Here are the topics that have been cooking in my tree. Commits prefixed with '+' are in 'next' (being in 'next' is a sign that a topic is stable enough to be used and is a candidate to be in a future release). Commits prefixed with '-' are only in 'seen', and aren't considered "accepted" at all. They may be annotated with a URL to a message that raises issues but they are by no means exhaustive. A topic without enough support may be discarded after a long period of no activity (of course, it can be resubmitted when new interest arises).

Git 2.56-rc1 has been tagged. We may merge last-minute fixes before the final Git 2.56 release, but otherwise I do not expect any new feature topics to be ready before the final, so most of the in-flight topics will stay cooking in 'next' until then. As discussed at the Git Contributors' Summit, the version after the upcoming Git 2.56 will be Git 2.98, scheduled near the end of this year.

Copies of the source code to Git live in many repositories, and the following is a list of the ones I push into or their mirrors. Some repositories have only a subset of branches.

With maint, master, next, seen, todo:

	git://git.kernel.org/pub/scm/git/git.git/
	git://repo.or.cz/alt-git.git/
	https://kernel.googlesource.com/pub/scm/git/git/
	https://github.com/git/git/
	https://gitlab.com/git-scm/git/

With all the integration branches and topics broken out:

	https://github.com/gitster/git/

Even though the preformatted documentation in HTML and man format are not sources, they are published in these repositories for convenience (replace "htmldocs" with "manpages" for the manual pages):

	git://git.kernel.org/pub/scm/git/git-htmldocs.git/
	https://github.com/gitster/git-htmldocs.git/

Release tarballs are available at:

	https://www.kernel.org/pub/software/scm/git/

-------------------------------------------------- [Graduated to 'master']

* tz/doc-pack-refs-and-refs-fixes (2026-09-15) 2 commits
  (merged to 'next' on 2026-09-16 at 127f3b13fa)
 + doc/refs: backtick-quote commands and options consistently
 + doc/pack-refs: convert synopsis and options to new style

 Doc updates.

 Graduated to 'master'.
 source: <20260912191509.844954-1-tmz@pobox.com>

* yt/pathspec-negative-prefix (2026-09-14) 2 commits
  (merged to 'next' on 2026-09-16 at f4f244ea28)
 + dir: preserve pathspec prefix optimization with leading excludes
 + dir: do not apply prefix to negative pathspecs
 + Merge branch 'jc/pathspec-match-const' into yt/pathspec-negative-prefix

 The pathspec matching logic has been updated to avoid out-of-bounds
 memory accesses when a negative pathspec is shorter than the common
 prefix of positive pathspecs.

 Graduated to 'master'.
 source: <7CB757FB-1F2D-4EE6-8C31-8C2CD6D42397@ytausch.de>

-------------------------------------------------- [New Topics]

* rr/upload-pack-swap-shallow-wanted-ref (2026-09-16) 1 commit
 - upload-pack: swap wanted-ref/shallow-info responses

 The server-side protocol v2 response order for 'wanted-refs' and
 'shallow-info' has been swapped to match the client's expectation,
 fixing a fetch failure when the server has 'uploadpack.allowRefInWant'
 enabled and the client performs a shallow fetch.

 Will merge to 'next'.
 source: <20260916203221.5265-1-royceremer@gmail.com>

* bs/runtime-prefix-obsd-getexecpath (2026-09-16) 1 commit
 - exec_cmd: RUNTIME_PREFIX on OpenBSD systems

 Git on OpenBSD historically lacked an authoritative mechanism to
 resolve its executable path, as it doesn't support the
 KERN_PROC_PATHNAME sysctl.  With OpenBSD 8.0 introducing
 getexecpath(3), it is now utilized to provide proper RUNTIME_PREFIX
 resolution instead of relying on argv[0] fallback.

 Waiting for response.
 cf. <xmqq4ifio83d.fsf@gitster.g>
 source: <aqthQ3u4eW1wHCn7@humpty.home.comstyle.com>

* js/coverity-fixes (2026-09-17) 7 commits
 - test-read-midx: check midx_fill_entry() result
 - oss-fuzz: handle reftable iterator initialization failures
 - t/unit-tests: check reftable iterator initialization
 - rerere: do not record failed conflict resolution data
 - midx: validate incremental MIDX pack IDs
 - gpg-interface: make signature-prefix matching length-aware
 - wrapper: guard writev_in_full() against signed overflow

 Assorted fixes for code paths that are not careful with boundary and
 error conditions.

