\n   core.warnambiguousrefs is set to false to squelch this warning, in\n   which case the cycles spent to look at the ref namespace were an\n   expensive no-op, as the result was discarded without being used.\n\n * The naming convention of the packfiles has been updated; it used to\n   be based on the enumeration of names of the objects that are\n   contained in the pack, but now it also depends on how the packed\n   result is represented---packing the same set of objects using\n   different settings (or delta order) would produce a pack with\n   different name.\n\n * \"git diff --no-index\" mode used to unnecessarily attempt to read\n   the index when there is one.\n\n * The deprecated parse-options macro OPT_BOOLEAN has been removed;\n   use OPT_BOOL or OPT_COUNTUP in new code.\n\n * A few duplicate implementations of prefix/suffix string comparison\n   functions have been unified to starts_with() and ends_with().\n\n * The new PERLLIB_EXTRA makefile variable can be used to specify\n   additional directories Perl modules (e.g. the ones necessary to run\n   git-svn) are installed on the platform when building.\n\n * \"git merge-base\" learned the \"--fork-point\" mode, that implements\n   the same logic used in \"git pull --rebase\" to find a suitable fork\n   point out of the reflog entries for the remote-tracking branch the\n   work has been based on.  \"git rebase\" has the same logic that can be\n   triggered with the \"--fork-point\" option.\n\n * A third-party \"receive-pack\" (the responder to \"git push\") can\n   advertise the \"no-thin\" capability to tell \"git push\" not to use\n   the thin-pack optimization. Our receive-pack has always been\n   capable of accepting and fattening a thin-pack, and will continue\n   not to ask \"git push\" to use a non-thin pack.\n\n\nAlso contains various documentation updates and code clean-ups.\n\n\nFixes since v1.8.5\n------------------\n\nUnless otherwise noted, all the fixes since v1.8.5 in the maintenance\ntrack are contained in this release (see the maintenance releases' notes\nfor details).\n\n * The pathspec matching code, while comparing two trees (e.g. \"git\n   diff A B -- path1 path2\") was too agrresive and failed to match\n   some paths when multiple pathspecs were involved.\n   (merge e4ddb05 as/tree-walk-fix-aggressive-short-cut later to maint).\n\n * \"git repack --max-pack-size=8g\" stopped being parsed correctly when\n   the command was reimplemented in C.\n   (merge b861e23 sb/repack-in-c later to maint).\n\n * An earlier update in v1.8.4.x to \"git rev-list --objects\" with\n   negative ref had performance regression.\n   (merge 200abe7 jk/mark-edges-uninteresting later to maint).\n\n * A recent update to \"git send-email\" broke platforms where\n   /etc/ssl/certs/ directory exists, but it cannot used as SSL_ca_path\n   (e.g. Fedora rawhide).\n   (merge 01645b7 rk/send-email-ssl-cert later to maint).\n\n * A handful of bugs around interpreting $branch@{upstream} notation\n   and its lookalike, when $branch part has interesting characters,\n   e.g. \"@\", and \":\", have been fixed.\n   (merge 9892d5d jk/interpret-branch-name-fix later to maint).\n\n * \"git clone\" would fail to clone from a repository that has a ref\n   directly under \"refs/\", e.g. \"refs/stash\", because different\n   validation paths do different things on such a refname.  Loosen the\n   client side's validation to allow such a ref.\n   (merge 4c22408 jk/allow-fetch-onelevel-refname later to maint).\n\n * \"git log --left-right A...B\" lost the \"leftness\" of commits\n   reachable from A when A is a tag as a side effect of a recent\n   bugfix.  This is a regression in 1.8.4.x series.\n   (merge a743528 jc/revision-range-unpeel later to maint).\n\n * documentations to \"git pull\" hinted there is an \"-m\" option because\n   it incorrectly shared the documentation with \"git merge\".\n   (merge 08f19cf jc/maint-pull-docfix later to maint).\n\n * \"git diff A B submod\" and \"git diff A B submod/\" ought to have done\n   the same for a submodule \"submod\", but didn't.\n\n * \"git clone $origin foo\\bar\\baz\" on Windows failed to create the\n   leading directories (i.e. a moral-equivalent of \"mkdir -p\").\n\n * \"submodule.*.update=checkout\", when propagated from .gitmodules to\n   .git/config, turned into a \"submodule.*.update=none\", which did not\n   make much sense.\n   (merge efa8fd7 fp/submodule-checkout-mode later to maint).\n\n * The implementation of 'git stash $cmd \"stash@{...}\"' did not quote\n   the stash argument properly and left it split at IFS whitespace.\n   (merge 2a07e43 ow/stash-with-ifs later to maint).