{"thread":{"id":"57506","subject":"What's cooking in git.git (Mar 2022, #01; Thu, 3)","startedAt":"2022-03-04T04:31:30Z","lastAt":"2022-03-17T15:54:34Z","messageCount":22,"participants":["Junio C Hamano","Derrick Stolee","Ævar Arnfjörð Bjarmason","Jonathan Nieder","Taylor Blau","Michael J Gruber"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"450353","messageId":"xmqqv8wu2vag.fsf@gitster.g","threadId":"57506","inReplyTo":null,"subject":"What's cooking in git.git (Mar 2022, #01; Thu, 3)","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-03-04T04:31:19Z","receivedAt":"2022-03-04T04:31:30Z","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 in my tree. Commits\nprefixed with '+' are in 'next' (being in 'next' is a sign that a\ntopic is stable enough to be used and are candidate to be in a\nfuture release). Commits prefixed with '-' are only in 'seen',\nwhich means nothing more than that I have found them of interest for\nsome reason (like \"it may have hard-to-resolve conflicts with\nanother topic already in flight\" or \"this may turn out to be\nuseful\"). Do not read too much into a topic being in (or not in)\n'seen'. The ones marked with '.' do not appear in any of the\nintegration branches, but I am still holding onto them.\n\nCopies of the source code to Git live in many repositories, and the\nfollowing is a list of the ones I push into or their mirrors. Some\nrepositories have only a subset of branches.\n\nWith maint, master, next, seen, todo:\n\n\tgit://git.kernel.org/pub/scm/git/git.git/\n\tgit://repo.or.cz/alt-git.git/\n\thttps://kernel.googlesource.com/pub/scm/git/git/\n\thttps://github.com/git/git/\n\thttps://gitlab.com/git-vcs/git/\n\nWith all the integration branches and topics broken out:\n\n\thttps://github.com/gitster/git/\n\nEven though the preformatted documentation in HTML and man format\nare not sources, they are published in these repositories for\nconvenience (replace \"htmldocs\" with \"manpages\" for the manual\npages):\n\n\tgit://git.kernel.org/pub/scm/git/git-htmldocs.git/\n\thttps://github.com/gitster/git-htmldocs.git/\n\nRelease tarballs are available at:\n\n\thttps://www.kernel.org/pub/software/scm/git/\n\n--------------------------------------------------\n[New Topics]\n\n* jd/userdiff-kotlin (2022-03-03) 1 commit\n - userdiff: add builtin diff driver for Kotlin language.\n\n A new built-in userdiff driver for kotlin.\n\n source: <20220303181517.70682-1-jaydeepjd.8914@gmail.com>\n\n\n* ps/fetch-mirror-optim (2022-03-01) 6 commits\n - refs/files-backend: optimize reading of symbolic refs\n - remote: read symbolic refs via `refs_read_symbolic_ref()`\n - refs: add ability for backends to special-case reading of symbolic refs\n - fetch: avoid lookup of commits when not appending to FETCH_HEAD\n - upload-pack: look up \"want\" lines via commit-graph\n - Merge branch 'ps/fetch-atomic' into ps/fetch-mirror-optim\n (this branch uses ps/fetch-atomic.)\n\n Various optimization for \"git fetch\".\n\n Will merge to 'next'.\n source: \n\n\n* tk/t7063-chmtime-dirs-too (2022-03-01) 2 commits\n - t7063: mtime-mangling instead of delays in untracked cache testing\n - t/helper/test-chmtime: update mingw to support chmtime on directories\n\n Teach \"test-chmtime\" to work on a directory and use it to avoid\n having to wait for a second in a few places in tests.\n\n Expecting a reroll.\n cf. <37f66fae-911a-62f1-4711-c9f6c7794911@jeffhostetler.com>\n source: \n\n\n* ab/hook-tests-updates (2022-03-02) 10 commits\n - tests: change \"cat && chmod +x\" to use \"test_hook\"\n - tests: change \"mkdir -p && write_script\" to use \"test_hook\"\n - test-lib-functions: add and use a \"test_hook\" wrapper\n - fetch+push tests: have tests clean up their own mess\n - hook tests: get rid of unnecessary sub-shells\n - tests: indent and add hook setup to \"test_expect_success\"\n - bugreport tests: tighten up \"git bugreport -s hooks\" test\n - tests: assume the hooks are disabled by default\n - t5540: don't rely on \"hook/post-update.sample\"\n - hook tests: turn exit code assertions into a loop\n\n Update tests around the use of hook scripts.\n\n Expecting a reroll.\n cf. \n cf. \n source: \n\n\n* ab/keep-git-exit-codes-in-tests (2022-03-02) 15 commits\n - rev-list simplify tests: don't ignore \"git\" exit code\n - checkout tests: don't ignore \"git \" exit code\n - apply tests: don't ignore \"git ls-files\" exit code, drop sub-shell\n - gettext tests: don't ignore \"test-tool regex\" exit code\n - rev-list tests: don't hide abort() in \"test_expect_failure\"\n - diff tests: don't ignore \"git rev-list\" exit code\n - notes tests: don't ignore \"git\" exit code\n - rev-parse tests: don't ignore \"git reflog\" exit code\n - merge tests: use \"test_must_fail\" instead of ad-hoc pattern\n - apply tests: use \"test_must_fail\" instead of ad-hoc pattern\n - diff tests: don't ignore \"git diff\" exit code in \"read\" loop\n - diff tests: don't ignore \"git diff\" exit code\n - read-tree tests: check \"diff-files\" exit code on failure\n - tests: use \"test_stdout_line_count\", not \"test $(git [...] | wc -l)\"\n - tests: change some 'test $(git) = \"x\"' to test_cmp\n\n Updates tests around the use of \"test $(git cmd) = constant\".\n\n Expecting a reroll.\n cf. \n source: \n\n\n* ab/plug-random-leaks (2022-03-02) 14 commits\n - repository.c: free the \"path cache\" in repo_clear()\n - range-diff: plug memory leak in read_patches()\n - range-diff: plug memory leak in common invocation\n - lockfile API users: simplify and don't leak \"path\"\n - commit-graph: stop fill_oids_from_packs() progress on error and free()\n - commit-graph: fix memory leak in misused string_list API\n - submodule--helper: fix trivial leak in module_add()\n - transport: stop needlessly copying bundle header references\n - bundle: call strvec_clear() on allocated strvec\n - remote-curl.c: free memory in cmd_main()\n - urlmatch.c: add and use a *_release() function\n - diff.c: free \"buf\" in diff_words_flush()\n - merge-base: free() allocated \"struct commit **\" list\n - index-pack: fix memory leaks\n\n Plug random memory leaks.\n\n Will merge to 'next'.\n source: \n\n\n* en/merge-ort-align-verbosity-with-recursive (2022-03-01) 1 commit\n - merge-ort: exclude messages from inner merges by default\n\n Align the level of verbose output from the ort backend during inner\n merge to that of the recursive backend.\n\n Will merge to 'next'.\n source: \n\n\n* ep/test-malloc-check-with-glibc-2.34 (2022-03-03) 1 commit\n - test-lib.sh: Use GLIBC_TUNABLES instead of MALLOC_CHECK_ on glibc >= 2.34\n\n The method to trigger malloc check used in our tests no longer work\n with newer versions of glibc.\n\n Expecting a reroll.\n cf. \n source: <20220303090640.190307-1-gitter.spiros@gmail.com>\n\n\n* gc/parse-tree-indirect-errors (2022-03-01) 1 commit\n - checkout, clone: die if tree cannot be parsed\n\n Check the return value from parse_tree_indirect() to turn segfaults\n into calls to die().\n\n Will merge to 'next'.\n source: <20220302003613.15567-1-chooglen@google.com>\n\n\n* jc/mailsplit-warn-on-tty (2022-03-03) 1 commit\n - am/apply: warn if we end up reading patches from terminal\n\n \"git am\" can read from the standard input when no mailbox is given\n on the command line, but the end-user gets no indication when it\n happens, making Git appear stuck.\n\n source: \n\n\n* jc/stash-drop (2022-03-02) 3 commits\n - stash: call reflog_delete() in reflog.c\n - reflog: libify delete reflog function and helpers\n - stash: add tests to ensure reflog --rewrite --updatref behavior\n\n \"git stash drop\" is reimplemented as an internal call to\n reflog_delete() function, instead of invoking \"git reflog delete\"\n via run_command() API.\n\n Will merge to 'next'?\n source: \n\n\n* nj/read-tree-doc-reffix (2022-03-03) 1 commit\n - Documentation: git-read-tree: separate links using commas\n\n Documentation mark-up fix.\n\n Will merge to 'next'.\n source: <20220303161542.11616-1-nihal@nihaljere.xyz>\n\n\n* ps/fetch-atomic-fixup (2022-03-03) 1 commit\n - t5503: simplify setup of test which exercises failure of backfill\n (this branch uses ps/fetch-atomic.)\n\n Test simplification.\n\n Will merge to 'next'.\n source: \n\n\n* tb/cruft-packs (2022-03-02) 17 commits\n - sha1-file.c: don't freshen cruft packs\n - builtin/gc.c: conditionally avoid pruning objects via loose\n - builtin/repack.c: add cruft packs to MIDX during geometric repack\n - builtin/repack.c: use named flags for existing_packs\n - builtin/repack.c: allow configuring cruft pack generation\n - builtin/repack.c: support generating a cruft pack\n - builtin/pack-objects.c: --cruft with expiration\n - reachable: report precise timestamps from objects in cruft packs\n - reachable: add options to add_unseen_recent_objects_to_traversal\n - builtin/pack-objects.c: --cruft without expiration\n - builtin/pack-objects.c: return from create_object_entry()\n - t/helper: add 'pack-mtimes' test-tool\n - pack-mtimes: support writing pack .mtimes files\n - chunk-format.h: extract oid_version()\n - pack-write: pass 'struct packing_data' to 'stage_tmp_packfiles'\n - pack-mtimes: support reading .mtimes files\n - Documentation/technical: add cruft-packs.txt\n\n A mechanism to pack unreachable objects into a \"cruft pack\",\n instead of ejecting them into loose form to be reclaimed later, has\n been introduced.\n\n Will merge to 'next'?\n source: \n\n\n* tb/rename-remote-progress (2022-03-03) 2 commits\n - builtin/remote.c: show progress when renaming remote references\n - builtin/remote.c: parse options in 'rename'\n\n \"git remote rename A B\", depending on the number of remote-tracking\n refs involved, takes long time renaming them. The command has been\n taught to show progress bar while making the user wait.\n\n Will merge to 'next'?\n source: \n\n--------------------------------------------------\n[Stalled]\n\n* tl/ls-tree-oid-only (2022-02-08) 13 commits\n - ls-tree.c: support --object-only option for \"git-ls-tree\"\n - ls-tree: introduce function \"fast_path()\"\n - ls-tree.c: introduce \"--format\" option\n - cocci: allow padding with `strbuf_addf()`\n - ls-tree: introduce struct \"show_tree_data\"\n - ls-tree: slightly refactor `show_tree()`\n - ls-tree: fix \"--name-only\" and \"--long\" combined use bug\n - ls-tree: simplify nesting if/else logic in \"show_tree()\"\n - ls-tree: rename \"retval\" to \"recurse\" in \"show_tree()\"\n - ls-tree: use \"size_t\", not \"int\" for \"struct strbuf\"'s \"len\"\n - ls-tree: use \"enum object_type\", not {blob,tree,commit}_type\n - ls-tree: add missing braces to \"else\" arms\n - ls-tree: remove commented-out code\n\n \"git ls-tree\" learns \"--oid-only\" option, similar to \"--name-only\",\n and more generalized \"--format\" option.\n source: \n\n\n* je/http-better-error-output (2021-12-03) 1 commit\n . http-backend: give a hint that web browser access is not supported\n\n When the http-backend program, which is the server-side component\n for the smart HTTP transport, sends a \"404 Not found\" error, we\n deliberately did not say anything to the requesting client. We now\n send a message back to the browser to tell the user that they do\n not want to visit the URL via their browser, instead of a totally\n blank page.