{"thread":{"id":"52829","subject":"What's cooking in git.git (Feb 2020, #04; Mon, 17)","startedAt":"2020-02-17T22:08:47Z","lastAt":"2020-02-26T20:04:21Z","messageCount":14,"participants":["Junio C Hamano","Elijah Newren","Johannes Schindelin","Taylor Blau","Eric Sunshine","Phillip Wood","Alban Gruin"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"391957","messageId":"xmqq36b8lvdl.fsf@gitster-ct.c.googlers.com","threadId":"52829","inReplyTo":null,"subject":"What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2020-02-17T22:08:38Z","receivedAt":"2020-02-17T22:08:47Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Here are the topics that have been cooking.  Commits prefixed with\n'-' are only in 'pu' (proposed updates) while commits prefixed with\n'+' are in 'next'.  The ones marked with '.' do not appear in any of\nthe integration branches, but I am still holding onto them.\n\nAccumulated fixes on the 'master' front have been flushed to 'maint'\nand a new maintenance release 2.25.1 has been tagged.\n\nYou can find the changes described here in the integration branches\nof the repositories listed at\n\n    http://git-blame.blogspot.com/p/git-public-repositories.html\n\n--------------------------------------------------\n[Graduated to \"master\"]\n\n* ag/edit-todo-drop-check (2020-01-28) 2 commits\n  (merged to 'next' on 2020-02-07 at 687c82cd3a)\n + rebase-interactive: warn if commit is dropped with `rebase --edit-todo'\n + sequencer: move check_todo_list_from_file() to rebase-interactive.c\n (this branch uses js/rebase-i-with-colliding-hash.)\n\n Allow the rebase.missingCommitsCheck configuration to kick in when\n \"rebase --edit-todo\" and \"rebase --continue\" restarts the procedure.\n\n\n* ag/rebase-avoid-unneeded-checkout (2020-01-24) 1 commit\n  (merged to 'next' on 2020-02-07 at b08da07534)\n + rebase -i: stop checking out the tip of the branch to rebase\n\n \"git rebase -i\" (and friends) used to unnecessarily check out the\n tip of the branch to be rebased, which has been corrected.\n\n\n* bc/hash-independent-tests-part-8 (2020-02-07) 21 commits\n  (merged to 'next' on 2020-02-11 at 127748e61a)\n + t6024: update for SHA-256\n + t6006: make hash size independent\n + t6000: abstract away SHA-1-specific constants\n + t5703: make test work with SHA-256\n + t5607: make hash size independent\n + t5318: update for SHA-256\n + t5515: make test hash independent\n + t5321: make test hash independent\n + t5313: make test hash independent\n + t5309: make test hash independent\n + t5302: make hash size independent\n + t4060: make test work with SHA-256\n + t4211: add test cases for SHA-256\n + t4211: move SHA-1-specific test cases into a directory\n + t4013: make test hash independent\n + t3311: make test work with SHA-256\n + t3310: make test work with SHA-256\n + t3309: make test work with SHA-256\n + t3308: make test work with SHA-256\n + t3206: make hash size independent\n + t/lib-pack: support SHA-256\n\n Preparation for SHA-256 migration continues.\n\n\n* dl/test-must-fail-fixes-2 (2020-01-27) 16 commits\n  (merged to 'next' on 2020-02-07 at 535e019692)\n + t4124: only mark git command with test_must_fail\n + t3507: use test_path_is_missing()\n + t3507: fix indentation\n + t3504: do check for conflict marker after failed cherry-pick\n + t3419: stop losing return code of git command\n + t3415: increase granularity of test_auto_{fixup,squash}()\n + t3415: stop losing return codes of git commands\n + t3310: extract common notes_merge_files_gone()\n + t3030: use test_path_is_missing()\n + t2018: replace \"sha\" with \"oid\"\n + t2018: don't lose return code of git commands\n + t2018: teach do_checkout() to accept `!` arg\n + t2018: be more discerning when checking for expected exit codes\n + t2018: improve style of if-statement\n + t2018: add space between function name and ()\n + t2018: remove trailing space from test description\n\n Test updates.\n\n\n* ds/sparse-checkout-harden (2020-01-31) 15 commits\n  (merged to 'next' on 2020-02-07 at 56c09b944f)\n + sparse-checkout: fix cone mode behavior mismatch\n + sparse-checkout: improve docs around 'set' in cone mode\n + sparse-checkout: escape all glob characters on write\n + sparse-checkout: use C-style quotes in 'list' subcommand\n + sparse-checkout: unquote C-style strings over --stdin\n + sparse-checkout: write escaped patterns in cone mode\n + sparse-checkout: properly match escaped characters\n + sparse-checkout: warn on globs in cone patterns\n + sparse-checkout: detect short patterns\n + sparse-checkout: cone mode does not recognize \"**\"\n + sparse-checkout: fix documentation typo for core.sparseCheckoutCone\n + clone: fix --sparse option with URLs\n + sparse-checkout: create leading directories\n + t1091: improve here-docs\n + t1091: use check_files to reduce boilerplate\n (this branch is used by ds/sparse-add.)\n\n Some rough edges in the sparse-checkout feature, especially around\n the cone mode, have been cleaned up.\n\n\n* es/outside-repo-errmsg-hints (2020-02-16) 1 commit\n  (merged to 'next' on 2020-02-16 at ed78b064b4)\n + prefix_path: show gitdir when arg is outside repo\n\n Error message clarification.\n\n\n* hw/advice-add-nothing (2020-02-06) 2 commits\n  (merged to 'next' on 2020-02-07 at 3c2a5c1868)\n + add: change advice config variables used by the add API\n  (merged to 'next' on 2020-01-22 at 3fe4494fa6)\n + add: use advise function to display hints\n\n Two help messages given when \"git add\" notices the user gave it\n nothing to add have been updated to use advise() API.\n\n\n* jh/notes-fanout-fix (2020-02-04) 2 commits\n  (merged to 'next' on 2020-02-07 at a4ec94aee0)\n + notes.c: fix off-by-one error when decreasing notes fanout\n + t3305: check notes fanout more carefully and robustly\n\n The code to automatically shrink the fan-out in the notes tree had\n an off-by-one bug, which has been killed.\n\n\n* jk/alloc-cleanups (2020-02-04) 4 commits\n  (merged to 'next' on 2020-02-07 at f259621312)\n + tree-walk.c: break circular dependency with unpack-trees\n + traverse_trees(): use stack array for name entries\n + walker_fetch(): avoid raw array length computation\n + normalize_path_copy(): document \"dst\" size expectations\n\n Misc code clean-ups.\n\n\n* jk/diff-honor-wserrhighlight-in-plumbing (2020-01-31) 1 commit\n  (merged to 'next' on 2020-02-07 at acb077fdcb)\n + diff: move diff.wsErrorHighlight to \"basic\" config\n\n The diff-* plumbing family of subcommands now pay attention to the\n diff.wsErrorHighlight configuration, which has been ignored before;\n this allows \"git add -p\" to also show the whitespace problems to\n the end user.\n\n\n* jk/doc-credential-helper (2020-02-14) 1 commit\n  (merged to 'next' on 2020-02-16 at 0909d3f981)\n + doc: move credential helper info into gitcredentials(7)\n\n Docfix.\n\n\n* jk/get-oid-error-message-i18n (2020-01-30) 3 commits\n  (merged to 'next' on 2020-02-07 at 6bfe9235e5)\n + sha1-name: mark get_oid() error messages for translation\n + t1506: drop space after redirection operator\n + t1400: avoid \"test\" string comparisons\n\n A low-level API function get_oid(), that accepts various ways to\n name an object, used to issue end-user facing error messages\n without l10n, which has been updated to be translatable.\n\n\n* jk/index-pack-dupfix (2020-02-04) 1 commit\n  (merged to 'next' on 2020-02-07 at e6d79d432b)\n + index-pack: downgrade twice-resolved REF_DELTA to die()\n\n The index-pack code now diagnoses a bad input packstream that\n records the same object twice when it is used as delta base; the\n code used to declare a software bug when encountering such an\n input, but it is an input error.\n\n\n* jk/mailinfo-cleanup (2020-02-11) 4 commits\n  (merged to 'next' on 2020-02-11 at ab9890fbcb)\n + mailinfo: factor out some repeated header handling\n + mailinfo: be more liberal with header whitespace\n + mailinfo: simplify parsing of header values\n + mailinfo: treat header values as C strings\n (this branch uses rs/strbuf-insertstr.)\n\n Code clean-up.