\n\n * The \"--[no-]informative-errors\" options to \"git daemon\" were parsed\n   a bit too loosely, allowing any other string after these option\n   names.\n   (merge 82246b7 nd/daemon-informative-errors-typofix later to maint).\n\n * There is no reason to have a hardcoded upper limit of the number of\n   parents for an octopus merge, created via the graft mechanism, but\n   there was.\n   (merge e228c17 js/lift-parent-count-limit later to maint).\n\n * The basic test used to leave unnecessary trash directories in the\n   t/ directory.\n   (merge 738a8be jk/test-framework-updates later to maint).\n\n * \"git merge-base --octopus\" used to leave cleaning up suboptimal\n   result to the caller, but now it does the clean-up itself.\n   (merge 8f29299 bm/merge-base-octopus-dedup later to maint).\n\n * A \"gc\" process running as a different user should be able to stop a\n   new \"gc\" process from starting, but it didn't.\n   (merge ed7eda8 km/gc-eperm later to maint).\n\n * An earlier \"clean-up\" introduced an unnecessary memory leak.\n   (merge e1c1a32 jk/credential-plug-leak later to maint).\n\n * \"git add -A\" (no other arguments) in a totally empty working tree\n   used to emit an error.\n   (merge 64ed07c nd/add-empty-fix later to maint).\n\n * \"git log --decorate\" did not handle a tag pointed by another tag\n   nicely.\n   (merge 5e1361c bc/log-decoration later to maint).\n\n * When we figure out how many file descriptors to allocate for\n   keeping packfiles open, a system with non-working getrlimit() could\n   cause us to die(), but because we make this call only to get a\n   rough estimate of how many is available and we do not even attempt\n   to use up all file descriptors available ourselves, it is nicer to\n   fall back to a reasonable low value rather than dying.\n   (merge 491a8de jh/rlimit-nofile-fallback later to maint).\n\n * read_sha1_file(), that is the workhorse to read the contents given\n   an object name, honoured object replacements, but there was no\n   corresponding mechanism to sha1_object_info() that was used to\n   obtain the metainfo (e.g. type & size) about the object.  This led\n   callers to weird inconsistencies.\n   (merge 663a856 cc/replace-object-info later to maint).\n\n * \"git cat-file --batch=\", an admittedly useless command, did not\n   behave very well.\n   (merge 6554dfa jk/cat-file-regression-fix later to maint).\n\n * \"git rev-parse <revs> -- <paths>\" did not implement the usual\n   disambiguation rules the commands in the \"git log\" family used in\n   the same way.\n   (merge 62f162f jk/rev-parse-double-dashes later to maint).\n\n * \"git mv A B/\", when B does not exist as a directory, should error\n   out, but it didn't.\n   (merge c57f628 mm/mv-file-to-no-such-dir-with-slash later to maint).\n\n * A workaround to an old bug in glibc prior to glibc 2.17 has been\n   retired; this would remove a side effect of the workaround that\n   corrupts system error messages in non-C locales.\n\n * SSL-related options were not passed correctly to underlying socket\n   layer in \"git send-email\".\n   (merge 5508f3e tr/send-email-ssl later to maint).\n\n * \"git commit -v\" appends the patch to the log message before\n   editing, and then removes the patch when the editor returned\n   control. However, the patch was not stripped correctly when the\n   first modified path was a submodule.\n   (merge 1a72cfd jl/commit-v-strip-marker later to maint).\n\n * \"git fetch --depth=0\" was a no-op, and was silently ignored.\n   Diagnose it as an error.\n   (merge 5594bca nd/transport-positive-depth-only later to maint).\n\n * Remote repository URL expressed in scp-style host:path notation are\n   parsed more carefully (e.g. \"foo/bar:baz\" is local, \"[::1]:/~user\" asks\n   to connect to user's home directory on host at address ::1.\n   (merge a2036d7 tb/clone-ssh-with-colon-for-port later to maint).\n\n * \"git diff -- ':(icase)makefile'\" was unnecessarily rejected at the\n   command line parser.\n   (merge 887c6c1 nd/magic-pathspec later to maint).\n\n * \"git cat-file --batch-check=ok\" did not check the existence of\n   the named object.\n   (merge 4ef8d1d sb/sha1-loose-object-info-check-existence later to maint).\n\n * \"git am --abort\" sometimes complained about not being able to write\n   a tree with an 0{40} object in it.\n   (merge 77b43ca jk/two-way-merge-corner-case-fix later to maint).\n\n * Two processes creating loose objects at the same time could have\n   failed unnecessarily when the name of their new objects started\n   with the same byte value, due to a race condition.\n   (merge b2476a6 jh/loose-object-dirs-creation-race later to maint).\n\n----------------------------------------------------------------\n\nChanges since v1.9-rc2 are as follows:\n"}]}