 Will merge to 'next'?
 cf. <xmqq5x03u0xz.fsf@gitster.g>
 cf. <xmqq1parorz4.fsf@gitster.g>
 source: <pull.2231.git.1789667556.gitgitgadget@gmail.com>

* yt/winansi-die-lasterr-fix (2026-09-20) 1 commit
 - compat/winansi: fix die_lasterr() argument formatting

 The error reporting machinery in the WinANSI compatibility layer has
 been simplified to pass the exact Windows error code and correctly
 format arguments for fatal errors.

 Waiting for response.
 cf. <xmqqwlsemsjt.fsf@gitster.g>
 source: <20260921062114.14450-1-yqtian668@gmail.com>

* hd/diff-no-index-reverse-fix (2026-09-18) 1 commit
 - diff --no-index: fix -R with file/directory conflicts

 A bug in git diff --no-index that mishandled reverse (-R) output
 when conflicts existed between a file and a directory has been
 fixed.

 Will merge to 'next'?
 cf. <9b97c14b-1d25-409b-a72c-d8caf298bf87@web.de>
 cf. <0c82d50e-f2c0-4db6-ade8-7a403cac73da@intel.com>
 source: <pull.2232.git.1789715946888.gitgitgadget@gmail.com>

* js/gitlab-ci-windows-rust (2026-09-19) 4 commits
 - ci(gitlab,windows): provide GNU Rust's host-linker support
 - ci(gitlab,windows): fix Rust setup for GitLab's MinGW build
 - ci(gitlab,windows): preserve exclusions during dependency setup
 - ci(gitlab,windows): provision GNU Rust for SDK-based MinGW builds

 The Windows GitLab CI job has been updated to provision and use the
 GNU Rust toolchain for MinGW builds, fixing job failures caused by
 incomplete Rust setup and missing linker support.

 Will merge to 'next'?
 cf. <CAOLa=ZQkJui77Xz2HL4sAWsaYLAzU6EPvBk+RzKkKxoiY_8aKw@mail.gmail.com>
 source: <pull.2233.git.1789819933.gitgitgadget@gmail.com>

-------------------------------------------------- [Cooking]

* jc/advice-config-set-global (2026-09-14) 1 commit
 - advice: give cut-and-pasteable advice to squelch

 The advice subsystem has been updated to suggest using the
 '--global' option when recommending a command snippet to squelch
 future advice messages, since global configuration is generally
 more appropriate for user-level preferences than per-repository
 settings.

 Needs review.
 source: <xmqq33vb4hma.fsf@gitster.g>

* sg/precompile-git-compat-util (2026-09-14) 4 commits
  (merged to 'next' on 2026-09-21 at 68acceee5b)
 + Makefile: precompile "git-compat-util.h"
 + Makefile: reintroduce REFTABLE_OBJS
 + cmake: remove any "$(*_OBJS)" variables when parsing Makefile for sources
 + Makefile: remove XDIFF_OBJS initialization

 The 'Makefile' has been taught to precompile 'git-compat-util.h' to
 speed up overall compilation, while excluding sources that do not
 include the compatibility header.

 Will cook in 'next'.
 source: <20260915060952.569535-1-szeder.dev@gmail.com>

* of/commit-reach-repo-awareness (2026-09-16) 1 commit
 - commit-reach: parse commits in the given repository

 The can_all_from_reach() and can_all_from_reach_with_flag()
 functions have been updated to accept a repository context,
 preventing bugs where submodule merging incorrectly reads from the
 superproject's commit-graph.

 Will merge to 'next'.
 source: <20260916134632.1424829-1-orestisflo@gmail.com>

* hn/range-diff-matched-only (2026-09-15) 1 commit
 - range-diff: add --matched-only to skip one-sided commits

 The 'git range-diff' command has been augmented with a
 '--matched-only' option to skip commits that are only present on
 one side, allowing users to easily focus on only the commits that
 have been retained.

 Needs review.
 source: <pull.2401.v3.git.git.1789458703432.gitgitgadget@gmail.com>

* jc/cocci-free-updates (2026-09-11) 2 commits
  (merged to 'next' on 2026-09-16 at 35b8bafa06)
 + cocci: FREE_AND_NULL(E) is safe to call on NULL
 + cocci: remove risky "if (!E) free(E)" conversion

 Updates to Coccinelle semantic patches to correctly handle the
 'FREE_AND_NULL()' macro and avoid generating broken transformations
 for negated pointer checks.