\n\n Expecting a reroll.\n Breaks its self tests.\n cf. <7r23s082-o3q0-479o-srqn-r45q778s5nq7@vanv.qr>\n source: <20211202102855.23907-1-jengelh@inai.de>\n\n\n* cb/save-term-across-editor-invocation (2021-12-01) 3 commits\n - fixup! editor: allow for saving/restoring terminal state\n - editor: allow for saving/restoring terminal state\n - terminal: teach save_term to fail when not foreground\n\n Some editors are reported to leave the terminal in funny state\n after they exit on Windows. Work it around by saving and restoring\n the terminal state when needed.\n\n Expecting a reroll.\n cf. \n source: <20211202035446.1154-1-carenas@gmail.com>\n\n--------------------------------------------------\n[Cooking]\n\n* js/ci-github-workflow-markup (2022-03-01) 9 commits\n - ci: call `finalize_test_case_output` a little later\n - ci: use `--github-workflow-markup` in the GitHub workflow\n - ci: optionally mark up output in the GitHub workflow\n - test(junit): avoid line feeds in XML attributes\n - tests: refactor --write-junit-xml code\n - ci/run-build-and-tests: add some structure to the GitHub workflow output\n - ci: make it easier to find failed tests' logs in the GitHub workflow\n - ci/run-build-and-tests: take a more high-level view\n - ci: fix code style\n\n Update the GitHub workflow support to make it quicker to get to the\n failing test.\n\n Will merge to 'next'?\n cf. <220302.86mti87cj2.gmgdl@evledraar.gmail.com>\n cf. <30dbc8fb-a1db-05bc-3dcb-070e11cf4715@gmail.com>\n source: \n\n\n* jk/name-rev-w-genno (2022-02-28) 1 commit\n - name-rev: use generation numbers if available\n\n \"git name-rev\" learned to use the generation numbers when setting\n the lower bound of searching commits used to explain the revision,\n when available, instead of committer time.\n\n Will merge to 'next'?\n source: <20220228215025.325904-2-jacob.e.keller@intel.com>\n\n\n* et/xdiff-indirection (2022-02-17) 1 commit\n - xdiff: provide indirection to git functions\n\n Insert a layer of preprocessor macros for common functions in xdiff\n codebase.\n\n Expecting a (hopefully final) reroll.\n cf. \n source: <20220217225408.GB7@edef91d97c94>\n\n\n* jt/ls-files-stage-recurse (2022-02-23) 1 commit\n (merged to 'next' on 2022-02-28 at 4ad82bfdf6)\n + ls-files: support --recurse-submodules --stage\n\n Many output modes of \"ls-files\" does not work with its\n \"--recurse-submodules\" option, but the \"-s\" mode has taught to work\n with it.\n\n Will merge to 'master'.\n source: <20220224002303.2363189-1-jonathantanmy@google.com>\n\n\n* ab/c99-variadic-macros (2022-02-21) 2 commits\n (merged to 'next' on 2022-03-01 at 0419a86e23)\n + C99: remove hardcoded-out !HAVE_VARIADIC_MACROS code\n + git-compat-util.h: clarify GCC v.s. C99-specific in comment\n\n Remove the escape hatch we added when we introduced the weather\n balloon to use variadic macros unconditionally, to make it official\n that we now have a hard dependency on the feature.\n\n Will merge to 'master'.\n source: \n\n\n* ab/hook-tests (2022-02-19) 2 commits\n (merged to 'next' on 2022-02-28 at bb6b7d1c24)\n + hook tests: use a modern style for \"pre-push\" tests\n + hook tests: test for exact \"pre-push\" hook input\n\n Test modernization.\n\n Will merge to 'master'.\n source: \n\n\n* ab/test-leak-diag (2022-02-28) 4 commits\n (merged to 'next' on 2022-02-28 at 3019aad7fe)\n + test-lib: add \"fast_unwind_on_malloc=0\" to LSAN_OPTIONS\n + test-lib: make $GIT_BUILD_DIR an absolute path\n + test-lib: correct and assert TEST_DIRECTORY overriding\n + test-lib: add GIT_SAN_OPTIONS, inherit [AL]SAN_OPTIONS\n\n Random test-framework clean-up.\n\n Will merge to 'master'.\n source: \n\n\n* ds/worktree-docs (2022-02-23) 11 commits\n (merged to 'next' on 2022-02-28 at b7d23cb1a5)\n + worktree: use 'worktree' over 'working tree'\n + worktree: use 'worktree' over 'working tree'\n + worktree: use 'worktree' over 'working tree'\n + worktree: use 'worktree' over 'working tree'\n + worktree: use 'worktree' over 'working tree'\n + worktree: use 'worktree' over 'working tree'\n + worktree: use 'worktree' over 'working tree'\n + worktree: extract checkout_worktree()\n + worktree: extract copy_sparse_checkout()\n + worktree: extract copy_filtered_worktree_config()\n + worktree: combine two translatable messages\n\n Tighten the language around \"working tree\" and \"worktree\" in the\n docs.\n\n Will merge to 'master'.\n source: \n\n\n* en/merge-ort-plug-leaks (2022-02-20) 2 commits\n (merged to 'next' on 2022-02-28 at d7afb6b624)\n + merge-ort: fix small memory leak in unique_path()\n + merge-ort: fix small memory leak in detect_and_process_renames()\n\n Leakfix.\n\n Will merge to 'master'.\n source: \n\n\n* jc/rerere-train-modernise (2022-02-27) 1 commit\n (merged to 'next' on 2022-02-28 at d86d29733d)\n + rerere-train: two fixes to the use of \"git show -s\"\n\n Small modernization of the rerere-train script (in contrib/).\n\n Will merge to 'master'.\n source: <20220227220924.2144325-1-gitster@pobox.com>\n\n\n* ab/help-fixes (2022-02-23) 9 commits\n (merged to 'next' on 2022-03-01 at 215f5dbbfb)\n + help: don't print \"\\n\" before single-section output\n + help: add --no-[external-commands|aliases] for use with --all\n + help: error if [-a|-g|-c] and [-i|-m|-w] are combined\n + help: correct usage & behavior of \"git help --all\"\n + help: note the option name on option incompatibility\n + help.c: split up list_all_cmds_help() function\n + help tests: test \"git\" and \"git help [-a|-g] spacing\n + help.c: use puts() instead of printf{,_ln}() for consistency\n + help doc: add missing \"]\" to \"[-a|--all]\"\n\n Updates to how command line options to \"git help\" are handled.\n\n Will merge to 'master'.\n source: \n\n\n* ac/usage-string-fixups (2022-02-23) 1 commit\n (merged to 'next' on 2022-02-28 at 5e8fa3f846)\n + amend remaining usage strings according to style guide\n\n Usage-string normalization.\n\n Will merge to 'master'.\n source: \n\n\n* mc/index-pack-report-max-size (2022-02-23) 1 commit\n (merged to 'next' on 2022-02-28 at 31b907844b)\n + index-pack: clarify the breached limit\n\n When \"index-pack\" dies due to incoming data exceeding the maximum\n allowed input size, include the value of the limit in the error\n message.\n\n Will merge to 'master'.\n source: \n\n\n* sm/no-git-in-upstream-of-pipe-in-tests (2022-02-23) 1 commit\n - t0003: avoid pipes with Git on LHS\n\n Test fixes.\n\n Will merge to 'next'.\n source: <20220223115347.3083-3-shivam828787@gmail.com>\n\n\n* ab/c99-designated-initializers (2022-02-24) 9 commits\n (merged to 'next' on 2022-02-28 at 5ff6d5b26f)\n + fast-import.c: use designated initializers for \"partial\" struct assignments\n + refspec.c: use designated initializers for \"struct refspec_item\"\n + convert.c: use designated initializers for \"struct stream_filter*\"\n + userdiff.c: use designated initializers for \"struct userdiff_driver\"\n + archive-*.c: use designated initializers for \"struct archiver\"\n + object-file: use designated initializers for \"struct git_hash_algo\"\n + trace2: use designated initializers for \"struct tr2_dst\"\n + trace2: use designated initializers for \"struct tr2_tgt\"\n + imap-send.c: use designated initializers for \"struct imap_server_conf\"\n\n Use designated initializers we started using in mid 2017 in more\n parts of the codebase that are relatively quiescent.\n\n Will merge to 'master'.\n source: \n\n\n* ab/http-gcc-12-workaround (2022-02-25) 1 commit\n - http API: fix dangling pointer issue noted by GCC 12.0\n\n Work around false warning pre-release of GCC 12.\n source: \n\n\n* ab/make-optim-noop (2022-03-03) 8 commits\n - Makefiles: add and use wildcard \"mkdir -p\" template\n - Makefile: add \"$(QUIET)\" boilerplate to shared.mak\n - Makefile: move $(comma), $(empty) and $(space) to shared.mak\n - Makefile: move \".SUFFIXES\" rule to shared.mak\n - Makefile: define $(LIB_H) in terms of $(FIND_SOURCE_FILES)\n - Makefile: disable GNU make built-in wildcard rules\n - Makefiles: add \"shared.mak\", move \".DELETE_ON_ERROR\" to it\n - scalar Makefile: use \"The default target of...\" pattern\n\n Makefile refactoring with a bit of suffixes rule stripping to\n optimize the runtime overhead.\n\n Will merge to 'next'.\n source: \n\n\n* ah/advice-switch-requires-detach-to-detach (2022-02-25) 1 commit\n (merged to 'next' on 2022-02-28 at 941ee621cf)\n + switch: mention the --detach option when dying due to lack of a branch\n\n The error message given by \"git switch HEAD~4\" has been clarified\n to suggest the \"--detach\" option that is required.\n\n Will merge to 'master'.\n source: <20220226061213.1590341-1-alexhenrie24@gmail.com>\n\n\n* ds/commit-graph-gen-v2-fixes (2022-03-01) 5 commits\n - commit-graph: fix generation number v2 overflow values\n - commit-graph: start parsing generation v2 (again)\n - commit-graph: fix ordering bug in generation numbers\n - t5318: extract helpers to lib-commit-graph.sh\n - test-read-graph: include extra post-parse info\n\n Fixes to the way generation number v2 in the commit-graph files are\n (not) handled.\n source: \n\n\n* ds/partial-bundles (2022-02-24) 11 commits\n - bundle: unbundle promisor packs\n - bundle: create filtered bundles\n - rev-list: move --filter parsing into revision.c\n - bundle: parse filter capability\n - bundle: safely handle --objects option\n - MyFirstObjectWalk: update recommended usage\n - list-objects: consolidate traverse_commit_list[_filtered]\n - pack-bitmap: drop filter in prepare_bitmap_walk()\n - pack-objects: use rev.filter when possible\n - revision: put object filter into struct rev_info\n - index-pack: document and test the --promisor option\n (this branch is used by ds/bundle-uri.)\n\n Bundle file format gets extended to allow a partial bundle,\n filtered by similar criteria you would give when making a\n partial/lazy clone.\n source: \n\n\n* fs/gpgsm-update (2022-03-02) 3 commits\n - t/lib-gpg: kill all gpg components, not just gpg-agent\n - t/lib-gpg: reload gpg components after updating trustlist\n - gpg-interface/gpgsm: fix for v2.3\n\n Newer version of GPGSM changed its output in a backward\n incompatible way to break our code that parses its output. It also\n added more processes our tests need to kill when cleaning up.\n Adjustments have been made to accomodate these changes.\n\n Will merge to 'next'.\n source: <20220224100628.612789-1-fs@gigacodes.de>\n\n\n* gc/stash-on-branch-with-multi-level-name (2022-02-24) 1 commit\n (merged to 'next' on 2022-02-28 at 1a94306611)\n + stash: strip \"refs/heads/\" with skip_prefix\n\n \"git checkout -b branch/with/multi/level/name && git stash\" only\n recorded the last level component of the branch name, which has\n been corrected.\n\n Will merge to 'master'.