\n\n\n* jk/packfile-reuse-cleanup (2020-01-23) 12 commits\n  (merged to 'next' on 2020-02-07 at c543bc27fb)\n + pack-bitmap: don't rely on bitmap_git->reuse_objects\n + pack-objects: add checks for duplicate objects\n + pack-objects: improve partial packfile reuse\n + builtin/pack-objects: introduce obj_is_packed()\n + pack-objects: introduce pack.allowPackReuse\n + csum-file: introduce hashfile_total()\n + pack-bitmap: simplify bitmap_has_oid_in_uninteresting()\n + pack-bitmap: uninteresting oid can be outside bitmapped packfile\n + pack-bitmap: introduce bitmap_walk_contains()\n + ewah/bitmap: introduce bitmap_word_alloc()\n + packfile: expose get_delta_base()\n + builtin/pack-objects: report reused packfile objects\n\n The way \"git pack-objects\" reuses objects stored in existing pack\n to generate its result has been improved.\n\n\n* jn/promote-proto2-to-default (2020-01-15) 5 commits\n  (merged to 'next' on 2020-02-07 at 5597fd726b)\n + fetch: default to protocol version 2\n + protocol test: let protocol.version override GIT_TEST_PROTOCOL_VERSION\n + test: request GIT_TEST_PROTOCOL_VERSION=0 when appropriate\n + config doc: protocol.version is not experimental\n + fetch test: use more robust test for filtered objects\n\n The transport protocol version 2 becomes the default one.\n\n\n* js/mingw-open-in-gdb (2020-02-14) 1 commit\n  (merged to 'next' on 2020-02-16 at 6670f36d73)\n + mingw: add a helper function to attach GDB to the current process\n\n Dev support.\n\n\n* js/rebase-i-with-colliding-hash (2020-01-23) 3 commits\n  (merged to 'next' on 2020-02-07 at 670ba07c4d)\n + rebase -i: also avoid SHA-1 collisions with missingCommitsCheck\n + rebase -i: re-fix short SHA-1 collision\n + parse_insn_line(): improve error message when parsing failed\n (this branch is used by ag/edit-todo-drop-check.)\n\n \"git rebase -i\" identifies existing commits in its todo file with\n their abbreviated object name, which could become ambigous as it\n goes to create new commits, and has a mechanism to avoid ambiguity\n in the main part of its execution.  A few other cases however were\n not covered by the protection against ambiguity, which has been\n corrected.\n\n\n* js/test-avoid-pipe (2020-02-14) 1 commit\n  (merged to 'next' on 2020-02-16 at 090cac9c23)\n + t9001, t9116: avoid pipes\n\n Test clean-up.\n\n\n* js/test-unc-fetch (2020-02-14) 1 commit\n  (merged to 'next' on 2020-02-16 at a4ca01efef)\n + t5580: test cloning without file://, test fetching via UNC paths\n\n Test updates.\n\n\n* js/test-write-junit-xml-fix (2020-02-12) 1 commit\n  (merged to 'next' on 2020-02-16 at 2deb247550)\n + tests: fix --write-junit-xml with subshells\n\n Testfix.\n\n\n* jt/connectivity-check-optim-in-partial-clone (2020-01-30) 2 commits\n  (merged to 'next' on 2020-02-07 at ca5102fa58)\n + fetch: forgo full connectivity check if --filter\n + connected: verify promisor-ness of partial clone\n\n Unneeded connectivity check is now disabled in a partial clone when\n fetching into it.\n\n\n* kw/fsmonitor-watchman-racefix (2020-01-23) 4 commits\n  (merged to 'next' on 2020-02-07 at 3c42195a34)\n + fsmonitor: update documentation for hook version and watchman hooks\n + fsmonitor: add fsmonitor hook scripts for version 2\n + fsmonitor: handle version 2 of the hooks that will use opaque token\n + fsmonitor: change last update timestamp on the index_state to opaque token\n\n A new version of fsmonitor-watchman hook has been introduced, to\n avoid races.\n\n\n* ld/p4-cleanup-processes (2020-01-30) 7 commits\n  (merged to 'next' on 2020-02-07 at e8c573f01c)\n + git-p4: avoid leak of file handle when cloning\n + git-p4: check for access to remote host earlier\n + git-p4: cleanup better on error exit\n + git-p4: create helper function importRevisions()\n + git-p4: disable some pylint warnings, to get pylint output to something manageable\n + git-p4: add P4CommandException to report errors talking to Perforce\n + git-p4: make closeStreams() idempotent\n\n p4 updates.\n\n\n* mr/show-config-scope (2020-02-10) 10 commits\n  (merged to 'next' on 2020-02-11 at 904bca086e)\n + config: add '--show-scope' to print the scope of a config value\n + submodule-config: add subomdule config scope\n + config: teach git_config_source to remember its scope\n + config: preserve scope in do_git_config_sequence\n + config: clarify meaning of command line scoping\n + config: split repo scope to local and worktree\n + config: make scope_name non-static and rename it\n + t1300: create custom config file without special characters\n + t1300: fix over-indented HERE-DOCs\n + config: fix typo in variable name\n (this branch is used by bw/remote-rename-update-config and es/bugreport.)\n\n \"git config\" learned to show in which \"scope\", in addition to in\n which file, each config setting comes from.\n\n\n* mt/threaded-grep-in-object-store (2020-01-17) 12 commits\n  (merged to 'next' on 2020-02-07 at e86bab09e1)\n + grep: use no. of cores as the default no. of threads\n + grep: move driver pre-load out of critical section\n + grep: re-enable threads in non-worktree case\n + grep: protect packed_git [re-]initialization\n + grep: allow submodule functions to run in parallel\n + submodule-config: add skip_if_read option to repo_read_gitmodules()\n + grep: replace grep_read_mutex by internal obj read lock\n + object-store: allow threaded access to object reading\n + replace-object: make replace operations thread-safe\n + grep: fix racy calls in grep_objects()\n + grep: fix race conditions at grep_submodule()\n + grep: fix race conditions on userdiff calls\n\n Traditionally, we avoided threaded grep while searching in objects\n (as opposed to files in the working tree) as accesses to the object\n layer is not thread-safe.  This limitation is getting lifted.\n\n\n* mt/use-passed-repo-more-in-funcs (2020-01-31) 7 commits\n  (merged to 'next' on 2020-02-07 at 536709ab4b)\n + sha1-file: allow check_object_signature() to handle any repo\n + sha1-file: pass git_hash_algo to hash_object_file()\n + sha1-file: pass git_hash_algo to write_object_file_prepare()\n + streaming: allow open_istream() to handle any repo\n + pack-check: use given repo's hash_algo at verify_packfile()\n + cache-tree: use given repo's hash_algo at verify_one()\n + diff: make diff_populate_filespec() honor its repo argument\n\n Some codepaths were given a repository instance as a parameter to\n work in the repository, but passed the_repository instance to its\n callees, which has been cleaned up (somewhat).\n\n\n* pk/status-of-uncloned-submodule (2020-02-04) 3 commits\n  (merged to 'next' on 2020-02-07 at f6349f5f5f)\n + t7400: testcase for submodule status on unregistered inner git repos\n + submodule: fix status of initialized but not cloned submodules\n + t7400: add a testcase for submodule status on empty dirs\n\n The way \"git submodule status\" reports an initialized but not yet\n populated submodule has not been reimplemented correctly when a\n part of the \"git submodule\" command was rewritten in C, which has\n been corrected.\n\n\n* rs/name-rev-memsave (2020-02-05) 11 commits\n  (merged to 'next' on 2020-02-11 at 17616cb9b1)\n + name-rev: sort tip names before applying\n + name-rev: release unused name strings\n + name-rev: generate name strings only if they are better\n + name-rev: pre-size buffer in get_parent_name()\n + name-rev: factor out get_parent_name()\n + name-rev: put struct rev_name into commit slab\n + name-rev: don't _peek() in create_or_update_name()\n + name-rev: don't leak path copy in name_ref()\n + name-rev: respect const qualifier\n + name-rev: remove unused typedef\n + name-rev: rewrite create_or_update_name()\n\n Memory footprint and performance of \"git name-rev\" has been\n improved.\n\n\n* rs/parse-options-concat-dup (2020-02-10) 4 commits\n  (merged to 'next' on 2020-02-11 at a1685687ab)\n + parse-options: simplify parse_options_dup()\n + parse-options: const parse_options_concat() parameters\n + parse-options: factor out parse_options_count()\n + parse-options: use COPY_ARRAY in parse_options_concat()\n\n Code clean-up.