 Will cook in 'next'.
 cf. <96aca004-0df4-4e21-b60a-0288239122cc@web.de>
 source: <xmqqld978mok.fsf@gitster.g>
 source: <xmqqh5jv8m4w.fsf@gitster.g>

* jt/object-file-batch-fsync-fix (2026-09-13) 2 commits
 - object-file: flush transaction packfile before migrating objects
 - object-file: lift ODB reprepare out of packfile flush

 When 'core.fsyncMethod' is set to 'batch', the ODB transaction
 failed to properly flush large blob packfiles residing in the
 temporary directory before migrating the directory's contents to the
 main object store.  The execution sequence has been corrected by
 performing the packfile flush before the temporary directory
 migration, averting failure.

 Waiting for review.
 cf. <CAOLa=ZRCyowPgMABwsQBYTbW1cEf8PBBSszOEYQ4TKwLVKQfFA@mail.gmail.com>
 cf. <aqkGPcJdw3QagN0B@jtobler--20250820-SHC54>
 source: <cover.1789328612.git.jltobler@gmail.com>

* ak/refs-files-root-ref-lock (2026-09-11) 1 commit
  (merged to 'next' on 2026-09-16 at c8568ea3e1)
 + refs/files: avoid packed-refs lock for root ref deletion

 The files backend has been updated to avoid unconditionally locking
 the 'packed-refs' file when deleting a root ref (which are never
 packed).

 Will cook in 'next'.
 cf. <aqeLMG7lTJy3bM-h@pks.im>
 source: <20260912014609.535922-1-skariel@gmail.com>

* tb/rerere-wait-for-merge-rr-lock (2026-09-14) 2 commits
 - rerere: go on at a conflict when the lock stays busy
 - rerere: wait for MERGE_RR.lock, and let the gc skip it

 Instead of failing to record conflicts to be resolved immediately,
 wait while "rerere gc" is ongoing.

 Needs review.
 source: <pull.2214.v4.git.1789373061.gitgitgadget@gmail.com>

* pp/midx-write-skip-empty (2026-09-08) 1 commit
 - midx-write: skip writes with no object entries

 The `git multi-pack-index write` command has been updated to
 silently return success when there are no object entries to index.
 This avoids writing empty `multi-pack-index` layers, which
 previously caused subsequent incremental midx writes using the
 `--bitmap` option to fail when attempting to load the missing
 reverse index.

 Waiting for review.
 cf. <aqAkfGZtLJ97nG1m@com-79390>
 cf. <CAOWp8q5UqPrJQosjypdcq=KX1TKcVenAOoZUYfTBDKbRUwazTw@mail.gmail.com>
 source: <eef33827000cf106544174ed000129c2989af1cd.1788851232.git.pia@pierre.co>

* jk/merge-ll-tempfile-cleanup (2026-09-11) 3 commits
 - merge-ll: use tempfile API for external driver files
 - merge-ll: catch close() errors when writing external tempfiles
 - merge-ll: use strbuf to read back external merge result

 The external merge driver in 'git merge' now uses the tempfile API
 to create its temporary files.  This ensures that these temporaries
 are reliably cleaned up even when the merge driver or its parent Git
 process is terminated abruptly.

 Expecting a reroll.
 cf. <20260914165350.GA32247@peff.net>
 source: <20260911171044.GA1609692@coredump.intra.peff.net>

* mh/rust-crate-subdir (2026-09-16) 1 commit
 - move rust gitcore crate to a different subdirectory

 The Rust code and its 'Cargo.toml' file have been moved from the
 top-level repository root and 'src/' directory into a dedicated
 'rust/' subdirectory.  This avoids confusing Cargo's packaging
 mechanism when the Git repository is included as a submodule in
 other Rust projects.