\n source: <20220124205342.41450-1-chooglen@google.com>\n\n\n* jd/prompt-upstream-mark (2022-02-25) 4 commits\n - git-prompt: put upstream comments together\n - git-prompt: make long upstream state indicator consistent\n - git-prompt: make upstream state indicator location consistent\n - git-prompt: rename `upstream` to `upstream_type`\n\n Tweaks in the command line prompt (in contrib/) code around its\n GIT_PS1_SHOWUPSTREAM feature.\n\n Needs reviews.\n source: \n\n\n* pw/worktree-list-with-z (2022-02-25) 1 commit\n - worktree: add -z option for list subcommand\n\n \"git worktree list --porcelain\" did not c-quote pathnames and lock\n reasons with unsafe bytes correctly, which is worked around by\n introducing NUL terminated output format with \"-z\".\n\n Expecting a reroll.\n cf. \n source: \n\n\n* rc/fetch-repair (2022-02-24) 8 commits\n - doc/partial-clone: mention --repair fetch option\n - fetch: after repair, encourage auto gc repacking\n - maintenance: add ability to pass config options\n - t5615-partial-clone: add test for fetch --repair\n - fetch: add --repair option\n - builtin/fetch-pack: add --repair option\n - fetch-pack: add repairing\n - fetch-negotiator: add specific noop initializor\n\n \"git fetch --repair\" learned to fetch everything without telling\n the other side what we already have, which is useful when you\n cannot trust what you have in the local object store.\n source: \n\n\n* tk/empty-untracked-cache (2022-02-28) 3 commits\n - untracked-cache: write index when populating empty untracked cache\n - t7519: populate untracked cache before test\n - t7519: avoid file to index mtime race for untracked cache\n\n The untracked cache newly computed weren't written back to the\n on-disk index file when there is no other change to the index,\n which has been corrected.\n\n Will merge to 'next'.\n source: \n\n\n* tk/simple-autosetupmerge (2022-02-25) 2 commits\n - t3200: tests for new branch.autosetupmerge option \"simple\"\n - merge: new autosetupmerge option 'simple' for matching branches\n\n \"git -c branch.autosetupmerge=simple branch $A $B\" will set the $B\n as $A's upstream only when $A and $B shares the same name, and \"git\n -c push.default=simple\" on branch $A would push to update the\n branch $A at the remote $B came from.\n\n Needs review.\n source: \n\n\n* tk/untracked-cache-with-uall (2022-02-25) 1 commit\n - untracked-cache: support '--untracked-files=all' if configured\n\n The performance of the \"untracked cache\" feature has been improved\n when \"--untracked-files=\" and \"status.showUntrackedFiles\"\n are combined.\n\n Expecting a reroll.\n source: \n\n\n* vd/sparse-read-tree (2022-03-01) 8 commits\n - read-tree: make three-way merge sparse-aware\n - read-tree: make two-way merge sparse-aware\n - read-tree: narrow scope of index expansion for '--prefix'\n - read-tree: integrate with sparse index\n - read-tree: expand sparse checkout test coverage\n - read-tree: explicitly disallow prefixes with a leading '/'\n - status: fix nested sparse directory diff in sparse index\n - sparse-index: prevent repo root from becoming sparse\n\n \"git read-tree\" has been made to be aware of the sparse-index\n feature.\n\n Will merge to 'next'.\n source: \n\n\n* en/sparse-checkout-fixes (2022-02-20) 5 commits\n (merged to 'next' on 2022-02-28 at 7307431c62)\n + sparse-checkout: reject arguments in cone-mode that look like patterns\n + sparse-checkout: error or warn when given individual files\n + sparse-checkout: pay attention to prefix for {set, add}\n + sparse-checkout: correctly set non-cone mode when expected\n + sparse-checkout: correct reapply's handling of options\n\n Further polishing of \"git sparse-checkout\".\n\n Will merge to 'master'.\n source: \n\n\n* ab/commit-plug-leaks (2022-02-16) 2 commits\n - commit: use strbuf_release() instead of UNLEAK()\n - commit: fix \"author_ident\" leak\n\n Leakfixes in the top-level called-once function.\n\n Expecting a reroll.\n I think UNLEAK->strbuf_release() is a regression.\n source: \n\n\n* pw/single-key-interactive (2022-02-23) 4 commits\n - add -p: disable stdin buffering when interactive.singlekey is set\n - terminal: set VMIN and VTIME in non-canonical mode\n - terminal: pop signal handler when terminal is restored\n - terminal: always reset terminal when reading without echo\n\n The single-key interactive operation used by \"git add -p\" has been\n made more robust.\n\n Will merge to 'next'?\n source: \n\n\n* jh/builtin-fsmonitor-part3 (2022-03-01) 23 commits\n - t/perf/p7527: add perf test for builtin FSMonitor\n - t7527: FSMonitor tests for directory moves\n - fsmonitor: optimize processing of directory events\n - fsm-listen-darwin: shutdown daemon if worktree root is moved/renamed\n - fsm-health-win32: force shutdown daemon if worktree root moves\n - fsm-health-win32: add framework to monitor daemon health\n - fsmonitor--daemon: stub in health thread\n - fsmonitor--daemon: rename listener thread related variables\n - fsmonitor--daemon: prepare for adding health thread\n - fsmonitor--daemon: cd out of worktree root\n - fsm-listen-darwin: ignore FSEvents caused by xattr changes on macOS\n - unpack-trees: initialize fsmonitor_has_run_once in o->result\n - fsmonitor-settings: remote repos on Windows are incompatible with FSMonitor\n - fsmonitor-settings: remote repos on macOS are incompatible with FSMonitor\n - fsmonitor-settings: stub in macOS-specific incompatibility checking\n - fsmonitor-settings: virtual repos are incompatible with FSMonitor\n - fsmonitor-settings: stub in platform-specific incompatibility checking\n - fsmonitor-settings: bare repos are incompatible with FSMonitor\n - t/helper/fsmonitor-client: create stress test\n - t7527: test builtin FSMonitor watching repos with unicode paths\n - t7527: test FS event reporing on macOS WRT case and Unicode\n - fsm-listen-win32: handle shortnames\n - Merge branch 'jh/builtin-fsmonitor-part2' into jh/builtin-fsmonitor-part3\n (this branch uses jh/builtin-fsmonitor-part2.)\n\n More fsmonitor--daemon.\n\n Expecting a reroll on the updated \"part 2\".\n source: \n\n\n* hn/reftable-no-empty-keys (2022-02-23) 7 commits\n (merged to 'next' on 2022-03-01 at dc4aa7e796)\n + reftable: rename writer_stats to reftable_writer_stats\n + reftable: add test for length of disambiguating prefix\n + reftable: ensure that obj_id_len is >= 2 on writing\n + reftable: avoid writing empty keys at the block layer\n + reftable: add a test that verifies that writing empty keys fails\n + reftable: reject 0 object_id_len\n + Documentation: object_id_len goes up to 31\n\n General clean-up in reftable implementation, including\n clarification of the API documentation, tightening the code to\n honor documented length limit, etc.\n\n Will merge to 'master'.\n source: \n\n\n* ps/fetch-atomic (2022-02-17) 7 commits\n (merged to 'next' on 2022-02-18 at 3824153b23)\n + fetch: make `--atomic` flag cover pruning of refs\n + fetch: make `--atomic` flag cover backfilling of tags\n + refs: add interface to iterate over queued transactional updates\n + fetch: report errors when backfilling tags fails\n + fetch: control lifecycle of FETCH_HEAD in a single place\n + fetch: backfill tags before setting upstream\n + fetch: increase test coverage of fetches\n (this branch is used by ps/fetch-atomic-fixup and ps/fetch-mirror-optim.)\n\n \"git fetch\" can make two separate fetches, but ref updates coming\n from them were in two separate ref transactions under \"--atomic\",\n which has been corrected.\n\n source: \n\n\n* pw/xdiff-alloc-fail (2022-02-16) 4 commits\n (merged to 'next' on 2022-03-01 at 8a10912059)\n + xdiff: handle allocation failure when merging\n + xdiff: refactor a function\n + xdiff: handle allocation failure in patience diff\n + xdiff: fix a memory leak\n\n Improve failure case behaviour of xdiff library when memory\n allocation fails.\n\n Will merge to 'master'.\n source: \n\n\n* gc/recursive-fetch-with-unused-submodules (2022-02-28) 12 commits\n - SQUASH???\n - SQUASH???\n - submodule: fix latent check_has_commit() bug\n - fetch: fetch unpopulated, changed submodules\n - submodule: move logic into fetch_task_create()\n - submodule: extract get_fetch_task()\n - submodule: store new submodule commits oid_array in a struct\n - submodule: inline submodule_commits() into caller\n - submodule: make static functions read submodules from commits\n - t5526: create superproject commits with test helper\n - t5526: stop asserting on stderr literally\n - t5526: introduce test helper to assert on fetches\n\n When \"git fetch --recurse-submodules\" grabbed submodule commits\n that would be needed to recursively check out newly fetched commits\n in the superproject, it only paid attention to submodules that are\n in the current checkout of the superproject. We now do so for all\n submodules that have been run \"git submodule init\" on.\n\n Expecting a reroll.\n source: <20220224100842.95827-1-chooglen@google.com>\n\n\n* cg/t3903-modernize (2022-02-23) 3 commits\n (merged to 'next' on 2022-02-23 at 0d5a844b40)\n + tests: make the code more readable\n + tests: allow testing if a path is truly a file or a directory\n + t/t3903-stash.sh: replace test [-d|-f] with test_path_is_*\n\n Test modernization.\n\n Will merge to 'master'.\n source: <20220222215430.605254-1-cogoni.guillaume@gmail.com>\n\n\n* js/bisect-in-c (2022-02-23) 14 commits\n - bisect: no longer try to clean up left-over `.git/head-name` files\n - bisect: remove Cogito-related code\n - bisect: turn `git bisect` into a full built-in\n - bisect: move even the option parsing to `bisect--helper`\n - bisect--helper: return only correct exit codes in `cmd_*()`\n - bisect--helper: move the `BISECT_STATE` case to the end\n - bisect--helper: make `--bisect-state` optional\n - bisect--helper: align the sub-command order with git-bisect.sh\n - bisect--helper: using `--bisect-state` without an argument is a bug\n - bisect--helper: really retire `--bisect-autostart`\n - bisect--helper: really retire --bisect-next-check\n - bisect--helper: retire the --no-log option\n - bisect: avoid double-quoting when printing the failed command\n - bisect run: fix the error message\n\n Final bits of \"git bisect.sh\" have been rewritten in C.\n\n Will merge to 'next'?\n source: \n\n\n* js/scalar-diagnose (2022-02-06) 6 commits\n - scalar: teach `diagnose` to gather loose objects information\n - scalar: teach `diagnose` to gather packfile info\n - scalar diagnose: include disk space information\n - scalar: add `diagnose`\n - scalar: validate the optional enlistment argument\n - archive: optionally add \"virtual\" files\n\n Implementation of \"scalar diagnose\" subcommand.\n\n On hold.\n cf. \n source: \n\n\n* jc/cat-file-batch-commands (2022-02-18) 4 commits\n (merged to 'next' on 2022-03-01 at 8764b7963c)\n + cat-file: add --batch-command mode\n + cat-file: add remove_timestamp helper\n + cat-file: introduce batch_mode enum to replace print_contents\n + cat-file: rename cmdmode to transform_mode\n\n \"git cat-file\" learns \"--batch-command\" mode, which is a more\n flexible interface than the existing \"--batch\" or \"--batch-check\"\n modes, to allow different kinds of inquiries made.