\n\n\n* rs/strbuf-insertstr (2020-02-10) 2 commits\n  (merged to 'next' on 2020-02-11 at 5c89aa002f)\n + mailinfo: don't insert header prefix for handle_content_type()\n + strbuf: add and use strbuf_insertstr()\n (this branch is used by jk/mailinfo-cleanup.)\n\n Code clean-up.\n\n\n* tb/commit-graph-object-dir (2020-02-04) 5 commits\n  (merged to 'next' on 2020-02-07 at 5c85173df4)\n + commit-graph.h: use odb in 'load_commit_graph_one_fd_st'\n + commit-graph.c: remove path normalization, comparison\n + commit-graph.h: store object directory in 'struct commit_graph'\n + commit-graph.h: store an odb in 'struct write_commit_graph_context'\n + t5318: don't pass non-object directory to '--object-dir'\n (this branch is used by tb/commit-graph-split-merge.)\n\n The code to compute the commit-graph has been taught to use a more\n robust way to tell if two object directories refer to the same\n thing.\n\n--------------------------------------------------\n[New Topics]\n\n* jk/object-filter-with-bitmap (2020-02-14) 15 commits\n  (merged to 'next' on 2020-02-16 at 42425c3658)\n + pack-objects: support filters with bitmaps\n + pack-bitmap: implement BLOB_LIMIT filtering\n + pack-bitmap: implement BLOB_NONE filtering\n + bitmap: add bitmap_unset() function\n + rev-list: use bitmap filters for traversal\n + pack-bitmap: basic noop bitmap filter infrastructure\n + rev-list: allow commit-only bitmap traversals\n + t5310: factor out bitmap traversal comparison\n + rev-list: allow bitmaps when counting objects\n + rev-list: make --count work with --objects\n + rev-list: factor out bitmap-optimized routines\n + pack-bitmap: refuse to do a bitmap traversal with pathspecs\n + rev-list: fallback to non-bitmap traversal when filtering\n + pack-bitmap: fix leak of haves/wants object lists\n + pack-bitmap: factor out type iterator initialization\n\n The object reachability bitmap machinery and the partial cloning\n machinery were not prepared to work well together, because some\n object-filtering criteria partial clones use inherently rely on\n object traversal, but the bitmap machinery is an optimization to\n bypass that object traversal.  There however are some cases that\n they can work together, and they were taught about this case.\n\n Will merge to 'master'.\n\n\n* hw/advise-ng (2020-02-16) 2 commits\n - advice: extract vadvise() from advise()\n - advice: refactor advise API\n\n Revamping of the advise API to allow more systematic enumeration of\n advice knobs in the future.\n\n--------------------------------------------------\n[Stalled]\n\n* pw/advise-rebase-skip (2019-12-06) 9 commits\n - rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\n - rebase: fix advice when a fixup creates an empty commit\n - commit: give correct advice for empty commit during a rebase\n - commit: encapsulate determine_whence() for sequencer\n - commit: use enum value for multiple cherry-picks\n - sequencer: write CHERRY_PICK_HEAD for reword and edit\n - cherry-pick: check commit error messages\n - cherry-pick: add test for `--skip` advice in `git commit`\n - t3404: use test_cmp_rev\n\n The mechanism to prevent \"git commit\" from making an empty commit\n or amending during an interrupted cherry-pick was broken during the\n rewrite of \"git rebase\" in C, which has been corrected.\n\n What's the status of this one?\n The tip two are still RFC.\n\n\n* vn/reset-deleted-ita (2019-07-26) 1 commit\n - reset: unstage empty deleted ita files\n\n \"git reset HEAD [<pathspec>]\" did not reset an empty file that was\n added with the intent-to-add bit.\n\n Expecting a reroll.\n\n\n* mk/use-size-t-in-zlib (2018-10-15) 1 commit\n - zlib.c: use size_t for size\n\n The wrapper to call into zlib followed our long tradition to use\n \"unsigned long\" for sizes of regions in memory, which have been\n updated to use \"size_t\".\n\n--------------------------------------------------\n[Cooking]\n\n* ds/sparse-add (2020-02-11) 4 commits\n  (merged to 'next' on 2020-02-14 at 3799757bbe)\n + sparse-checkout: work with Windows paths\n + sparse-checkout: create 'add' subcommand\n + sparse-checkout: extract pattern update from 'set' subcommand\n + sparse-checkout: extract add_patterns_from_input()\n\n \"git sparse-checkout\" learned a new \"add\" subcommand.\n\n Will merge to 'master'.\n\n\n* at/rebase-fork-point-regression-fix (2020-02-11) 1 commit\n - rebase: --fork-point regression fix\n\n The \"--fork-point\" mode of \"git rebase\" regressed when the command\n was rewritten in C back in 2.20 era, which has been corrected.\n\n Was waiting for discussion to settle.\n cf. <CAPig+cQ-3Ds41hr91fRo_GvuFMTP7zNVJtaSqi-Yccq4Pk-8Qg@mail.gmail.com>\n cf. <xmqq36dtwcvw.fsf@gitster-ct.c.googlers.com>\n\n\n* hn/reftable (2020-02-10) 5 commits\n . Reftable support for git-core\n . Add reftable library\n . refs: document how ref_iterator_advance_fn should handle symrefs\n . create .git/refs in files-backend.c\n . refs.h: clarify reflog iteration order\n\n A new refs backend \"reftable\" to replace the traditional\n combination of packed-refs files and one-file-per-ref loose refs\n has been implemented and integrated for improved performance and\n atomicity.\n\n At v5.\n\n\n* es/bugreport (2020-02-14) 15 commits\n - bugreport: summarize contents of alternates file\n - bugreport: list contents of $OBJDIR/info\n - bugreport: add packed object summary\n - bugreport: count loose objects\n - bugreport: collect list of populated hooks\n - bugreport: add config values from safelist\n - bugreport: generate config safelist based on docs\n - bugreport: include user interactive shell\n - bugreport: add git-remote-https version\n - bugreport: add compiler info\n - bugreport: add uname info\n - bugreport: gather git version and build info\n - bugreport: add tool to generate debugging info\n - help: add shell-path to --build-options\n - help: move list_config_help to builtin/help\n\n The \"bugreport\" tool.\n\n at v7.\n\n\n* gs/commit-graph-path-filter (2020-02-12) 12 commits\n - (bytesex breakage band-aid)\n - commit-graph: add GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS test flag\n - revision.c: use Bloom filters to speed up path based revision walks\n - commit-graph: add --changed-paths option to write subcommand\n - commit-graph: reuse existing Bloom filters during write.\n - commit-graph: write Bloom filters to commit graph file\n - commit-graph: examine commits by generation number\n - commit-graph: examine changed-path objects in pack order\n - commit-graph: compute Bloom filters for changed paths\n - diff: halt tree-diff early after max_changes\n - bloom: core Bloom filter implementation for changed paths\n - commit-graph: use MAX_NUM_CHUNKS\n\n Introduce an extension to the commit-graph to make it efficient to\n check for the paths that were modified at each commit using Bloom\n filters.\n\n Will be rerolled with bytesex fixes squashed in.\n Breakage due to byte-order dependency reported.\n\n\n* mr/bisect-in-c-1 (2020-02-10) 13 commits\n - bisect: libify `bisect_next_all`\n - bisect: libify `handle_bad_merge_base` and its dependents\n - bisect: libify `check_good_are_ancestors_of_bad` and its dependents\n - bisect: libify `check_merge_bases` and its dependents\n - bisect: libify `bisect_checkout`\n - bisect: libify `exit_if_skipped_commits` to `error_if_skipped*` and its dependents\n - bisect--helper: return error codes from `cmd_bisect__helper()`\n - bisect: add enum to represent bisect returning codes\n - bisect--helper: introduce new `decide_next()` function\n - run-command: make `exists_in_PATH()` non-static\n - bisect: use the standard 'if (!var)' way to check for 0\n - bisect--helper: change `retval` to `res`\n - bisect--helper: convert `vocab_*` char pointers to char arrays\n\n Underlying machinery of \"git bisect--helper\" is being refactored\n into pieces that are more easily reused.