 Waiting for review.
 cf. <yc4bgjoyxnm6o7q4gwols4d6zvdrq3ydw2c65wjdwof3hjyde6@2osv37jtuxx3>
 source: <20260917060415.2986259-1-mh@glandium.org>

* ps/ref-storage-format (2026-09-09) 13 commits
  (merged to 'next' on 2026-09-16 at 8937a6240b)
 + setup: allow "--ref-storage-format=" to specify a payload
 + setup: rename "init.defaultRefFormat" to "init.defaultRefStorageFormat"
 + t: rename GIT_TEST_DEFAULT_REF_FORMAT
 + setup: rename ref storage format environment variables
 + setup: refactor how we configure the ref storage format
 + refs: expose function to parse reference URIs
 + help: rename "default-ref-format" to "default-ref-storage-format"
 + builtin/rev-parse: rename "--show-ref-format" to "--show-ref-storage-format"
 + builtin/submodule: rename "--ref-format=" to "--ref-storage-format="
 + builtin/refs: rename "--ref-format=" to "--ref-storage-format="
 + builtin/clone: rename "--ref-format=" to "--ref-storage-format="
 + builtin/init: rename "--ref-format=" to "--ref-storage-format="
 + parse-options: allow for hidden aliases

 The terminology regarding reference storage formats has been unified
 across command-line options, environment variables, configuration
 variables, and source code, standardizing on the phrase "ref storage
 format" (e.g., `--ref-storage-format`, `'GIT_REF_STORAGE_FORMAT'`).
 Additionally, the `--ref-storage-format` option has been updated to
 accept payloads in the form `<format>://<payload>`.

 Will cook in 'next'.
 cf. <CAOLa=ZRxXimh8W-QBJB3VhbHOZWMEfWXB0ST0a=dOi+PNiR6sQ@mail.gmail.com>
 cf. <5e062976-d584-496d-84e0-e4b59c8f6876@gmail.com>
 source: <20260909-b4-pks-unify-ref-storage-format-v3-0-ca041fb40ad8@pks.im>

* ta/command-list-guides-sync-lint (2026-09-10) 2 commits
 - lint-docs: check the guide list in command-list.txt
 - command-list.txt: add gitformat-loose(5) and gitpacking(7)
 - Merge branch 'kh/doc-datamodel' into ta/command-list-guides-sync-lint

 A new linter test has been added to Documentation/lint-manpages.sh
 to ensure that all non-command manual pages (guides and developer
 interfaces) listed in Documentation/Makefile are present in
 command-list.txt, replacing an older comment that reminded
 developers to keep them in sync.

 Needs review.
 cf. <xmqq1pb2s33e.fsf@gitster.g>
 source: <20260910194351.20809-1-taahol@utu.fi>

* ap/var-broken-down-idents (2026-09-14) 1 commit
 - var: support broken-down idents, signing key, multiple args, and -z

 The 'git var' command has been extended to expose individual
 identity components ('GIT_AUTHOR_NAME', etc.) and the commit
 signing key, and can now accept multiple variables to query at
 once, safely formatting the output with a new '-z' option.

 Expecting a reroll.
 cf. <20260915220228.42819-1-andrewpleeter@gmail.com>
 source: <pull.2388.v8.git.git.1789426226860.gitgitgadget@gmail.com>

* tc/push-force-if-includes-fixes (2026-09-17) 3 commits
 - push: --force-if-includes should allow fast-forward
 - push: fix --force-if-includes non-branch advice
 - push: check pushed ref for --force-if-includes

 The '--force-if-includes' protection for 'git push' has been updated
 to consult the reflog of the local branch being pushed, rather than
 incorrectly checking the reflog of a local branch that shares the name
 of the remote destination branch.  The push advice for detached HEAD
 scenarios has also been adjusted to indicate that the remote ref
 cannot be verified locally.  A regression that caused perfectly valid
 fast-forward pushes to be rejected when reflogs were expired has been
 fixed.

 Needs review.
 source: <20260917224351.57171-1-tyler@tylercipriani.com>

* as/push-force-if-includes-no-reflog (2026-09-05) 1 commit
 - push: fix --force-if-includes when remote-tracking ref has no reflog

 The timestamp used for checking the reflog of a remote-tracking
 branch during 'git push --force-if-includes' was left uninitialized
 when the reflog was completely empty, which has been corrected.

 Waiting for review.
 cf. <xmqqjyowz9oq.fsf@gitster.g>
 cf. <20260909065639.47316-1-f@lex.la>
 source: <20260905171330.34646-1-f@lex.la>

* tb/rerere-lock-grace (2026-09-17) 3 commits
 - sequencer: disable auto maintenance in spawned commands
 - rebase, cherry-pick, revert: run auto maintenance when done
 - config: add git_config_append_parameter()

 The sequencer machinery (used by 'git rebase', 'git cherry-pick', and
 'git revert') has been updated to defer automatic maintenance tasks
 until the end of the operation, preventing nested 'git commit', 'git
 merge', and 'exec' commands from triggering GC operations that could
 contend for locks or delete open packs while the sequence is in
 progress.