\n\n Will merge to 'master'.\n source: \n\n\n* ar/submodule-update (2022-03-02) 13 commits\n - submodule--helper update-clone: check for --filter and --init\n - submodule update: add tests for --filter\n - submodule--helper update-clone: learn --init\n - submodule--helper: allow setting superprefix for init_submodule()\n - submodule--helper: refactor get_submodule_displaypath()\n - submodule--helper run-update-procedure: learn --remote\n - submodule--helper: don't use bitfield indirection for parse_options()\n - submodule--helper: get remote names from any repository\n - submodule--helper: remove ensure-core-worktree\n - submodule--helper run-update-procedure: remove --suboid\n - submodule--helper: reorganize code for sh to C conversion\n - submodule--helper: remove update-module-mode\n - submodule tests: test for init and update failure output\n\n Rewrite of \"git submodule update\" in C (early part).\n\n cf. \n source: <20220303005727.69270-1-chooglen@google.com>\n\n\n* rj/receive-pack-abort-upon-disconnect (2022-01-28) 1 commit\n - receive-pack: check if client is alive before completing the push\n\n \"git push\" may be killed by the user when the server side has\n finished receiving all data and is about to commit the result.\n Give the latter a better chance to notice such situation and abort\n processing the ref updates.\n\n Needs more work?\n cf. <220204.864k5e4yvf.gmgdl@evledraar.gmail.com>\n source: <20220128194811.3396281-1-robin.jarry@6wind.com>\n\n\n* ab/object-file-api-updates (2022-02-25) 12 commits\n - object-file API: pass an enum to read_object_with_reference()\n - object-file.c: add a literal version of write_object_file_prepare()\n - object-file API: have hash_object_file() take \"enum object_type\"\n - object API: rename hash_object_file_literally() to write_*()\n - object-file API: split up and simplify check_object_signature()\n - object API users + docs: check <0, not !0 with check_object_signature()\n - object API docs: move check_object_signature() docs to cache.h\n - object API: correct \"buf\" v.s. \"map\" mismatch in *.c and *.h\n - object-file API: have write_object_file() take \"enum object_type\"\n - object-file API: add a format_object_header() function\n - object-file API: return \"void\", not \"int\" from hash_object_file()\n - object-file.c: split up declaration of unrelated variables\n\n Object-file API shuffling.\n\n Will merge to 'next'.\n source: \n\n\n* en/merge-tree (2022-02-23) 13 commits\n - git-merge-tree.txt: add a section on potentional usage mistakes\n - merge-tree: add a --allow-unrelated-histories flag\n - merge-tree: allow `ls-files -u` style info to be NUL terminated\n - merge-tree: provide easy access to `ls-files -u` style info\n - merge-tree: provide a list of which files have conflicts\n - merge-ort: provide a merge_get_conflicted_files() helper function\n - merge-tree: support including merge messages in output\n - merge-ort: split out a separate display_update_messages() function\n - merge-tree: implement real merges\n - merge-tree: add option parsing and initial shell for real merge function\n - merge-tree: move logic for existing merge into new function\n - merge-tree: rename merge_trees() to trivial_merge_trees()\n - Merge branch 'en/remerge-diff' into en/merge-trees\n\n A new command is introduced that takes two commits and computes a\n tree that would be contained in the resulting merge commit, if the\n histories leading to these two commits were to be merged, and is\n added as a new mode of \"git merge-tree\" subcommand.\n\n On hold.\n cf. \n source: \n\n\n* jh/p4-various-fixups (2022-02-10) 22 commits\n - git-p4: sort imports\n - git-p4: seperate multiple statements onto seperate lines\n - git-p4: move inline comments to line above\n - git-p4: only seperate code blocks by a single empty line\n - git-p4: compare to singletons with \"is\" and \"is not\"\n - git-p4: normalize indentation of lines in conditionals\n - git-p4: ensure there is a single space around all operators\n - git-p4: ensure every comment has a single #\n - git-p4: remove spaces between dictionary keys and colons\n - git-p4: remove redundant backslash-continuations inside brackets\n - git-p4: remove extraneous spaces before function arguments\n - git-p4: place a single space after every comma\n - git-p4: removed brackets when assigning multiple return values\n - git-p4: remove spaces around default arguments\n - git-p4: remove padding from lists, tuples and function arguments\n - git-p4: sort and de-duplcate pylint disable list\n - git-p4: remove commented code\n - git-p4: convert descriptive class and function comments into docstrings\n - git-p4: improve consistency of docstring formatting\n - git-p4: indent with 4-spaces\n - git-p4: remove unneeded semicolons from statements\n - git-p4: add blank lines between functions and class definitions\n\n Various cleanups to \"git p4\".\n source: <20220210164627.279520-1-jholdsworth@nvidia.com>\n\n\n* rs/bisect-executable-not-found (2022-01-19) 4 commits\n (merged to 'next' on 2022-02-28 at c74bff7954)\n + bisect--helper: double-check run command on exit code 126 and 127\n + bisect: document run behavior with exit codes 126 and 127\n + bisect--helper: release strbuf and strvec on run error\n + bisect--helper: report actual bisect_state() argument on error\n\n A not-so-common mistake is to write a script to feed \"git bisect\n run\" without making it executable, in which case all tests will\n exit with 126 or 127 error codes, even on revisions that are marked\n as good. Try to recognize this situation and stop iteration early.\n\n Will merge to 'master'.\n source: \n\n\n* en/present-despite-skipped (2022-03-01) 7 commits\n (merged to 'next' on 2022-03-01 at 4982287a31)\n + repo_read_index: add config to expect files outside sparse patterns\n (merged to 'next' on 2022-02-15 at 960873fdad)\n + Accelerate clear_skip_worktree_from_present_files() by caching\n + Update documentation related to sparsity and the skip-worktree bit\n + repo_read_index: clear SKIP_WORKTREE bit from files present in worktree\n + unpack-trees: fix accidental loss of user changes\n + t1011: add testcase demonstrating accidental loss of user modifications\n + Merge branch 'vd/sparse-clean-etc' into en/present-despite-skipped\n\n In sparse-checkouts, files mis-marked as missing from the working tree\n could lead to later problems. Such files were hard to discover, and\n harder to correct. Automatically detecting and correcting the marking\n of such files has been added to avoid these problems.\n\n Will merge to 'master'.\n source: \n\n\n* js/use-builtin-add-i (2021-12-01) 2 commits\n - add -i: default to the built-in implementation\n - t2016: require the PERL prereq only when necessary\n\n \"git add -i\" was rewritten in C some time ago and has been in\n testing; the reimplementation is now exposed to general public by\n default.\n\n On hold.\n\n What's the status of the \"known breakage\"?\n Are we ready to switch if we wanted to?\n There are known breakages on macOS.\n cf. \n source: \n\n\n* jh/builtin-fsmonitor-part2 (2022-03-01) 30 commits\n - update-index: convert fsmonitor warnings to advise\n - t7527: test status with untracked-cache and fsmonitor--daemon\n - fsmonitor: force update index after large responses\n - fsmonitor--daemon: use a cookie file to sync with file system\n - fsmonitor--daemon: periodically truncate list of modified files\n - t/perf/p7519: add fsmonitor--daemon test cases\n - t/perf/p7519: speed up test on Windows\n - t/helper/test-chmtime: skip directories on Windows\n - t/perf: avoid copying builtin fsmonitor files into test repo\n - t7527: create test for fsmonitor--daemon\n - t/helper/fsmonitor-client: create IPC client to talk to FSMonitor Daemon\n - help: include fsmonitor--daemon feature flag in version info\n - fsmonitor--daemon: implement handle_client callback\n - compat/fsmonitor/fsm-listen-darwin: implement FSEvent listener on MacOS\n - compat/fsmonitor/fsm-listen-darwin: add MacOS header files for FSEvent\n - compat/fsmonitor/fsm-listen-win32: implement FSMonitor backend on Windows\n - fsmonitor--daemon: create token-based changed path cache\n - fsmonitor--daemon: define token-ids\n - fsmonitor--daemon: add pathname classification\n - fsmonitor--daemon: implement 'start' command\n - fsmonitor--daemon: implement 'run' command\n - compat/fsmonitor/fsm-listen-darwin: stub in backend for Darwin\n - compat/fsmonitor/fsm-listen-win32: stub in backend for Windows\n - fsmonitor--daemon: implement 'stop' and 'status' commands\n - fsmonitor--daemon: add a built-in fsmonitor daemon\n - fsmonitor: document builtin fsmonitor\n - fsmonitor: use IPC to query the builtin FSMonitor daemon\n - fsmonitor: config settings are repository-specific\n - fsmonitor-ipc: create client routines for git-fsmonitor--daemon\n - fsmonitor: enhance existing comments, clarify trivial response handling\n (this branch is used by jh/builtin-fsmonitor-part3.)\n\n Built-in fsmonitor (part 2).\n\n Will merge to 'next'.\n source: \n\n\n* es/superproject-aware-submodules (2022-02-28) 3 commits\n - rev-parse: short-circuit superproject worktree when config unset\n - introduce submodule.hasSuperproject record\n - t7400-submodule-basic: modernize inspect() helper\n\n A configuration variable in a repository tells if it is (or is not)\n a submodule of a superproject.\n\n Expecting a reroll.\n cf. \n source: <20220301002613.1459916-1-emilyshaffer@google.com>\n\n------------------------------------------------------------\n[Discarded]\n\n* jc/parse-options-check (2022-03-01) 1 commit\n . parse-options: make parse_options_check() test-only\n\n Bypass sanity checks on the options array at runtime, and trigger\n them only during the tests.\n source: \n\n\n* jc/merge-continue-doc (2022-02-28) 1 commit\n . merge: 'git merge --continue' is merely 'git commit'\n\n \"git merge\" documentation clarifies that \"git commit\" is sufficient\n to conclude an interrupted merge.\n source: \n"},{"id":"450394","messageId":"5f8740bd-c0b7-a8e1-4d73-b79b40cc02e9@github.com","threadId":"57506","inReplyTo":"xmqqv8wu2vag.fsf@gitster.g","subject":"ab/plug-random-leaks (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Derrick Stolee","fromEmail":"derrickstolee@github.com","sentAt":"2022-03-04T13:25:56Z","receivedAt":"2022-03-04T13:26:01Z","isPatch":false,"sender":{"key":"stolee@gmail.com","avatar":"https://avatars.githubusercontent.com/u/570044?v=4"},"body":"On 3/3/2022 11:31 PM, Junio C Hamano wrote:\n\n> * ab/plug-random-leaks (2022-03-02) 14 commits\n> - repository.c: free the \"path cache\" in repo_clear()\n> - range-diff: plug memory leak in read_patches()\n> - range-diff: plug memory leak in common invocation\n> - lockfile API users: simplify and don't leak \"path\"\n> - commit-graph: stop fill_oids_from_packs() progress on error and free()\n> - commit-graph: fix memory leak in misused string_list API\n> - submodule--helper: fix trivial leak in module_add()\n> - transport: stop needlessly copying bundle header references\n> - bundle: call strvec_clear() on allocated strvec\n> - remote-curl.c: free memory in cmd_main()\n> - urlmatch.c: add and use a *_release() function\n> - diff.c: free \"buf\" in diff_words_flush()\n> - merge-base: free() allocated \"struct commit **\" list\n> - index-pack: fix memory leaks\n> \n> Plug random memory leaks.