\n\n\n* bw/remote-rename-update-config (2020-02-10) 6 commits\n  (merged to 'next' on 2020-02-14 at 30b1a477de)\n + remote rename/remove: gently handle remote.pushDefault config\n + config: provide access to the current line number\n + remote rename/remove: handle branch.<name>.pushRemote config values\n + remote: clean-up config callback\n + remote: clean-up by returning early to avoid one indentation\n + pull --rebase/remote rename: document and honor single-letter abbreviations rebase types\n\n \"git remote rename X Y\" needs to adjust configuration variables\n (e.g. branch.<name>.remote) whose value used to be X to Y.\n branch.<name>.pushRemote is now also updated.\n\n Will merge to 'master'.\n\n\n* es/bright-colors (2020-02-11) 3 commits\n  (merged to 'next' on 2020-02-14 at cc17dbc7e8)\n + color.c: alias RGB colors 8-15 to aixterm colors\n + color.c: support bright aixterm colors\n + color.c: refactor color_output arguments\n\n The basic 7 colors learned the brighter counterparts\n (e.g. \"brightred\").\n\n Will merge to 'master'.\n\n\n* es/doc-mentoring (2020-02-14) 2 commits\n  (merged to 'next' on 2020-02-17 at e3f3e316eb)\n + MyFirstContribution: rephrase contact info\n  (merged to 'next' on 2020-02-05 at 3b05aabbcd)\n + MyFirstContribution: add avenues for getting help\n\n Doc for new contributors.\n\n Will merge to 'master'.\n cf. <xmqqr1yxnkab.fsf@gitster-ct.c.googlers.com>\n\n\n* bk/p4-pre-edit-changelist (2020-02-14) 7 commits\n - git-p4: add RCS keyword status message\n - git-p4: add p4 submit hooks\n - git-p4: restructure code in submit\n - git-p4: add --no-verify option\n - git-p4: add p4-pre-submit exit text\n - git-p4: create new function run_git_hook\n - git-p4: rewrite prompt to be Windows compatible\n\n \"git p4\" learned four new hooks and also \"--no-verify\" option to\n bypass them (and the existing \"p4-pre-submit\" hook).\n\n Waiting for response to RFH from Windows/Python folks.\n cf. <pull.698.v6.git.git.1581691486.gitgitgadget@gmail.com>\n\n\n* en/fill-directory-exponential (2020-01-31) 6 commits\n - t7063: blindly accept diffs\n - dir: replace exponential algorithm with a linear one\n - dir: refactor treat_directory to clarify control flow\n - dir: fix confusion based on variable tense\n - dir: fix broken comment\n - dir: consolidate treat_path() and treat_one_path()\n\n The directory traversal code had redundant recursive calls which\n made its performance characteristics exponential wrt the depth of\n the tree, which was corrected.\n\n Still RFC?\n cf. <pull.700.v2.git.git.1580495486.gitgitgadget@gmail.com>\n cf. <CABPp-BF0SO92wxY-dM14Vxv6BGA5=T=CicdwbiaZ7c9w6-j2jQ@mail.gmail.com>\n\n\n* tb/commit-graph-split-merge (2020-02-12) 3 commits\n - builtin/commit-graph.c: support '--input=none'\n - builtin/commit-graph.c: introduce '--input=<source>'\n - builtin/commit-graph.c: support '--split[=<strategy>]'\n\n The code to write out the commit-graph has been taught a few\n options to control if the resulting graph chains should be merged\n or a single new incremental graph is created.\n\n Waiting for the discussion to settle.\n cf. <20200212205028.GE4364@syl.local>\n\n\n* am/pathspec-f-f-more (2020-02-10) 8 commits\n - stash push: support the --pathspec-from-file option\n - stash: eliminate crude option parsing\n - doc: stash: synchronize <pathspec> description\n - doc: stash: document more options\n - doc: stash: split options from description (2)\n - doc: stash: split options from description (1)\n - rm: support the --pathspec-from-file option\n - doc: rm: synchronize <pathspec> description\n\n \"git rm\" and \"git stash\" learns the new \"--pathspec-from-file\"\n option.\n\n\n* dl/merge-autostash (2020-01-13) 17 commits\n - pull: pass --autostash to merge\n - t5520: make test_pull_autostash() accept expect_parent_num\n - merge: teach --autostash option\n - sequencer: unlink autostash in apply_autostash()\n - sequencer: extract perform_autostash() from rebase\n - rebase: generify create_autostash()\n - rebase: extract create_autostash()\n - reset: extract reset_head() from rebase\n - rebase: generify reset_head()\n - rebase: use apply_autostash() from sequencer.c\n - sequencer: make apply_rebase() accept a path\n - rebase: use read_oneliner()\n - sequencer: make read_oneliner() extern\n - sequencer: configurably warn on non-existent files\n - sequencer: use file strbuf for read_oneliner()\n - t7600: use test_write_lines()\n - Makefile: alphabetically sort += lists\n\n \"git merge\" learns the \"--autostash\" option.\n\n Expecting a reroll.\n cf. <20200123042906.GA29009@generichostname>\n\n\n* yz/p4-py3 (2020-02-13) 14 commits\n - ci: use python3 in linux-gcc and osx-gcc and python2 elsewhere\n - git-p4: use python3's input() everywhere\n - git-p4: simplify regex pattern generation for parsing diff-tree\n - git-p4: use dict.items() iteration for python3 compatibility\n - git-p4: use functools.reduce instead of reduce\n - git-p4: fix freezing while waiting for fast-import progress\n - git-p4: use marshal format version 2 when sending to p4\n - git-p4: open .gitp4-usercache.txt in text mode\n - git-p4: convert path to unicode before processing them\n - git-p4: encode/decode communication with git for python3\n - git-p4: encode/decode communication with p4 for python3\n - git-p4: remove string type aliasing\n - git-p4: change the expansion test from basestring to list\n - git-p4: make python2.7 the oldest supported version\n\n Update \"git p4\" to work with Python 3.\n\n Hold.\n Waiting for SoB on the last step.\n cf. <xmqqk14qqj7n.fsf@gitster-ct.c.googlers.com>\n\n\n* en/rebase-backend (2020-02-16) 20 commits\n - rebase: rename the two primary rebase backends\n - rebase: change the default backend from \"am\" to \"merge\"\n - rebase: make the backend configurable via config setting\n - rebase tests: repeat some tests using the merge backend instead of am\n - rebase tests: mark tests specific to the am-backend with --am\n - rebase: drop '-i' from the reflog for interactive-based rebases\n - git-prompt: change the prompt for interactive-based rebases\n - rebase: add an --am option\n - rebase: move incompatibility checks between backend options a bit earlier\n - git-rebase.txt: add more details about behavioral differences of backends\n - rebase: allow more types of rebases to fast-forward\n - t3432: make these tests work with either am or merge backends\n - rebase: fix handling of restrict_revision\n - rebase: make sure to pass along the quiet flag to the sequencer\n - rebase, sequencer: remove the broken GIT_QUIET handling\n - t3406: simplify an already simple test\n - rebase (interactive-backend): fix handling of commits that become empty\n - rebase (interactive-backend): make --keep-empty the default\n - t3404: directly test the behavior of interest\n - git-rebase.txt: update description of --allow-empty-message\n\n \"git rebase\" has learned to use the sequencer backend by default,\n while allowing \"--am\" option to go back to the traditional \"am\"\n backend.\n\n The last step may be rushing things a bit and may want to be\n taken separately.\n cf. <pull.679.v5.git.git.1581802602.gitgitgadget@gmail.com>\n"},{"id":"391960","messageId":"CABPp-BEFQs4LL0uwkZSpXb_wxo7Wk-gEQSsv-=4B5ja5H1xXTg@mail.gmail.com","threadId":"52829","inReplyTo":"xmqq36b8lvdl.fsf@gitster-ct.c.googlers.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Elijah Newren","fromEmail":"newren@gmail.com","sentAt":"2020-02-18T03:30:53Z","receivedAt":"2020-02-18T03:31:08Z","isPatch":false,"sender":{"key":"newren@gmail.com","avatar":"https://avatars.githubusercontent.com/u/5455730?v=4"},"body":"On Mon, Feb 17, 2020 at 2:09 PM Junio C Hamano <gitster@pobox.com> wrote:\n\n> * en/fill-directory-exponential (2020-01-31) 6 commits\n>  - t7063: blindly accept diffs\n>  - dir: replace exponential algorithm with a linear one\n>  - dir: refactor treat_directory to clarify control flow\n>  - dir: fix confusion based on variable tense\n>  - dir: fix broken comment\n>  - dir: consolidate treat_path() and treat_one_path()\n>\n>  The directory traversal code had redundant recursive calls which\n>  made its performance characteristics exponential wrt the depth of\n>  the tree, which was corrected.