 Needs review.
 source: <pull.2217.v5.git.1789670534.gitgitgadget@gmail.com>

* cc/lazy-fetch-trusted-bit (2026-09-08) 5 commits
 - builtin/upload-pack: set GIT_NO_LAZY_FETCH to 0 on trusted repo
 - promisor-remote: prevent infinite recursion when lazy fetching
 - upload-pack: read uploadpack.lazyFetchTrusted
 - setup: extract path_allowlist_apply()
 - promisor-remote: factor out lazy_fetch_objects()

 A new 'uploadpack.lazyFetchTrusted' configuration variable has been
 introduced to allow 'upload-pack' to lazily fetch missing objects from
 configured promisor remotes when serving trusted repositories.

 Waiting for response.
 cf. <xmqq7bkvy74h.fsf@gitster.g>
 cf. <xmqqqzj3wr24.fsf@gitster.g>
 cf. <xmqqmrtrwq0k.fsf@gitster.g>
 source: <20260908164129.560396-1-christian.couder@gmail.com>

* cc/early-scan-options (2026-09-02) 6 commits
 - fast-import: use early_scan_options() for --allow-unsafe-features
 - parse-options: build early scan options from a struct option array
 - parse-options: add parse_options_takes_argument()
 - rev-parse: fix "--" detection when it is an option value
 - bisect: fix "--" detection when a term name is "--"
 - parse-options: add early_scan_options()

 The process of parsing command-line options in commands that
 perform an early scan over their arguments (such as 'git bisect',
 'git rev-parse', and 'git fast-import') has been unified using a
 new early-scan sub-API, which parses and skips known options taking
 separate values to prevent logic bugs.

 Waiting for response.
 cf. <xmqqpkyviizc.fsf@gitster.g>
 source: <20260902161047.476753-1-christian.couder@gmail.com>

* ec/commit-fixup-options (2026-05-26) 2 commits
 - commit: allow -c/-C for all kinds of --fixup
 - commit: allow -m/-F for all kinds of --fixup

 Support for '-m', '-F', '-c', or '-C' options to supply a commit log
 message from outside the editor has been added for all 'git commit
 --fixup' variations.

 Expecting a reroll.
 cf. <CA+JQ7M__GOnM9LHt0txry-G2z2CKhdZr0b-rU=Yd_A0gCEwmaQ@mail.gmail.com>
 source: <cover.1779792311.git.erik@cervined.in>

* jc/checkout-refactor (2026-08-30) 8 commits
 - checkout: move post_checkout_hook() to checkout.c
 - checkout: wrap overly long lines
 - checkout: restructure switch, restore, and checkout entrypoints
 - checkout: extract branch setup and tracking helpers
 - checkout: extract option validation and pathspec helpers
 - checkout: validate stage and merge option compatibility in checkout_paths()
 - checkout: validate new branch name in checkout_branch()
 - checkout: pass cb_option explicitly to branch name parsers

 The front-end code for 'git checkout', 'git switch', and 'git
 restore' has been restructured to cleanly separate their pathspec
 and branch handling, eliminating a common bottleneck and paving the
 way to libify utility helpers.

 Waiting for response.
 cf. <xmqqo6el1xz0.fsf@gitster.g>
 cf. <xmqqse3x1y20.fsf@gitster.g>
 source: <20260830204835.1040408-1-gitster@pobox.com>

* ll/doc-pushcert-if-asked (2026-08-29) 1 commit
 - doc: remote-helpers: option pushcert if-asked

 The remote helper documentation for the 'pushcert' option has been
 updated to mention that it can also take 'if-asked', reflecting the
 existing implementation in the code.

 Needs review.
 source: <20260829183659.29947-1-lorenz.leutgeb@posteo.eu>

* ws/squelch-svn-migrate (2026-08-27) 2 commits
 - Makefile: add NO_GIT_SVN knob to skip building/installing git-svn
 - git-svn: don't print v1-layout migration noise when there's nothing to migrate

 Needs review.
 source: <20260827234345.1037130-1-wesleys@opperschaap.net>

* dw/config-read-both-global (2026-08-23) 3 commits
 - config: read global scope via config_sequence
 - config: let sequence require a successful file
 - path: use forward slashes in XDG config on Windows

 The git config --global read operations have been updated to respect
 both $HOME/.gitconfig and $XDG_CONFIG_HOME/git/config, fixing an
 inconsistency where only the former was read when both configuration
 files are present.