\n> \n> Will merge to 'next'.\n> source: \n\nThis series needs a re-roll before it gets merged. There is\na correctness issue in this patch:\n\n> - transport: stop needlessly copying bundle header references\n\nThanks,\n-Stolee\n"},{"id":"450411","messageId":"0870b8f0-976a-cf2f-f34f-7e966b9c426f@github.com","threadId":"57506","inReplyTo":"xmqqv8wu2vag.fsf@gitster.g","subject":"tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Derrick Stolee","fromEmail":"derrickstolee@github.com","sentAt":"2022-03-04T15:35:34Z","receivedAt":"2022-03-04T15:35:39Z","isPatch":false,"sender":{"key":"stolee@gmail.com","avatar":"https://avatars.githubusercontent.com/u/570044?v=4"},"body":"On 3/3/2022 11:31 PM, Junio C Hamano wrote:\n> * tb/cruft-packs (2022-03-02) 17 commits\n> - sha1-file.c: don't freshen cruft packs\n> - builtin/gc.c: conditionally avoid pruning objects via loose\n> - builtin/repack.c: add cruft packs to MIDX during geometric repack\n> - builtin/repack.c: use named flags for existing_packs\n> - builtin/repack.c: allow configuring cruft pack generation\n> - builtin/repack.c: support generating a cruft pack\n> - builtin/pack-objects.c: --cruft with expiration\n> - reachable: report precise timestamps from objects in cruft packs\n> - reachable: add options to add_unseen_recent_objects_to_traversal\n> - builtin/pack-objects.c: --cruft without expiration\n> - builtin/pack-objects.c: return from create_object_entry()\n> - t/helper: add 'pack-mtimes' test-tool\n> - pack-mtimes: support writing pack .mtimes files\n> - chunk-format.h: extract oid_version()\n> - pack-write: pass 'struct packing_data' to 'stage_tmp_packfiles'\n> - pack-mtimes: support reading .mtimes files\n> - Documentation/technical: add cruft-packs.txt\n> \n> A mechanism to pack unreachable objects into a \"cruft pack\",\n> instead of ejecting them into loose form to be reclaimed later, has\n> been introduced.\n> \n> Will merge to 'next'?\n> source: \n\nThis version looks good to me. There were some other comments that\nseem to have resolved without a need to change the patches.\n\nThanks,\n-Stolee\n\n"},{"id":"450429","messageId":"220304.861qzhftzc.gmgdl@evledraar.gmail.com","threadId":"57506","inReplyTo":"5f8740bd-c0b7-a8e1-4d73-b79b40cc02e9@github.com","subject":"Re: ab/plug-random-leaks (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2022-03-04T18:33:05Z","receivedAt":"2022-03-04T18:33:50Z","isPatch":false,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"\nOn Fri, Mar 04 2022, Derrick Stolee wrote:\n\n> On 3/3/2022 11:31 PM, Junio C Hamano wrote:\n>\n>> * ab/plug-random-leaks (2022-03-02) 14 commits\n>> - repository.c: free the \"path cache\" in repo_clear()\n>> - range-diff: plug memory leak in read_patches()\n>> - range-diff: plug memory leak in common invocation\n>> - lockfile API users: simplify and don't leak \"path\"\n>> - commit-graph: stop fill_oids_from_packs() progress on error and free()\n>> - commit-graph: fix memory leak in misused string_list API\n>> - submodule--helper: fix trivial leak in module_add()\n>> - transport: stop needlessly copying bundle header references\n>> - bundle: call strvec_clear() on allocated strvec\n>> - remote-curl.c: free memory in cmd_main()\n>> - urlmatch.c: add and use a *_release() function\n>> - diff.c: free \"buf\" in diff_words_flush()\n>> - merge-base: free() allocated \"struct commit **\" list\n>> - index-pack: fix memory leaks\n>> \n>> Plug random memory leaks.\n>> \n>> Will merge to 'next'.\n>> source: \n>\n> This series needs a re-roll before it gets merged. There is\n> a correctness issue in this patch:\n>\n>> - transport: stop needlessly copying bundle header references\n\nThanks. I just sent in that re-roll at:\nhttps://lore.kernel.org/git/cover-v2-00.14-00000000000-20220304T182902Z-avarab@gmail.com/\n"},{"id":"450495","messageId":"220305.86y21o327b.gmgdl@evledraar.gmail.com","threadId":"57506","inReplyTo":"xmqqv8wu2vag.fsf@gitster.g","subject":"jc/stash-drop (was: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2022-03-05T14:25:19Z","receivedAt":"2022-03-05T14:26:38Z","isPatch":false,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"\nOn Thu, Mar 03 2022, Junio C Hamano wrote:\n\n> * jc/stash-drop (2022-03-02) 3 commits\n> - stash: call reflog_delete() in reflog.c\n> - reflog: libify delete reflog function and helpers\n> - stash: add tests to ensure reflog --rewrite --updatref behavior\n>\n> \"git stash drop\" is reimplemented as an internal call to\n> reflog_delete() function, instead of invoking \"git reflog delete\"\n> via run_command() API.\n>\n> Will merge to 'next'?\n> source: \n\nI've been reviewing this closely & it looks good to me, all feedback has\nbeen addressed, and there's nothing outstanding (there's one sub-thread\non v4 answering a question from Phillip Wood asking if the new test\nneeds REFFILES, it does).\n"},{"id":"450573","messageId":"3683c549-6efc-2c67-1130-d7008100ee55@github.com","threadId":"57506","inReplyTo":"xmqqv8wu2vag.fsf@gitster.g","subject":"ds/commit-graph-gen-v2-fixes (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Derrick Stolee","fromEmail":"derrickstolee@github.com","sentAt":"2022-03-07T13:49:43Z","receivedAt":"2022-03-07T13:49:50Z","isPatch":false,"sender":{"key":"stolee@gmail.com","avatar":"https://avatars.githubusercontent.com/u/570044?v=4"},"body":"On 3/3/2022 11:31 PM, Junio C Hamano wrote:\n\n> * ds/commit-graph-gen-v2-fixes (2022-03-01) 5 commits\n> - commit-graph: fix generation number v2 overflow values\n> - commit-graph: start parsing generation v2 (again)\n> - commit-graph: fix ordering bug in generation numbers\n> - t5318: extract helpers to lib-commit-graph.sh\n> - test-read-graph: include extra post-parse info\n> \n> Fixes to the way generation number v2 in the commit-graph files are\n> (not) handled.\n> source: \n\nAfter discussing the issues in detail with Patrick, it seems\nthat an additional patch is required (submitted in [1], though\nI incorrectly numbered it \"5/4\" when it should be \"6/5\"). I can\nalso re-roll to make sure that the patch is visible inline with\nthe rest of the series, if you want.\n\nThanks,\n-Stolee\n\n[1] https://lore.kernel.org/git/db5eb248-1b54-9f24-8f8a-28c19a1eee6a@github.com/\n"},{"id":"450620","messageId":"xmqqee3du1eg.fsf@gitster.g","threadId":"57506","inReplyTo":"3683c549-6efc-2c67-1130-d7008100ee55@github.com","subject":"Re: ds/commit-graph-gen-v2-fixes (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-03-07T17:18:31Z","receivedAt":"2022-03-07T17:18:39Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Derrick Stolee writes:\n\n> On 3/3/2022 11:31 PM, Junio C Hamano wrote:\n>\n>> * ds/commit-graph-gen-v2-fixes (2022-03-01) 5 commits\n>> - commit-graph: fix generation number v2 overflow values\n>> - commit-graph: start parsing generation v2 (again)\n>> - commit-graph: fix ordering bug in generation numbers\n>> - t5318: extract helpers to lib-commit-graph.sh\n>> - test-read-graph: include extra post-parse info\n>> \n>> Fixes to the way generation number v2 in the commit-graph files are\n>> (not) handled.\n>> source: \n>\n> After discussing the issues in detail with Patrick, it seems\n> that an additional patch is required (submitted in [1], though\n> I incorrectly numbered it \"5/4\" when it should be \"6/5\"). I can\n> also re-roll to make sure that the patch is visible inline with\n> the rest of the series, if you want.\n>\n> Thanks,\n> -Stolee\n>\n> [1] https://lore.kernel.org/git/db5eb248-1b54-9f24-8f8a-28c19a1eee6a@github.com/\n\nI do recall seeing the follow-up patch. Let me queue it on top.\n\nThanks.\n\n\n"},{"id":"450627","messageId":"YiZJiPVMZwPXbfrK@google.com","threadId":"57506","inReplyTo":"0870b8f0-976a-cf2f-f34f-7e966b9c426f@github.com","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2022-03-07T18:06:00Z","receivedAt":"2022-03-07T18:06:06Z","isPatch":false,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Hi,\n\nDerrick Stolee wrote:\n> On 3/3/2022 11:31 PM, Junio C Hamano wrote:\n\n>> * tb/cruft-packs (2022-03-02) 17 commits\n>> - sha1-file.c: don't freshen cruft packs\n>> - builtin/gc.c: conditionally avoid pruning objects via loose\n>> - builtin/repack.c: add cruft packs to MIDX during geometric repack\n>> - builtin/repack.c: use named flags for existing_packs\n>> - builtin/repack.c: allow configuring cruft pack generation\n>> - builtin/repack.c: support generating a cruft pack\n>> - builtin/pack-objects.c: --cruft with expiration\n>> - reachable: report precise timestamps from objects in cruft packs\n>> - reachable: add options to add_unseen_recent_objects_to_traversal\n>> - builtin/pack-objects.c: --cruft without expiration\n>> - builtin/pack-objects.c: return from create_object_entry()\n>> - t/helper: add 'pack-mtimes' test-tool\n>> - pack-mtimes: support writing pack .mtimes files\n>> - chunk-format.h: extract oid_version()\n>> - pack-write: pass 'struct packing_data' to 'stage_tmp_packfiles'\n>> - pack-mtimes: support reading .mtimes files\n>> - Documentation/technical: add cruft-packs.txt\n>>\n>> A mechanism to pack unreachable objects into a \"cruft pack\",\n>> instead of ejecting them into loose form to be reclaimed later, has\n>> been introduced.\n>>\n>> Will merge to 'next'?\n>> source: \n>\n> This version looks good to me. There were some other comments that\n> seem to have resolved without a need to change the patches.\n\nI'm excited about this work! I just sent a quick review to the\ndesign doc.\n\nBefore merging to 'next', I'd be interested in two things:\n\n 1. Marking the feature as experimental so we can learn from experience.\n Clarifying what aspects we consider to be stable / set in stone and\n what are subject to modification.\n\n 2. Marking this as a repository format extension so it doesn't interact\n poorly with Git implementations (including older versions of Git\n itself) that are not aware of the new feature\n\nThoughts?\n\nThanks,\nJonathan\n"},{"id":"450630","messageId":"YiZMhuI/DdpvQ/ED@nand.local","threadId":"57506","inReplyTo":"YiZJiPVMZwPXbfrK@google.com","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Taylor Blau","fromEmail":"me@ttaylorr.com","sentAt":"2022-03-07T18:18:46Z","receivedAt":"2022-03-07T18:18:49Z","isPatch":false,"sender":{"key":"me@ttaylorr.com","avatar":"https://avatars.githubusercontent.com/u/301000140?v=4"},"body":"On Mon, Mar 07, 2022 at 10:06:00AM -0800, Jonathan Nieder wrote:\n> I'm excited about this work! I just sent a quick review to the\n> design doc.\n\nThanks! I haven't had a chance to look at the design doc, but let me\nrespond quickly to this message:\n\n> Before merging to 'next', I'd be interested in two things:\n>\n> 1. Marking the feature as experimental so we can learn from experience.\n> Clarifying what aspects we consider to be stable / set in stone and\n> what are subject to modification.