\n>\n>  Still RFC?\n>  cf. <pull.700.v2.git.git.1580495486.gitgitgadget@gmail.com>\n>  cf. <CABPp-BF0SO92wxY-dM14Vxv6BGA5=T=CicdwbiaZ7c9w6-j2jQ@mail.gmail.com>\n\nYes, sorry.  I was hoping someone familiar with untracked-cache (or\neven just the index format) could comment and provide some pointers,\nbut it looks like that won't be happening.  I'll try to dig in this\nweek to figure out what to do there.\n\n\n> * en/rebase-backend (2020-02-16) 20 commits\n>  - rebase: rename the two primary rebase backends\n>  - rebase: change the default backend from \"am\" to \"merge\"\n>  - rebase: make the backend configurable via config setting\n>  - rebase tests: repeat some tests using the merge backend instead of am\n>  - rebase tests: mark tests specific to the am-backend with --am\n>  - rebase: drop '-i' from the reflog for interactive-based rebases\n>  - git-prompt: change the prompt for interactive-based rebases\n>  - rebase: add an --am option\n>  - rebase: move incompatibility checks between backend options a bit earlier\n>  - git-rebase.txt: add more details about behavioral differences of backends\n>  - rebase: allow more types of rebases to fast-forward\n>  - t3432: make these tests work with either am or merge backends\n>  - rebase: fix handling of restrict_revision\n>  - rebase: make sure to pass along the quiet flag to the sequencer\n>  - rebase, sequencer: remove the broken GIT_QUIET handling\n>  - t3406: simplify an already simple test\n>  - rebase (interactive-backend): fix handling of commits that become empty\n>  - rebase (interactive-backend): make --keep-empty the default\n>  - t3404: directly test the behavior of interest\n>  - git-rebase.txt: update description of --allow-empty-message\n>\n>  \"git rebase\" has learned to use the sequencer backend by default,\n>  while allowing \"--am\" option to go back to the traditional \"am\"\n>  backend.\n>\n>  The last step may be rushing things a bit and may want to be\n>  taken separately.\n>  cf. <pull.679.v5.git.git.1581802602.gitgitgadget@gmail.com>\n\nThanks.  I'm curious, though, what you mean by this last bit -- was it\njust a reflection of my request for folks to comment on whether the\nlast patch was wanted?\n\nIn my view, taking the last patch separately does not make sense; it\nshould either be dropped entirely or taken close to the same time as\nthe series.  My reasoning for this is as follows: The only place the\ncurrent backend names exist is in the documentation.  The cost of\nchanging the names is thus low.  Once this series lands, the backend\nnames will be exposed in the user interface.  If we cut a release with\nthose names, then changing the names will require a bunch of\ntransition work.  So, if we're going to change the backend names, it'd\nbe better to do it sooner (while the cost is negligible) rather than\nlater.\n\nSo far, Phillip has voiced an opinion on this (cf.\n<e2863381-174c-a55c-bb22-0c7aec9cabf4@gmail.com>) -- \"I think it is a\ngood idea - merge and apply make much more sense that interactive and\nam\"; haven't heard from anyone else yet.\n"},{"id":"391991","messageId":"nycvar.QRO.7.76.6.2002181659540.46@tvgsbejvaqbjf.bet","threadId":"52829","inReplyTo":"CABPp-BEFQs4LL0uwkZSpXb_wxo7Wk-gEQSsv-=4B5ja5H1xXTg@mail.gmail.com","subject":"en/rebase-backend, was Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2020-02-18T16:00:55Z","receivedAt":"2020-02-18T16:01:06Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi Elijah,\n\nOn Mon, 17 Feb 2020, Elijah Newren wrote:\n\n> On Mon, Feb 17, 2020 at 2:09 PM Junio C Hamano <gitster@pobox.com> wrote:\n>\n> > * en/rebase-backend (2020-02-16) 20 commits\n> >  - rebase: rename the two primary rebase backends\n> >  - rebase: change the default backend from \"am\" to \"merge\"\n> >  - rebase: make the backend configurable via config setting\n> >  - rebase tests: repeat some tests using the merge backend instead of am\n> >  - rebase tests: mark tests specific to the am-backend with --am\n> >  - rebase: drop '-i' from the reflog for interactive-based rebases\n> >  - git-prompt: change the prompt for interactive-based rebases\n> >  - rebase: add an --am option\n> >  - rebase: move incompatibility checks between backend options a bit earlier\n> >  - git-rebase.txt: add more details about behavioral differences of backends\n> >  - rebase: allow more types of rebases to fast-forward\n> >  - t3432: make these tests work with either am or merge backends\n> >  - rebase: fix handling of restrict_revision\n> >  - rebase: make sure to pass along the quiet flag to the sequencer\n> >  - rebase, sequencer: remove the broken GIT_QUIET handling\n> >  - t3406: simplify an already simple test\n> >  - rebase (interactive-backend): fix handling of commits that become empty\n> >  - rebase (interactive-backend): make --keep-empty the default\n> >  - t3404: directly test the behavior of interest\n> >  - git-rebase.txt: update description of --allow-empty-message\n> >\n> >  \"git rebase\" has learned to use the sequencer backend by default,\n> >  while allowing \"--am\" option to go back to the traditional \"am\"\n> >  backend.\n> >\n> >  The last step may be rushing things a bit and may want to be\n> >  taken separately.\n> >  cf. <pull.679.v5.git.git.1581802602.gitgitgadget@gmail.com>\n>\n> Thanks.  I'm curious, though, what you mean by this last bit -- was it\n> just a reflection of my request for folks to comment on whether the\n> last patch was wanted?\n>\n> In my view, taking the last patch separately does not make sense; it\n> should either be dropped entirely or taken close to the same time as\n> the series.  My reasoning for this is as follows: The only place the\n> current backend names exist is in the documentation.  The cost of\n> changing the names is thus low.  Once this series lands, the backend\n> names will be exposed in the user interface.  If we cut a release with\n> those names, then changing the names will require a bunch of\n> transition work.  So, if we're going to change the backend names, it'd\n> be better to do it sooner (while the cost is negligible) rather than\n> later.\n>\n> So far, Phillip has voiced an opinion on this (cf.\n> <e2863381-174c-a55c-bb22-0c7aec9cabf4@gmail.com>) -- \"I think it is a\n> good idea - merge and apply make much more sense that interactive and\n> am\"; haven't heard from anyone else yet.\n\nFWIW I am also in favor of renaming the backends. Unfortunately, I won't\nhave time to review this patch series for the foreseeable time.\n\nCiao,\nDscho\n"},{"id":"392100","messageId":"20200219195822.GA5101@syl.local","threadId":"52829","inReplyTo":"xmqq36b8lvdl.fsf@gitster-ct.c.googlers.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Taylor Blau","fromEmail":"me@ttaylorr.com","sentAt":"2020-02-19T19:58:22Z","receivedAt":"2020-02-19T19:58:30Z","isPatch":false,"sender":{"key":"me@ttaylorr.com","avatar":"https://avatars.githubusercontent.com/u/301000140?v=4"},"body":"On Mon, Feb 17, 2020 at 02:08:38PM -0800, Junio C Hamano wrote:\n\n> * tb/commit-graph-split-merge (2020-02-12) 3 commits\n>  - builtin/commit-graph.c: support '--input=none'\n>  - builtin/commit-graph.c: introduce '--input=<source>'\n>  - builtin/commit-graph.c: support '--split[=<strategy>]'\n>\n>  The code to write out the commit-graph has been taught a few\n>  options to control if the resulting graph chains should be merged\n>  or a single new incremental graph is created.\n>\n>  Waiting for the discussion to settle.\n>  cf. <20200212205028.GE4364@syl.local>\n\nI think that this discussion is still \"settling\", although the thread\nhasn't had any activity for a few days now.\n\nI wanted to get feedback from my response to Eric Sunshine before saying\nthat we have consensus, but otherwise I think this is moving along.