 Expecting a reroll.
 cf. <aqIvJhLLcCSnyaL4-delilahwu@linux.microsoft.com>
 source: <20260823-fix-config-list-global-home-and-xdg-v2-0-b29cc63f017b@microsoft.com>

* vv/branch-recurse-no-start-ref (2026-08-21) 2 commits
 - branch: allow recursion with no tracking name
 - branch: do not track a start point with no ref

 The --recurse-submodules option in 'git branch' has been fixed to
 avoid a crash when the start point is not a reference (e.g., a raw
 object ID).  The creation path now skips setting up tracking and
 properly forwards the absent tracking name to the submodule helper.

 Needs review.
 source: <20260822-vv-branch-recurse-no-start-ref-v1-0-46dc140acaa8@zitro.id>

* ps/odb-alternates-at-creation (2026-09-10) 9 commits
  (merged to 'next' on 2026-09-15 at 4acaa6a8aa)
 + odb/source: remove the ability to write alternates
 + builtin/clone: write alternates via `odb_create_on_disk()`
 + odb/source: support writing alternates when creating the database
 + builtin/clone: move setup of alternates for non-shared local clones
 + builtin/clone: move setup of alternates for shared local clones
 + builtin/clone: refactor handling of "--reference{,-if-able}"
 + builtin/clone: move around `setup_reference()`
 + builtin/clone: defer setup of the object database
 + setup: split up concerns of `init_db()`
 + Merge branch 'ps/odb-eagerly-load-alternates' into ps/odb-alternates-at-creation

 The setup of alternates has been deferred to object database
 creation time during clone, which drops the unused ad-hoc alternate
 writing API, simplifying the object database backend interface.

 Will cook in 'next'.
 cf. <CAOLa=ZRYsJL_0sKnfHD0PJO+5c+BKSMiuN20PeQHKJin82TJDw@mail.gmail.com>
 source: <20260910-pks-odb-write-alternates-at-creation-time-v5-0-8d10c4238edc@pks.im>

* as/utimensat-utimes (2026-08-21) 3 commits
 - compat/posix: drop legacy <utime.h> header and shims
 - treewide: use utimensat(2) instead of legacy utime(3p)
 - compat/posix: introduce utimensat(2) wrapper

 The codebase has been updated to use the newer utimensat() POSIX
 function instead of the obsolescent utime(), allowing
 high-precision timestamps while preserving fallback compatibility.

 Waiting for response for too long, stalled
 cf. <aonIVn-ZQoMKWCAd@fruit.crustytoothpaste.net>
 source: <pull.2209.git.1787322203.gitgitgadget@gmail.com>

* kn/receive-report-hook (2026-09-14) 5 commits
  (merged to 'next' on 2026-09-15 at aa6cbdb87e)
 + receive-pack: coccinelle fix
 + hook: introduce the receive-report hook
 + receive-pack: move message generation to separate function
 + receive-pack: drop static variables to track report status version
 + doc: add proc-receive hook info in 'git-receive-pack.adoc'

 A new hook 'report' is added to 'git receive-pack', which runs after
 reference updates and allows the server to filter or modify the
 packet-line status report sent back to the client.

 Will cook in 'next'.
 source: <20260910-758-introduce-hook-v10-0-06f9c506631c@gmail.com>
 source: <xmqqwlsn31gq.fsf_-_@gitster.g>

* ap/http-preserve-wwwauth-redirect (2026-08-19) 1 commit
 - http: preserve wwwauth_headers across redirects

 When an HTTP request triggers a redirect and the target yields an
 authentication challenge, the WWW-Authenticate headers received
 during the redirect are now explicitly preserved across the
 credential URL update, fixing an issue where they were incorrectly
 cleared.

 Needs review.
 source: <20260819-http-preserve-wwwauth-redirect-v2-1-4c61039432b0@nvidia.com>

* kh/format-rev-more-options (2026-08-18) 5 commits
 - format-rev: learn --abbrev, --color, and --date
 - doc: rev-list-options.adoc: factor out --date alts
 - format-rev: factor option variables into a struct
 - format-rev: place BUG calls first in callback
 - format-rev: use lower case for opts description

 The experimental 'git format-rev' has been taught a few more
 formatting options.