\n\nI'm not sure there is much practical benefit to marking this feature as\nexperimental. The only new file format here is the .mtimes one, which\nshould make it easy for us to modify the format in a\nbackwards-compatible way.\n\nIf there are other benefits you had in mind, I'm curious to hear them.\nBut I think we should be fine to \"lock in\" the first version of the\n.mtimes format since we have an easy-ish mechanism to change it in the\nfuture.\n\n> 2. Marking this as a repository format extension so it doesn't interact\n> poorly with Git implementations (including older versions of Git\n> itself) that are not aware of the new feature\n\nThe design of cruft packs was done intentionally to avoid needing a\nformat extension. The cruft pack is \"just a pack\" to any older version\nof Git. The only thing an older version of Git wouldn't understand is\nhow to interpret the .mtimes file. But that's no different than the\ncurrent behavior without cruft packs, where any unreachable object\ninherits the mtime of its containing pack.\n\nSo an older version of Git might prune a different set of objects than a\nversion that understands cruft packs depending on the contents of the\n.mtimes file, the mtime of the cruft pack, and the width of the grace\nperiod. But I think by downgrading you are more or less buying into the\nexisting behavior. So I don't think there is a compelling reason to\nintroduce a format extension here.\n\nThanks,\nTaylor\n"},{"id":"450631","messageId":"xmqqh789sjuq.fsf@gitster.g","threadId":"57506","inReplyTo":"220305.86y21o327b.gmgdl@evledraar.gmail.com","subject":"Re: jc/stash-drop","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-03-07T18:22:53Z","receivedAt":"2022-03-07T18:22:59Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Ævar Arnfjörð Bjarmason writes:\n\n> On Thu, Mar 03 2022, Junio C Hamano wrote:\n>\n>> * jc/stash-drop (2022-03-02) 3 commits\n>> - stash: call reflog_delete() in reflog.c\n>> - reflog: libify delete reflog function and helpers\n>> - stash: add tests to ensure reflog --rewrite --updatref behavior\n>>\n>> \"git stash drop\" is reimplemented as an internal call to\n>> reflog_delete() function, instead of invoking \"git reflog delete\"\n>> via run_command() API.\n>>\n>> Will merge to 'next'?\n>> source: \n>\n> I've been reviewing this closely & it looks good to me, all feedback has\n> been addressed, and there's nothing outstanding (there's one sub-thread\n> on v4 answering a question from Phillip Wood asking if the new test\n> needs REFFILES, it does).\n\nI think this one is good, and I agree that we should leave the \"old\nside is not even visible to regular users and tests need to peek\ninto the implementation detail of the log files---do we need the\n--rewrite option at all, instead of doing so always?\" outside the\ntopic.\n\nThanks.\n"},{"id":"450632","messageId":"ebfac323-7567-6327-f5e6-0fd9e0356550@github.com","threadId":"57506","inReplyTo":"YiZMhuI/DdpvQ/ED@nand.local","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Derrick Stolee","fromEmail":"derrickstolee@github.com","sentAt":"2022-03-07T18:32:19Z","receivedAt":"2022-03-07T18:32:26Z","isPatch":false,"sender":{"key":"stolee@gmail.com","avatar":"https://avatars.githubusercontent.com/u/570044?v=4"},"body":"On 3/7/2022 1:18 PM, Taylor Blau wrote:\n> On Mon, Mar 07, 2022 at 10:06:00AM -0800, Jonathan Nieder wrote:\n>> I'm excited about this work! I just sent a quick review to the\n>> design doc.\n> \n> Thanks! I haven't had a chance to look at the design doc, but let me\n> respond quickly to this message:\n> \n>> Before merging to 'next', I'd be interested in two things:\n>>\n>> 1. Marking the feature as experimental so we can learn from experience.\n>> Clarifying what aspects we consider to be stable / set in stone and\n>> what are subject to modification.\n> \n> I'm not sure there is much practical benefit to marking this feature as\n> experimental. The only new file format here is the .mtimes one, which\n> should make it easy for us to modify the format in a\n> backwards-compatible way.\n> \n> If there are other benefits you had in mind, I'm curious to hear them.\n> But I think we should be fine to \"lock in\" the first version of the\n> .mtimes format since we have an easy-ish mechanism to change it in the\n> future.\n\nI feel similarly to Taylor here.\n\n>> 2. Marking this as a repository format extension so it doesn't interact\n>> poorly with Git implementations (including older versions of Git\n>> itself) that are not aware of the new feature\n> \n> The design of cruft packs was done intentionally to avoid needing a\n> format extension. The cruft pack is \"just a pack\" to any older version\n> of Git. The only thing an older version of Git wouldn't understand is\n> how to interpret the .mtimes file. But that's no different than the\n> current behavior without cruft packs, where any unreachable object\n> inherits the mtime of its containing pack.\n> \n> So an older version of Git might prune a different set of objects than a\n> version that understands cruft packs depending on the contents of the\n> .mtimes file, the mtime of the cruft pack, and the width of the grace\n> period. But I think by downgrading you are more or less buying into the\n> existing behavior. So I don't think there is a compelling reason to\n> introduce a format extension here.\n\nIn particular, older versions would first explode unreachable objects\nout of the cruft pack and into loose objects before expiring any of\nthem based on the loose object mtime. There is no risk here of causing\nproblems with older versions of Git and does not need an extension.\n\nThanks,\n-Stolee\n"},{"id":"450640","messageId":"YiZoojeTodMr+Ypw@google.com","threadId":"57506","inReplyTo":"ebfac323-7567-6327-f5e6-0fd9e0356550@github.com","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2022-03-07T20:18:42Z","receivedAt":"2022-03-07T20:18:49Z","isPatch":false,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Derrick Stolee wrote:\n> On 3/7/2022 1:18 PM, Taylor Blau wrote:\n>> On Mon, Mar 07, 2022 at 10:06:00AM -0800, Jonathan Nieder wrote:\n\n>>> 2. Marking this as a repository format extension so it doesn't interact\n>>> poorly with Git implementations (including older versions of Git\n>>> itself) that are not aware of the new feature\n>>\n>> The design of cruft packs was done intentionally to avoid needing a\n>> format extension. The cruft pack is \"just a pack\" to any older version\n>> of Git. The only thing an older version of Git wouldn't understand is\n>> how to interpret the .mtimes file. But that's no different than the\n>> current behavior without cruft packs, where any unreachable object\n>> inherits the mtime of its containing pack.\n>>\n>> So an older version of Git might prune a different set of objects than a\n>> version that understands cruft packs depending on the contents of the\n>> .mtimes file, the mtime of the cruft pack, and the width of the grace\n>> period. But I think by downgrading you are more or less buying into the\n>> existing behavior. So I don't think there is a compelling reason to\n>> introduce a format extension here.\n>\n> In particular, older versions would first explode unreachable objects\n> out of the cruft pack and into loose objects before expiring any of\n> them based on the loose object mtime. There is no risk here of causing\n> problems with older versions of Git and does not need an extension.\n\nSurely when older and versions are acting on the same repository, they\nwould fight by exploding out unreachable objects, packing them back\ninto a cruft pack, etc, no?\n\nThanks,\nJonathan\n"},{"id":"450647","messageId":"332f9b54-edfb-50e7-46a7-2e46d3b574e4@github.com","threadId":"57506","inReplyTo":"YiZoojeTodMr+Ypw@google.com","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Derrick Stolee","fromEmail":"derrickstolee@github.com","sentAt":"2022-03-07T20:51:40Z","receivedAt":"2022-03-07T20:51:47Z","isPatch":false,"sender":{"key":"stolee@gmail.com","avatar":"https://avatars.githubusercontent.com/u/570044?v=4"},"body":"On 3/7/2022 3:18 PM, Jonathan Nieder wrote:\n> Derrick Stolee wrote:\n>> On 3/7/2022 1:18 PM, Taylor Blau wrote:\n>>> On Mon, Mar 07, 2022 at 10:06:00AM -0800, Jonathan Nieder wrote:\n> \n>>>> 2. Marking this as a repository format extension so it doesn't interact\n>>>> poorly with Git implementations (including older versions of Git\n>>>> itself) that are not aware of the new feature\n>>>\n>>> The design of cruft packs was done intentionally to avoid needing a\n>>> format extension. The cruft pack is \"just a pack\" to any older version\n>>> of Git. The only thing an older version of Git wouldn't understand is\n>>> how to interpret the .mtimes file. But that's no different than the\n>>> current behavior without cruft packs, where any unreachable object\n>>> inherits the mtime of its containing pack.\n>>>\n>>> So an older version of Git might prune a different set of objects than a\n>>> version that understands cruft packs depending on the contents of the\n>>> .mtimes file, the mtime of the cruft pack, and the width of the grace\n>>> period. But I think by downgrading you are more or less buying into the\n>>> existing behavior. So I don't think there is a compelling reason to\n>>> introduce a format extension here.\n>>\n>> In particular, older versions would first explode unreachable objects\n>> out of the cruft pack and into loose objects before expiring any of\n>> them based on the loose object mtime. There is no risk here of causing\n>> problems with older versions of Git and does not need an extension.\n> \n> Surely when older and versions are acting on the same repository, they\n> would fight by exploding out unreachable objects, packing them back\n> into a cruft pack, etc, no?\n\nYou are referring to a situation where there are multiple possible\nversions responsible for maintaining a repository. Git does not\nsupport parallel writers doing significant updates like full\nrepacks and GCs and instead relies on the user to control the\nconcurrency there. The standard we keep to is that parallel readers\ncan still access the repo during this time.\n\nIf someone was running a case where they had these parallel\nmaintenance processes, then they would already be risking failure\nwith existing features (though actually in the case of the old\nversions breaking the new ones): what if the new/old versions\ndiffer in their understanding of the commit-graph? The old one\ncould remove commits but not update the commit-graph, leaving\nextra commits in that file that the new one would fail to verify.\nHow about the multi-pack-index? The new version would try loading\nobjects from missing pack-files since the old version deleted\nthose packs without updating the multi-pack-index.\n\nAt least in the cruft packs the worst case is that no objects are\never expired because they are toggling between loose objects and\ncruft packs.