\n\nThanks,\nTaylor\n"},{"id":"392109","messageId":"CAPig+cQQdLaD58y+OLPELOWP_91ypHep68HMjwBBHYkZE_=6XA@mail.gmail.com","threadId":"52829","inReplyTo":"20200219195822.GA5101@syl.local","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Eric Sunshine","fromEmail":"sunshine@sunshineco.com","sentAt":"2020-02-19T20:50:10Z","receivedAt":"2020-02-19T20:50:24Z","isPatch":false,"sender":{"key":"sunshine@sunshineco.com","avatar":"https://avatars.githubusercontent.com/u/163641?v=4"},"body":"On Wed, Feb 19, 2020 at 2:58 PM Taylor Blau <me@ttaylorr.com> wrote:\n> On Mon, Feb 17, 2020 at 02:08:38PM -0800, Junio C Hamano wrote:\n> > * tb/commit-graph-split-merge (2020-02-12) 3 commits\n> >\n> >  Waiting for the discussion to settle.\n> >  cf. <20200212205028.GE4364@syl.local>\n>\n> I wanted to get feedback from my response to Eric Sunshine before saying\n> that we have consensus, but otherwise I think this is moving along.\n\nErm, I don't think Eric Sunshine participated in the discussion about\nthat series (or did I overlook something?).\n"},{"id":"392120","messageId":"xmqqy2syckju.fsf@gitster-ct.c.googlers.com","threadId":"52829","inReplyTo":"20200219195822.GA5101@syl.local","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2020-02-19T21:51:49Z","receivedAt":"2020-02-19T21:51:56Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Taylor Blau <me@ttaylorr.com> writes:\n\n> On Mon, Feb 17, 2020 at 02:08:38PM -0800, Junio C Hamano wrote:\n>\n>> * tb/commit-graph-split-merge (2020-02-12) 3 commits\n>>  - builtin/commit-graph.c: support '--input=none'\n>>  - builtin/commit-graph.c: introduce '--input=<source>'\n>>  - builtin/commit-graph.c: support '--split[=<strategy>]'\n>>\n>>  The code to write out the commit-graph has been taught a few\n>>  options to control if the resulting graph chains should be merged\n>>  or a single new incremental graph is created.\n>>\n>>  Waiting for the discussion to settle.\n>>  cf. <20200212205028.GE4364@syl.local>\n>\n> I think that this discussion is still \"settling\", although the thread\n> hasn't had any activity for a few days now.\n>\n> I wanted to get feedback from my response to Eric Sunshine before saying\n> that we have consensus, but otherwise I think this is moving along.\n\nDid you mean Szeder?\n\nhttps://lore.kernel.org/git/cover.1581486293.git.me@ttaylorr.com/\n\n"},{"id":"392122","messageId":"20200219220718.GD5101@syl.local","threadId":"52829","inReplyTo":"xmqqy2syckju.fsf@gitster-ct.c.googlers.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Taylor Blau","fromEmail":"me@ttaylorr.com","sentAt":"2020-02-19T22:07:18Z","receivedAt":"2020-02-19T22:07:22Z","isPatch":false,"sender":{"key":"me@ttaylorr.com","avatar":"https://avatars.githubusercontent.com/u/301000140?v=4"},"body":"On Wed, Feb 19, 2020 at 01:51:49PM -0800, Junio C Hamano wrote:\n> Taylor Blau <me@ttaylorr.com> writes:\n>\n> > On Mon, Feb 17, 2020 at 02:08:38PM -0800, Junio C Hamano wrote:\n> >\n> >> * tb/commit-graph-split-merge (2020-02-12) 3 commits\n> >>  - builtin/commit-graph.c: support '--input=none'\n> >>  - builtin/commit-graph.c: introduce '--input=<source>'\n> >>  - builtin/commit-graph.c: support '--split[=<strategy>]'\n> >>\n> >>  The code to write out the commit-graph has been taught a few\n> >>  options to control if the resulting graph chains should be merged\n> >>  or a single new incremental graph is created.\n> >>\n> >>  Waiting for the discussion to settle.\n> >>  cf. <20200212205028.GE4364@syl.local>\n> >\n> > I think that this discussion is still \"settling\", although the thread\n> > hasn't had any activity for a few days now.\n> >\n> > I wanted to get feedback from my response to Eric Sunshine before saying\n> > that we have consensus, but otherwise I think this is moving along.\n>\n> Did you mean Szeder?\n\nI did, whoops. Thanks for noticing (and sorry to Eric, who I see has\nresponded just above :-)).\n\n> https://lore.kernel.org/git/cover.1581486293.git.me@ttaylorr.com/\n\nThanks,\nTaylor\n"},{"id":"392269","messageId":"e9984c53-7266-c438-87f5-7f5474e3adce@gmail.com","threadId":"52829","inReplyTo":"xmqq36b8lvdl.fsf@gitster-ct.c.googlers.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Phillip Wood","fromEmail":"phillip.wood123@gmail.com","sentAt":"2020-02-21T16:26:18Z","receivedAt":"2020-02-21T16:26:25Z","isPatch":false,"sender":{"key":"phillip.wood@dunelm.org.uk","avatar":null},"body":"On 17/02/2020 22:08, Junio C Hamano wrote:\n> Here are the topics that have been cooking.  Commits prefixed with\n> '-' are only in 'pu' (proposed updates) while commits prefixed with\n> '+' are in 'next'.  The ones marked with '.' do not appear in any of\n> the integration branches, but I am still holding onto them.\n> \n> Accumulated fixes on the 'master' front have been flushed to 'maint'\n> and a new maintenance release 2.25.1 has been tagged.\n> [...]\n> [Stalled]\n> \n> * pw/advise-rebase-skip (2019-12-06) 9 commits\n>   - rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\n>   - rebase: fix advice when a fixup creates an empty commit\n>   - commit: give correct advice for empty commit during a rebase\n>   - commit: encapsulate determine_whence() for sequencer\n>   - commit: use enum value for multiple cherry-picks\n>   - sequencer: write CHERRY_PICK_HEAD for reword and edit\n>   - cherry-pick: check commit error messages\n>   - cherry-pick: add test for `--skip` advice in `git commit`\n>   - t3404: use test_cmp_rev\n> \n>   The mechanism to prevent \"git commit\" from making an empty commit\n>   or amending during an interrupted cherry-pick was broken during the\n>   rewrite of \"git rebase\" in C, which has been corrected.\n> \n>   What's the status of this one?\n>   The tip two are still RFC.\n\nThe tip \"rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\" \nneeds reworking and can be dropped (cf \n<7e1b92f5-48df-e202-ebcc-5b15987a7d63@gmail.com>). The other RFC patch \n\"rebase: fix advice when a fixup creates an empty commit\" [1] could do \nwith someone commenting on it (I've Cc'd dscho and Elijah). I think the \nmessages in it could be improved, but if the idea of different messages \nfor fixups that make a commit empty is not popular then the rest of the \nseries can be simplified by dropping some earlier patches including \npatch 6 which you had some doubts about. (I tired to address those \n<141f95b0-cae0-06f6-2c29-618dc22ae000@gmail.com> but I don't know if I \nconvinced you)\n\nBest Wishes\n\nPhillip\n\n[1] \nhttps://lore.kernel.org/git/20191206160614.631724-9-phillip.wood123@gmail.com/\n"},{"id":"392292","messageId":"CABPp-BE_ktNmDB43N_qijzYzhXvsK8Fi7TJQ7goHu+MzGvdpBQ@mail.gmail.com","threadId":"52829","inReplyTo":"e9984c53-7266-c438-87f5-7f5474e3adce@gmail.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Elijah Newren","fromEmail":"newren@gmail.com","sentAt":"2020-02-22T03:03:04Z","receivedAt":"2020-02-22T03:03:15Z","isPatch":false,"sender":{"key":"newren@gmail.com","avatar":"https://avatars.githubusercontent.com/u/5455730?v=4"},"body":"On Fri, Feb 21, 2020 at 8:26 AM Phillip Wood <phillip.wood123@gmail.com> wrote:\n>\n> On 17/02/2020 22:08, Junio C Hamano wrote:\n> > Here are the topics that have been cooking.  Commits prefixed with\n> > '-' are only in 'pu' (proposed updates) while commits prefixed with\n> > '+' are in 'next'.  The ones marked with '.' do not appear in any of\n> > the integration branches, but I am still holding onto them.\n> >\n> > Accumulated fixes on the 'master' front have been flushed to 'maint'\n> > and a new maintenance release 2.25.1 has been tagged.\n> > [...]\n> > [Stalled]\n> >\n> > * pw/advise-rebase-skip (2019-12-06) 9 commits\n> >   - rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\n> >   - rebase: fix advice when a fixup creates an empty commit\n> >   - commit: give correct advice for empty commit during a rebase\n> >   - commit: encapsulate determine_whence() for sequencer\n> >   - commit: use enum value for multiple cherry-picks\n> >   - sequencer: write CHERRY_PICK_HEAD for reword and edit\n> >   - cherry-pick: check commit error messages\n> >   - cherry-pick: add test for `--skip` advice in `git commit`\n> >   - t3404: use test_cmp_rev\n> >\n> >   The mechanism to prevent \"git commit\" from making an empty commit\n> >   or amending during an interrupted cherry-pick was broken during the\n> >   rewrite of \"git rebase\" in C, which has been corrected.