\n\nThanks,\n-Stolee\n"},{"id":"450650","messageId":"xmqqpmmxqwe6.fsf@gitster.g","threadId":"57506","inReplyTo":"332f9b54-edfb-50e7-46a7-2e46d3b574e4@github.com","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-03-07T21:34:57Z","receivedAt":"2022-03-07T21:36:04Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Derrick Stolee writes:\n\n> ... Git does not\n> support parallel writers doing significant updates like full\n> repacks and GCs and instead relies on the user to control the\n> concurrency there.\n\nAt least when we set out to give our users Git, allowing such\nconcurrent writing without corrupting repositories was what we aimed\nto achieve. If you did two simultanenous repacks, one of the may\nfail while trying to acquire a lock or two, so from waste-avoidance\nperspective, there is a strong incentive on the user's side to make\nsure such housecleaning tasks are not triggered needlessly and\nsimultanously, but it shouldn't lead to repository corruption.\n\nHave we broken that over the course of adding new commits and\ndevelopers? If so, we have stuff that needs fixing, and also ensure\nthat the developer and end-user expectations are aligned.\n\nThanks.\n\n\n"},{"id":"450688","messageId":"xmqqv8wpp9ws.fsf@gitster.g","threadId":"57506","inReplyTo":"YiZMhuI/DdpvQ/ED@nand.local","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-03-08T00:25:55Z","receivedAt":"2022-03-08T00:26:00Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Taylor Blau writes:\n\n> On Mon, Mar 07, 2022 at 10:06:00AM -0800, Jonathan Nieder wrote:\n>> I'm excited about this work! I just sent a quick review to the\n>> design doc.\n>\n> Thanks! I haven't had a chance to look at the design doc, but let me\n> respond quickly to this message:\n>\n>> Before merging to 'next', I'd be interested in two things:\n>>\n>> 1. Marking the feature as experimental so we can learn from experience.\n>> Clarifying what aspects we consider to be stable / set in stone and\n>> what are subject to modification.\n>\n> I'm not sure there is much practical benefit to marking this feature as\n> experimental. The only new file format here is the .mtimes one, which\n> should make it easy for us to modify the format in a\n> backwards-compatible way.\n>\n> If there are other benefits you had in mind, I'm curious to hear them.\n> But I think we should be fine to \"lock in\" the first version of the\n> .mtimes format since we have an easy-ish mechanism to change it in the\n> future.\n\nHmph, how? For example, if it turns out that rewriting .mtimes file\nfor each object access turns out to be too much I/O churn and the\napproach to use the mtime of the cruft pack for expiration of the\nentire cruft pack (while ejecting objects that was used from the\ncruft pack out of it to resurrect them from expiration schedule) is\nmore preferrable, how do we back out of from the \"lock in\" once this\nseries is unleashed to the workd?\n"},{"id":"450689","messageId":"YiaoCivdcT3QE24f@nand.local","threadId":"57506","inReplyTo":"xmqqv8wpp9ws.fsf@gitster.g","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Taylor Blau","fromEmail":"me@ttaylorr.com","sentAt":"2022-03-08T00:49:14Z","receivedAt":"2022-03-08T00:49:18Z","isPatch":false,"sender":{"key":"me@ttaylorr.com","avatar":"https://avatars.githubusercontent.com/u/301000140?v=4"},"body":"On Mon, Mar 07, 2022 at 04:25:55PM -0800, Junio C Hamano wrote:\n> > If there are other benefits you had in mind, I'm curious to hear them.\n> > But I think we should be fine to \"lock in\" the first version of the\n> > .mtimes format since we have an easy-ish mechanism to change it in the\n> > future.\n>\n> Hmph, how? For example, if it turns out that rewriting .mtimes file\n> for each object access turns out to be too much I/O churn and the\n> approach to use the mtime of the cruft pack for expiration of the\n> entire cruft pack (while ejecting objects that was used from the\n> cruft pack out of it to resurrect them from expiration schedule) is\n> more preferrable, how do we back out of from the \"lock in\" once this\n> series is unleashed to the workd?\n\n(Note that this series does not propose rewriting the .mtimes file\nduring each object access, since we only need to update our view of\n\"last modified time\" when pruning or repacking. I think a more complete\nexplanation of why can be found in [1] and [2]).\n\nThat detail aside, if we suddenly decided that cruft packs were a bad\nidea and we should get rid of them, then we would be fine to drop all of\nthe cruft pack code. A future version of Git that didn't understand\ncruft packs would ignore the .mtimes file, and we would go back to\nhandling unreachable objects as we do today (by ejecting them loose when\npruning a too-new object that hasn't fallen out of the grace period).\n\nIn other words, this series is designed intentionally so that older\nversions of Git that don't understand cruft packs will continue to work\nfine even in the presence of cruft packs. If we backed out of cruft\npacks at a later date, it would be no different than using an older\nversion of Git that predates cruft packs.\n\nIn other words, I think Stolee's comparison to a feature like\ncommit-graphs (where older versions of Git that don't yet understand\ncommit-graphs work just fine even in repositories that have\ncommit-graphs written) is applicable to this series, too.\n\nThanks,\nTaylor\n\n[1]: https://lore.kernel.org/git/Yap5INmX2ACfjoda@nand.local/\n[2]: https://lore.kernel.org/git/YaqCZ7BPwuMGmkZY@nand.local/\n"},{"id":"450691","messageId":"Yiao1+Zdlj9zZgCQ@nand.local","threadId":"57506","inReplyTo":"xmqqpmmxqwe6.fsf@gitster.g","subject":"Re: tb/cruft-packs (was Re: What's cooking in git.git (Mar 2022, #01; Thu, 3))","fromName":"Taylor Blau","fromEmail":"me@ttaylorr.com","sentAt":"2022-03-08T00:52:39Z","receivedAt":"2022-03-08T00:52:46Z","isPatch":false,"sender":{"key":"me@ttaylorr.com","avatar":"https://avatars.githubusercontent.com/u/301000140?v=4"},"body":"On Mon, Mar 07, 2022 at 01:34:57PM -0800, Junio C Hamano wrote:\n> Derrick Stolee writes:\n>\n> > ... Git does not\n> > support parallel writers doing significant updates like full\n> > repacks and GCs and instead relies on the user to control the\n> > concurrency there.\n>\n> At least when we set out to give our users Git, allowing such\n> concurrent writing without corrupting repositories was what we aimed\n> to achieve. If you did two simultanenous repacks, one of the may\n> fail while trying to acquire a lock or two, so from waste-avoidance\n> perspective, there is a strong incentive on the user's side to make\n> sure such housecleaning tasks are not triggered needlessly and\n> simultanously, but it shouldn't lead to repository corruption.\n\nIt is not true that `git repack` does not support parallel writers.\n\nIndeed, `repack` doesn't hold any locks on the repository ahead of time,\nbut the concurrent writers situation will at worst leave us in a state\nwhere objects appear twice across multiple packs.\n\nSo yes, users are incentivized to limit multiple repack processes from\nstomping on each other and wasting effort, but multiple writers running\n`git repack` cannot corrupt a repository.\n\nThanks,\nTaylor\n"},{"id":"451557","messageId":"a5e5cdd4658d457ffbd80f7263e352cbf3141a1a.1647520853.git.git@grubix.eu","threadId":"57506","inReplyTo":"220304.861qzhftzc.gmgdl@evledraar.gmail.com","subject":"[PATCH] tests: test show --word-diff --color-moved","fromName":"Michael J Gruber","fromEmail":"git@grubix.eu","sentAt":"2022-03-17T12:46:06Z","receivedAt":"2022-03-17T12:58:41Z","isPatch":true,"sender":{"key":"git@grubix.eu","avatar":"https://avatars.githubusercontent.com/u/233215?v=4"},"body":"a18d66cefb (\"diff.c: free \"buf\" in diff_words_flush()\", 2022-03-04)\nintroduced a breakage to `show --word-diff --color-moved` which gives\n\nfree(): double free detected in tcache 2\nAborted (core dumped)\n\non every incarnation. This was not caught by the test suite because we\ntest `diff --word-diff --color-moved` only so far.\n\nTherefore, add a test for `show`, too.\n\nReverting a18d66cefb makes the test pass, but there might be a better\nfix.\n\nSigned-off-by: Michael J Gruber \n---\n t/t4015-diff-whitespace.sh | 8 +++++++-\n 1 file changed, 7 insertions(+), 1 deletion(-)\n\ndiff --git a/t/t4015-diff-whitespace.sh b/t/t4015-diff-whitespace.sh\nindex 9babf13bc9..ff8a0426ca 100755\n--- a/t/t4015-diff-whitespace.sh\n+++ b/t/t4015-diff-whitespace.sh\n@@ -1622,7 +1622,7 @@ test_expect_success 'cmd option assumes configured colored-moved' '\n \ttest_cmp expected actual\n '\n \n-test_expect_success 'no effect from --color-moved with --word-diff' '\n+test_expect_success 'no effect on diff from --color-moved with --word-diff' '\n \tcat <<-\\EOF >text.txt &&\n \tLorem Ipsum is simply dummy text of the printing and typesetting industry.\n \tEOF\n@@ -1636,6 +1636,12 @@ test_expect_success 'no effect from --color-moved with --word-diff' '\n \ttest_cmp expect actual\n '\n \n+test_expect_failure 'no effect on show from --color-moved with --word-diff' '\n+\tgit show --color-moved --word-diff >actual &&\n+\tgit show --word-diff >expect &&\n+\ttest_cmp expect actual\n+'\n+\n test_expect_success 'set up whitespace tests' '\n \tgit reset --hard &&\n \t# Note that these lines have no leading or trailing whitespace.\n-- \n2.35.1.981.gd72fe90d09\n\n"},{"id":"451560","messageId":"cover-v2-0.2-00000000000-20220317T144838Z-avarab@gmail.com","threadId":"57506","inReplyTo":"a5e5cdd4658d457ffbd80f7263e352cbf3141a1a.1647520853.git.git@grubix.eu","subject":"[PATCH v2 0/2] diff.c: fix a recent memory leak regression","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2022-03-17T14:55:33Z","receivedAt":"2022-03-17T14:55:41Z","isPatch":true,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"There's a really dumb think-o in a commit of mine in the recently\nlanded ccafbbfb4ee (Merge branch 'ab/plug-random-leaks', 2022-03-13),\nsorry about that, and thanks a lot to Michael J Gruber for the report.\n\nThis \"v2\" series is re-roll of his addition of a regression test in\nthe \"v1\"[1].\n\nAs the range-diff shows I took the libertay of adjusting the commit\nmessage a bit, mainly noting the regression, re-wording a bit, and\nreplacing the (presumably glibc?) output with the better\nSANITIZE=address report.\n\nThe 2/2 then fixes the issue, and changes the relevant regression\ntests to run under SANITIZE=leak (i.e. the \"linux-leaks\" job).\n\n1. https://lore.kernel.org/git/a5e5cdd4658d457ffbd80f7263e352cbf3141a1a.1647520853.git.git@grubix.eu/\n\nMichael J Gruber (1):\n tests: demonstrate \"show --word-diff --color-moved\" regression\n\nÆvar Arnfjörð Bjarmason (1):\n diff.c: fix a double-free regression in a18d66cefb\n\n diff.c | 11 +++++++++--\n t/t4015-diff-whitespace.sh | 12 ++++++++++--\n 2 files changed, 19 insertions(+), 4 deletions(-)\n\nRange-diff against v1:\n1: f6e7318b418 ! 1: 7f6a6450259 tests: test show --word-diff --color-moved\n @@ Metadata\n Author: Michael J Gruber \n \n ## Commit message ##\n - tests: test show --word-diff --color-moved\n + tests: demonstrate \"show --word-diff --color-moved\" regression\n \n - a18d66cefb (\"diff.c: free \"buf\" in diff_words_flush()\", 2022-03-04)\n - introduced a breakage to `show --word-diff --color-moved` which gives\n + Add a failing test which demonstrates a regression in\n + a18d66cefb (\"diff.c: free \"buf\" in diff_words_flush()\", 2022-03-04),\n + the regression is discussed in detail in the subsequent commit. With\n + it running `git show --word-diff --color-moved` with SANITIZE=address\n + would emit:\n \n - free(): double free detected in tcache 2\n - Aborted (core dumped)\n + ==31191==ERROR: AddressSanitizer: attempting double-free on 0x617000021100 in thread T0:\n + #0 0x49f0a2 in free (git+0x49f0a2)\n + #1 0x9b0e4d in diff_words_flush diff.c:2153:3\n + #2 0x9aed5d in fn_out_consume diff.c:2354:3\n + #3 0xe092ab in consume_one xdiff-interface.c:43:9\n + #4 0xe072eb in xdiff_outf xdiff-interface.c:76:10\n + #5 0xec7014 in xdl_emit_diffrec xdiff/xutils.c:53:6\n + [...]\n \n - on every incarnation. This was not caught by the test suite because we\n - test `diff --word-diff --color-moved` only so far.\n + 0x617000021100 is located 0 bytes inside of 768-byte region [0x617000021100,0x617000021400)\n + freed by thread T0 here:\n + #0 0x49f0a2 in free (git+0x49f0a2)\n + [...(same stacktrace)...]\n \n - Therefore, add a test for `show`, too.\n + previously allocated by thread T0 here:\n + #0 0x49f603 in __interceptor_realloc (git+0x49f603)\n + #1 0xde4da4 in xrealloc wrapper.c:126:8\n + #2 0x995dc5 in append_emitted_diff_symbol diff.c:794:2\n + #3 0x96c44a in emit_diff_symbol diff.c:1527:3\n + [...]\n +\n + This was not caught by the test suite because we test `diff\n + --word-diff --color-moved` only so far.\n \n - Reverting a18d66cefb makes the test pass, but there might be a better\n - fix.\n + Therefore, add a test for `show`, too.\n \n Signed-off-by: Michael J Gruber \n + Signed-off-by: Ævar Arnfjörð Bjarmason \n \n ## t/t4015-diff-whitespace.sh ##\n @@ t/t4015-diff-whitespace.sh: test_expect_success 'cmd option assumes configured colored-moved' '\n-: ----------- > 2: cae11491599 diff.c: fix a double-free regression in a18d66cefb\n-- \n2.35.1.1384.g7d2906948a1\n\n"},{"id":"451561","messageId":"patch-v2-2.2-cae11491599-20220317T144838Z-avarab@gmail.com","threadId":"57506","inReplyTo":"cover-v2-0.2-00000000000-20220317T144838Z-avarab@gmail.com","subject":"[PATCH v2 2/2] diff.c: fix a double-free regression in a18d66cefb","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2022-03-17T14:55:35Z","receivedAt":"2022-03-17T14:55:45Z","isPatch":true,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"My a18d66cefb9 (diff.c: free \"buf\" in diff_words_flush(), 2022-03-04)\nhas what it retrospect is a rather obvious bug (I don't know what I\nwas thinking, if it all): We use the \"emitted_symbols\" allocation in\nappend_emitted_diff_symbol() N times, but starting with a18d66cefb9\nwe'd free it after its first use!\n\nThe correct way to free this data would have been to add the free() to\nthe existing free_diff_words_data() function, so let's do that. The\n\"ecbdata->diff_words->opt->emitted_symbols\" might be NULL, so let's\nadd a trivial free_emitted_diff_symbols() helper next to the function\nthat appends to it.\n\nThis fixes the \"no effect on show from\" leak tested for in the\npreceding commit. Perhaps confusingly this change will skip that test\nunder SANITIZE=leak, but otherwise opt-in the\n\"t4015-diff-whitespace.sh\" test.\n\nThe reason is that a18d66cefb9 \"fixed\" the leak in the preceding \"no\neffect on diff\" test, but for the first call to diff_words_flush() the\n\"wol->buf\" would be NULL, so we wouldn't double-free (and\nSANITIZE=address would see nothing amiss). With this change we'll\nstill pass that test, showing that we've also fixed leaks on this\ncodepath.\n\nWe then have to skip the new \"no effect on show\" test because it\nhappens to trip over an unrelated memory leak (in revision.c). The\nsame goes for \"move detection with submodules\". Both of them pass with\nSANITIZE=address though, which would error on the \"no effect on show\"\ntest before this change.\n\nReported-by: Michael J Gruber \nSigned-off-by: Ævar Arnfjörð Bjarmason \n---\n diff.c | 11 +++++++++--\n t/t4015-diff-whitespace.sh | 6 ++++--\n 2 files changed, 13 insertions(+), 4 deletions(-)\n\ndiff --git a/diff.c b/diff.c\nindex 6b22946cd0e..ef7159968b6 100644\n--- a/diff.c\n+++ b/diff.c\n@@ -800,6 +800,14 @@ static void append_emitted_diff_symbol(struct diff_options *o,\n \tf->line = e->line ? xmemdupz(e->line, e->len) : NULL;\n }\n \n+static void free_emitted_diff_symbols(struct emitted_diff_symbols *e)\n+{\n+\tif (!e)\n+\t\treturn;\n+\tfree(e->buf);\n+\tfree(e);\n+}\n+\n struct moved_entry {\n \tconst struct emitted_diff_symbol *es;\n \tstruct moved_entry *next_line;\n@@ -2150,7 +2158,6 @@ static void diff_words_flush(struct emit_callback *ecbdata)\n \n \t\tfor (i = 0; i < wol->nr; i++)\n \t\t\tfree((void *)wol->buf[i].line);\n-\t\tfree(wol->buf);\n \n \t\twol->nr = 0;\n \t}\n@@ -2228,7 +2235,7 @@ static void free_diff_words_data(struct emit_callback *ecbdata)\n {\n \tif (ecbdata->diff_words) {\n \t\tdiff_words_flush(ecbdata);\n-\t\tfree (ecbdata->diff_words->opt->emitted_symbols);\n+\t\tfree_emitted_diff_symbols(ecbdata->diff_words->opt->emitted_symbols);\n \t\tfree (ecbdata->diff_words->opt);\n \t\tfree (ecbdata->diff_words->minus.text.ptr);\n \t\tfree (ecbdata->diff_words->minus.orig);\ndiff --git a/t/t4015-diff-whitespace.sh b/t/t4015-diff-whitespace.sh\nindex ff8a0426ca5..f3e20dd5bba 100755\n--- a/t/t4015-diff-whitespace.sh\n+++ b/t/t4015-diff-whitespace.sh\n@@ -6,6 +6,8 @@\n test_description='Test special whitespace in diff engine.\n \n '\n+\n+TEST_PASSES_SANITIZE_LEAK=true\n . ./test-lib.sh\n . \"$TEST_DIRECTORY\"/lib-diff.sh\n \n@@ -1636,7 +1638,7 @@ test_expect_success 'no effect on diff from --color-moved with --word-diff' '\n \ttest_cmp expect actual\n '\n \n-test_expect_failure 'no effect on show from --color-moved with --word-diff' '\n+test_expect_success !SANITIZE_LEAK 'no effect on show from --color-moved with --word-diff' '\n \tgit show --color-moved --word-diff >actual &&\n \tgit show --word-diff >expect &&\n \ttest_cmp expect actual\n@@ -2022,7 +2024,7 @@ test_expect_success '--color-moved rewinds for MIN_ALNUM_COUNT' '\n \ttest_cmp expected actual\n '\n \n-test_expect_success 'move detection with submodules' '\n+test_expect_success !SANITIZE_LEAK 'move detection with submodules' '\n \ttest_create_repo bananas &&\n \techo ripe >bananas/recipe &&\n \tgit -C bananas add recipe &&\n-- \n2.35.1.1384.g7d2906948a1\n\n"},{"id":"451562","messageId":"patch-v2-1.2-7f6a6450259-20220317T144838Z-avarab@gmail.com","threadId":"57506","inReplyTo":"cover-v2-0.2-00000000000-20220317T144838Z-avarab@gmail.com","subject":"[PATCH v2 1/2] tests: demonstrate \"show --word-diff --color-moved\" regression","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2022-03-17T14:55:34Z","receivedAt":"2022-03-17T14:55:47Z","isPatch":true,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"From: Michael J Gruber \n\nAdd a failing test which demonstrates a regression in\na18d66cefb (\"diff.c: free \"buf\" in diff_words_flush()\", 2022-03-04),\nthe regression is discussed in detail in the subsequent commit. With\nit running `git show --word-diff --color-moved` with SANITIZE=address\nwould emit:\n\n\t==31191==ERROR: AddressSanitizer: attempting double-free on 0x617000021100 in thread T0:\n\t #0 0x49f0a2 in free (git+0x49f0a2)\n\t #1 0x9b0e4d in diff_words_flush diff.c:2153:3\n\t #2 0x9aed5d in fn_out_consume diff.c:2354:3\n\t #3 0xe092ab in consume_one xdiff-interface.c:43:9\n\t #4 0xe072eb in xdiff_outf xdiff-interface.c:76:10\n\t #5 0xec7014 in xdl_emit_diffrec xdiff/xutils.c:53:6\n\t [...]\n\n\t0x617000021100 is located 0 bytes inside of 768-byte region [0x617000021100,0x617000021400)\n\tfreed by thread T0 here:\n\t #0 0x49f0a2 in free (git+0x49f0a2)\n\t [...(same stacktrace)...]\n\n\tpreviously allocated by thread T0 here:\n\t #0 0x49f603 in __interceptor_realloc (git+0x49f603)\n\t #1 0xde4da4 in xrealloc wrapper.c:126:8\n\t #2 0x995dc5 in append_emitted_diff_symbol diff.c:794:2\n\t #3 0x96c44a in emit_diff_symbol diff.c:1527:3\n\t [...]\n\nThis was not caught by the test suite because we test `diff\n--word-diff --color-moved` only so far.\n\nTherefore, add a test for `show`, too.\n\nSigned-off-by: Michael J Gruber \nSigned-off-by: Ævar Arnfjörð Bjarmason \n---\n t/t4015-diff-whitespace.sh | 8 +++++++-\n 1 file changed, 7 insertions(+), 1 deletion(-)\n\ndiff --git a/t/t4015-diff-whitespace.sh b/t/t4015-diff-whitespace.sh\nindex 9babf13bc9b..ff8a0426ca5 100755\n--- a/t/t4015-diff-whitespace.sh\n+++ b/t/t4015-diff-whitespace.sh\n@@ -1622,7 +1622,7 @@ test_expect_success 'cmd option assumes configured colored-moved' '\n \ttest_cmp expected actual\n '\n \n-test_expect_success 'no effect from --color-moved with --word-diff' '\n+test_expect_success 'no effect on diff from --color-moved with --word-diff' '\n \tcat <<-\\EOF >text.txt &&\n \tLorem Ipsum is simply dummy text of the printing and typesetting industry.\n \tEOF\n@@ -1636,6 +1636,12 @@ test_expect_success 'no effect from --color-moved with --word-diff' '\n \ttest_cmp expect actual\n '\n \n+test_expect_failure 'no effect on show from --color-moved with --word-diff' '\n+\tgit show --color-moved --word-diff >actual &&\n+\tgit show --word-diff >expect &&\n+\ttest_cmp expect actual\n+'\n+\n test_expect_success 'set up whitespace tests' '\n \tgit reset --hard &&\n \t# Note that these lines have no leading or trailing whitespace.\n-- \n2.35.1.1384.g7d2906948a1\n\n"},{"id":"451564","messageId":"xmqqtubwh8wf.fsf@gitster.g","threadId":"57506","inReplyTo":"patch-v2-1.2-7f6a6450259-20220317T144838Z-avarab@gmail.com","subject":"Re: [PATCH v2 1/2] tests: demonstrate \"show --word-diff --color-moved\" regression","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-03-17T15:54:24Z","receivedAt":"2022-03-17T15:54:34Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Ævar Arnfjörð Bjarmason writes:\n\n> +test_expect_failure 'no effect on show from --color-moved with --word-diff' '\n> +\tgit show --color-moved --word-diff >actual &&\n> +\tgit show --word-diff >expect &&\n> +\ttest_cmp expect actual\n> +'\n\nOK. Just for future reference ...\n\nIn this case it is OK to start with a \"document current failure\"\nthat gets turned into expecting success in a separate patch for two\nreasons, (1) it is by somebody other than the author of the patch\nthat fixes the breakage, and more importantly (2) the body of the\ntest is short enough.\n\nBut in all normal cases, please add a test that expects success in\nthe commit that implements a fix. A one-line change that turns\nexpects_failure to expect_success in the commit that implements a\nfix, when presented in a patch form with the standard 3-line\ncontext, does not often have enough post-context to show the\nbehaviour the test tries to exercise and makes reviewing harder. It\nalso makes it more cumbersome to cherry-pick the fix to a different\ncontext as the two patches must be kept together.\n\nThanks.\n\n"}]}