\n> >\n> >   What's the status of this one?\n> >   The tip two are still RFC.\n>\n> The tip \"rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\"\n> needs reworking and can be dropped (cf\n> <7e1b92f5-48df-e202-ebcc-5b15987a7d63@gmail.com>). The other RFC patch\n> \"rebase: fix advice when a fixup creates an empty commit\" [1] could do\n> with someone commenting on it (I've Cc'd dscho and Elijah). I think the\n\nI'll try to take a look early to middle of next week.\n\n> messages in it could be improved, but if the idea of different messages\n> for fixups that make a commit empty is not popular then the rest of the\n> series can be simplified by dropping some earlier patches including\n> patch 6 which you had some doubts about. (I tired to address those\n> <141f95b0-cae0-06f6-2c29-618dc22ae000@gmail.com> but I don't know if I\n> convinced you)\n>\n> Best Wishes\n>\n> Phillip\n>\n> [1]\n> https://lore.kernel.org/git/20191206160614.631724-9-phillip.wood123@gmail.com/\n"},{"id":"392293","messageId":"CABPp-BEnt4C_7XyxQKxk4aga=JjM9fXCE-7SFp7azO_v5-pQYw@mail.gmail.com","threadId":"52829","inReplyTo":"CABPp-BEFQs4LL0uwkZSpXb_wxo7Wk-gEQSsv-=4B5ja5H1xXTg@mail.gmail.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Elijah Newren","fromEmail":"newren@gmail.com","sentAt":"2020-02-22T03:11:05Z","receivedAt":"2020-02-22T03:11:19Z","isPatch":false,"sender":{"key":"newren@gmail.com","avatar":"https://avatars.githubusercontent.com/u/5455730?v=4"},"body":"On Mon, Feb 17, 2020 at 7:30 PM Elijah Newren <newren@gmail.com> wrote:\n>\n> On Mon, Feb 17, 2020 at 2:09 PM Junio C Hamano <gitster@pobox.com> wrote:\n>\n> > * en/fill-directory-exponential (2020-01-31) 6 commits\n> >  - t7063: blindly accept diffs\n> >  - dir: replace exponential algorithm with a linear one\n> >  - dir: refactor treat_directory to clarify control flow\n> >  - dir: fix confusion based on variable tense\n> >  - dir: fix broken comment\n> >  - dir: consolidate treat_path() and treat_one_path()\n> >\n> >  The directory traversal code had redundant recursive calls which\n> >  made its performance characteristics exponential wrt the depth of\n> >  the tree, which was corrected.\n> >\n> >  Still RFC?\n> >  cf. <pull.700.v2.git.git.1580495486.gitgitgadget@gmail.com>\n> >  cf. <CABPp-BF0SO92wxY-dM14Vxv6BGA5=T=CicdwbiaZ7c9w6-j2jQ@mail.gmail.com>\n>\n> Yes, sorry.  I was hoping someone familiar with untracked-cache (or\n> even just the index format) could comment and provide some pointers,\n> but it looks like that won't be happening.  I'll try to dig in this\n> week to figure out what to do there.\n\nAfter some digging around, at least part of the final patch is wrong,\nand it points to me having somehow broken something with the untracked\ncache in the earlier patches.  (t7063.8 sets up a test where dthree/\nand dtwo/ should both be shown as untracked directories by 'git status\n--porcelain' but are not.  'git ls-files -o --directory', in contrast,\nwill show them as untracked.)  I probably won't have time to work on\nthese until after the Git Merge conference, but I will clearly need to\nfind and fix the bugs before this series progresses.  Feel free to\ndrop this series in the meantime if it's easier for you.\n\n\n> > * en/rebase-backend (2020-02-16) 20 commits\n> >  - rebase: rename the two primary rebase backends\n> >  - rebase: change the default backend from \"am\" to \"merge\"\n> >  - rebase: make the backend configurable via config setting\n> >  - rebase tests: repeat some tests using the merge backend instead of am\n> >  - rebase tests: mark tests specific to the am-backend with --am\n> >  - rebase: drop '-i' from the reflog for interactive-based rebases\n> >  - git-prompt: change the prompt for interactive-based rebases\n> >  - rebase: add an --am option\n> >  - rebase: move incompatibility checks between backend options a bit earlier\n> >  - git-rebase.txt: add more details about behavioral differences of backends\n> >  - rebase: allow more types of rebases to fast-forward\n> >  - t3432: make these tests work with either am or merge backends\n> >  - rebase: fix handling of restrict_revision\n> >  - rebase: make sure to pass along the quiet flag to the sequencer\n> >  - rebase, sequencer: remove the broken GIT_QUIET handling\n> >  - t3406: simplify an already simple test\n> >  - rebase (interactive-backend): fix handling of commits that become empty\n> >  - rebase (interactive-backend): make --keep-empty the default\n> >  - t3404: directly test the behavior of interest\n> >  - git-rebase.txt: update description of --allow-empty-message\n> >\n> >  \"git rebase\" has learned to use the sequencer backend by default,\n> >  while allowing \"--am\" option to go back to the traditional \"am\"\n> >  backend.\n> >\n> >  The last step may be rushing things a bit and may want to be\n> >  taken separately.\n> >  cf. <pull.679.v5.git.git.1581802602.gitgitgadget@gmail.com>\n\nIn contrast, this whole series I think is complete and ready to merge\ndown.  Both Phillip and Dscho have commented that they agree with the\ndirection of the final patch, Phillip has reviewed multiple rounds and\nsounds happy with it, and most of this series already spent some time\nin 'next' during the 2.25 cycle.\n\n> Thanks.  I'm curious, though, what you mean by this last bit -- was it\n> just a reflection of my request for folks to comment on whether the\n> last patch was wanted?\n>\n> In my view, taking the last patch separately does not make sense; it\n> should either be dropped entirely or taken close to the same time as\n> the series...[snip]...if we're going to change the backend names, it'd\n> be better to do it sooner (while the cost is negligible) rather than\n> later.\n"},{"id":"392300","messageId":"xmqq7e0e7d9z.fsf@gitster-ct.c.googlers.com","threadId":"52829","inReplyTo":"CABPp-BE_ktNmDB43N_qijzYzhXvsK8Fi7TJQ7goHu+MzGvdpBQ@mail.gmail.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2020-02-22T17:16:56Z","receivedAt":"2020-02-22T17:17:02Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Elijah Newren <newren@gmail.com> writes:\n\n>> > [Stalled]\n>> >\n>> > * pw/advise-rebase-skip (2019-12-06) 9 commits\n>> >   - rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\n>> >   - rebase: fix advice when a fixup creates an empty commit\n>> > ...\n>> The tip \"rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\"\n>> needs reworking and can be dropped (cf\n>> <7e1b92f5-48df-e202-ebcc-5b15987a7d63@gmail.com>). The other RFC patch\n>> \"rebase: fix advice when a fixup creates an empty commit\" [1] could do\n>> with someone commenting on it (I've Cc'd dscho and Elijah). I think the\n>\n> I'll try to take a look early to middle of next week.\n\nThanks, both.  I think the early ones except for those two make\nsense by themselves, so perhaps I'll advance them to 'next'\nindependently of these two steps.\n\n"},{"id":"392301","messageId":"xmqq36b27d7l.fsf@gitster-ct.c.googlers.com","threadId":"52829","inReplyTo":"CABPp-BEnt4C_7XyxQKxk4aga=JjM9fXCE-7SFp7azO_v5-pQYw@mail.gmail.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2020-02-22T17:18:22Z","receivedAt":"2020-02-22T17:18:29Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Elijah Newren <newren@gmail.com> writes:\n\n>> > * en/rebase-backend (2020-02-16) 20 commits\n>> >  - rebase: rename the two primary rebase backends\n>> >...\n>> >  - t3404: directly test the behavior of interest\n>> >  - git-rebase.txt: update description of --allow-empty-message\n>> >\n>> >  \"git rebase\" has learned to use the sequencer backend by default,\n>> >  while allowing \"--am\" option to go back to the traditional \"am\"\n>> >  backend.\n>> >\n>> >  The last step may be rushing things a bit and may want to be\n>> >  taken separately.\n>> >  cf. <pull.679.v5.git.git.1581802602.gitgitgadget@gmail.com>\n>\n> In contrast, this whole series I think is complete and ready to merge\n> down.  Both Phillip and Dscho have commented that they agree with the\n> direction of the final patch, Phillip has reviewed multiple rounds and\n> sounds happy with it, and most of this series already spent some time\n> in 'next' during the 2.25 cycle.\n\nLet's merge it down then.  Thanks.\n"},{"id":"392536","messageId":"3b63cd85-93b4-b6c1-6b42-c079691267dd@gmail.com","threadId":"52829","inReplyTo":"nycvar.QRO.7.76.6.2002181659540.46@tvgsbejvaqbjf.bet","subject":"Re: en/rebase-backend, was Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Alban Gruin","fromEmail":"alban.gruin@gmail.com","sentAt":"2020-02-26T10:34:49Z","receivedAt":"2020-02-26T10:35:04Z","isPatch":false,"sender":{"key":"alban.gruin@gmail.com","avatar":"https://avatars.githubusercontent.com/u/6310153?v=4"},"body":"Hi,\n\nLe 18/02/2020 à 17:00, Johannes Schindelin a écrit :\n> Hi Elijah,\n> \n> On Mon, 17 Feb 2020, Elijah Newren wrote:\n> \n>> On Mon, Feb 17, 2020 at 2:09 PM Junio C Hamano <gitster@pobox.com> wrote:\n>>\n>>> * en/rebase-backend (2020-02-16) 20 commits\n>>>  - rebase: rename the two primary rebase backends\n>>>  - rebase: change the default backend from \"am\" to \"merge\"\n>>>  - rebase: make the backend configurable via config setting\n>>>  - rebase tests: repeat some tests using the merge backend instead of am\n>>>  - rebase tests: mark tests specific to the am-backend with --am\n>>>  - rebase: drop '-i' from the reflog for interactive-based rebases\n>>>  - git-prompt: change the prompt for interactive-based rebases\n>>>  - rebase: add an --am option\n>>>  - rebase: move incompatibility checks between backend options a bit earlier\n>>>  - git-rebase.txt: add more details about behavioral differences of backends\n>>>  - rebase: allow more types of rebases to fast-forward\n>>>  - t3432: make these tests work with either am or merge backends\n>>>  - rebase: fix handling of restrict_revision\n>>>  - rebase: make sure to pass along the quiet flag to the sequencer\n>>>  - rebase, sequencer: remove the broken GIT_QUIET handling\n>>>  - t3406: simplify an already simple test\n>>>  - rebase (interactive-backend): fix handling of commits that become empty\n>>>  - rebase (interactive-backend): make --keep-empty the default\n>>>  - t3404: directly test the behavior of interest\n>>>  - git-rebase.txt: update description of --allow-empty-message\n>>>\n>>>  \"git rebase\" has learned to use the sequencer backend by default,\n>>>  while allowing \"--am\" option to go back to the traditional \"am\"\n>>>  backend.\n>>>\n>>>  The last step may be rushing things a bit and may want to be\n>>>  taken separately.\n>>>  cf. <pull.679.v5.git.git.1581802602.gitgitgadget@gmail.com>\n>>\n>> Thanks.  I'm curious, though, what you mean by this last bit -- was it\n>> just a reflection of my request for folks to comment on whether the\n>> last patch was wanted?\n>>\n>> In my view, taking the last patch separately does not make sense; it\n>> should either be dropped entirely or taken close to the same time as\n>> the series.  My reasoning for this is as follows: The only place the\n>> current backend names exist is in the documentation.  The cost of\n>> changing the names is thus low.  Once this series lands, the backend\n>> names will be exposed in the user interface.  If we cut a release with\n>> those names, then changing the names will require a bunch of\n>> transition work.  So, if we're going to change the backend names, it'd\n>> be better to do it sooner (while the cost is negligible) rather than\n>> later.\n>>\n>> So far, Phillip has voiced an opinion on this (cf.\n>> <e2863381-174c-a55c-bb22-0c7aec9cabf4@gmail.com>) -- \"I think it is a\n>> good idea - merge and apply make much more sense that interactive and\n>> am\"; haven't heard from anyone else yet.\n> \n> FWIW I am also in favor of renaming the backends. Unfortunately, I won't\n> have time to review this patch series for the foreseeable time.\n> \n\nI do agree too, for the same reasons as Elijah expressed.\n\nAlban\n\n> Ciao,\n> Dscho\n> \n\n"},{"id":"392558","messageId":"CABPp-BG_qoGFYhEMAtq0J8ZjS8SxC7WA=FjzWnNpM84CZTSVNw@mail.gmail.com","threadId":"52829","inReplyTo":"CABPp-BE_ktNmDB43N_qijzYzhXvsK8Fi7TJQ7goHu+MzGvdpBQ@mail.gmail.com","subject":"Re: What's cooking in git.git (Feb 2020, #04; Mon, 17)","fromName":"Elijah Newren","fromEmail":"newren@gmail.com","sentAt":"2020-02-26T20:04:08Z","receivedAt":"2020-02-26T20:04:21Z","isPatch":false,"sender":{"key":"newren@gmail.com","avatar":"https://avatars.githubusercontent.com/u/5455730?v=4"},"body":"On Fri, Feb 21, 2020 at 7:03 PM Elijah Newren <newren@gmail.com> wrote:\n>\n> On Fri, Feb 21, 2020 at 8:26 AM Phillip Wood <phillip.wood123@gmail.com> wrote:\n> >\n> > On 17/02/2020 22:08, Junio C Hamano wrote:\n> > > Here are the topics that have been cooking.  Commits prefixed with\n> > > '-' are only in 'pu' (proposed updates) while commits prefixed with\n> > > '+' are in 'next'.  The ones marked with '.' do not appear in any of\n> > > the integration branches, but I am still holding onto them.\n> > >\n> > > Accumulated fixes on the 'master' front have been flushed to 'maint'\n> > > and a new maintenance release 2.25.1 has been tagged.\n> > > [...]\n> > > [Stalled]\n> > >\n> > > * pw/advise-rebase-skip (2019-12-06) 9 commits\n> > >   - rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\n> > >   - rebase: fix advice when a fixup creates an empty commit\n> > >   - commit: give correct advice for empty commit during a rebase\n> > >   - commit: encapsulate determine_whence() for sequencer\n> > >   - commit: use enum value for multiple cherry-picks\n> > >   - sequencer: write CHERRY_PICK_HEAD for reword and edit\n> > >   - cherry-pick: check commit error messages\n> > >   - cherry-pick: add test for `--skip` advice in `git commit`\n> > >   - t3404: use test_cmp_rev\n> > >\n> > >   The mechanism to prevent \"git commit\" from making an empty commit\n> > >   or amending during an interrupted cherry-pick was broken during the\n> > >   rewrite of \"git rebase\" in C, which has been corrected.\n> > >\n> > >   What's the status of this one?\n> > >   The tip two are still RFC.\n> >\n> > The tip \"rebase -i: leave CHERRY_PICK_HEAD when there are conflicts\"\n> > needs reworking and can be dropped (cf\n> > <7e1b92f5-48df-e202-ebcc-5b15987a7d63@gmail.com>). The other RFC patch\n> > \"rebase: fix advice when a fixup creates an empty commit\" [1] could do\n> > with someone commenting on it (I've Cc'd dscho and Elijah). I think the\n>\n> I'll try to take a look early to middle of next week.\n>\n> > messages in it could be improved, but if the idea of different messages\n> > for fixups that make a commit empty is not popular then the rest of the\n> > series can be simplified by dropping some earlier patches including\n> > patch 6 which you had some doubts about. (I tired to address those\n> > <141f95b0-cae0-06f6-2c29-618dc22ae000@gmail.com> but I don't know if I\n> > convinced you)\n\nSorry for the delay.  I can see how different messages might be useful\nin the case where a fixup creates an empty commit, but I think it\nwould need more care since users won't know whether instructions are\nreferring to the combined changes of the original commit plus the\nfixup, or if the instructions about skipping/dropping/whatever are\njust referring to the changes from the fixup.  It wasn't clear to me.\n\nAdded a few other comments on that patch too.  Most of the patches\nearlier in the series looked good to me; I was slightly\nconcerned/confused initially by patch 6 and part of patch 8 for\nsimilar reasons on the determine_whence stuff.  This was only a minor\nconcern and I wouldn't hold up the patches over this, but I did I\nthrow an idea out there as an alternative -- though it might possibly\nbe worse.  (Nothing like throwing a bad idea out there to get creative\njuices flowing, right?)  I skipped patch 9 since you already mentioned\nit needs fixes.\n"}]}