 Needs review.
 source: <V2_CV_format-rev_three_more_opts.bd3@msgid.xyz>

* gg/http-ssl-verify-status (2026-09-15) 1 commit
 - http: add http.sslVerifyStatus to check stapled OCSP responses

 The HTTP transport has been taught to check the revocation status of
 the server certificate using the stapled OCSP response during the
 TLS handshake via a new 'http.sslVerifyStatus' configuration
 variable.

 Will merge to 'next'?
 cf. <xmqqv785uha7.fsf@gitster.g>
 source: <20260915162348.97792-1-ggordon@gitlab.com>

* ty/repo-config-cleanups (2026-08-07) 3 commits
 - environment: remove inaccurate repo_config_values comments
 - environment: clarify repository config getter documentation
 - environment: drop redundant NULL checks in config getters

 Repository configuration getters in 'environment.c' have been
 simplified by removing redundant NULL checks.  The documentation for
 these getters in 'environment.h' has been clarified, and inaccurate
 section comments inside 'struct repo_config_values' have been removed.

 Waiting for response for too long, stalled
 cf. <aqOeHlPWer60LcoO@pks.im>
 source: <20260807085932.3958759-1-cat@malon.dev>

* dk/use-nsec-runtime (2026-09-11) 3 commits
  (merged to 'next' on 2026-09-15 at 67ef6d82ef)
 + core: convert build-time USE_NSEC into runtime core.useNanosec
 + environment: align repo_config_values_init with struct declaration
 + meson: expose knob for xmlto relative links in manuals

 The build-time knob 'USE_NSEC' for nanosecond stat precision has been
 converted to a runtime configuration 'core.useNanosec', allowing
 distributions to bundle one binary that adapts to filesystem
 capabilities dynamically.

 Will cook in 'next'.
 cf. <apVJCt4prIi2GgXp@pks.im>
 source: <cover.1789129924.git.ben.knoble@gmail.com>

* bc/restrict-hex-to-lowercase (2026-09-07) 7 commits
 - hex: allow only lowercase object IDs in breaking changes mode
 - t5324: adjust tests for corrupt commit-graph
 - object-name: use hexval
 - hex: label usages of hex parsing for object IDs
 - hex: make hex_to_bytes accept kind of hex to use
 - hex: allow specifying hex type with hex2chr
 - hex: add functionality for lowercase-only hex

 The parser for hex object names has been updated to reject uppercase
 hexadecimal characters when running in the breaking changes mode, in
 preparation for Git 3.0.

 Needs review.
 source: <20260907195941.1024289-1-sandals@crustytoothpaste.net>

* kj/repo-info-more-path-keys (2026-09-11) 7 commits
 - repo: add path.cdup
 - repo: add path.git-prefix
 - repo: add path.grafts with absolute and relative suffixes
 - repo: add path.index with absolute and relative suffixes
 - repo: add path.hooks with absolute and relative suffixes
 - repo: add path.superproject-root with absolute and relative suffixes
 - repo: add path.toplevel with absolute and relative suffix formatting

 The 'git repo info' command has been taught more keys to output
 paths of various repository components (such as the working tree
 root, superproject working tree, object database, etc.), supporting
 both absolute and relative path formats.

 Waiting for response.
 cf. <xmqqse3fbudk.fsf@gitster.g>
 cf. <xmqqcxujbser.fsf@gitster.g>
 source: <20260911144519.1011780-1-jayatheerthkulkarni2005@gmail.com>

* tc/last-modified-bloom (2026-09-01) 6 commits
 - last-modified: keep per-path Bloom filters for wildcard pathspecs
 - last-modified: check pathspec against Bloom filter first
 - revision: add Bloom check that includes parent directories
 - bloom: add helper to check if any key in a vector is present
 - revision: expose check for paths maybe changed in Bloom filter
 - revision: move bloom keyvec precondition into function

 The 'git last-modified' command has been optimized by using Bloom
 filters.  It now reuses revision walk filtering logic from 'git log'
 to pre-filter commits, and maintains per-path Bloom filters even when
 wildcard pathspecs are used.

 Waiting for response.
 cf. <aqJWihcFmX7tPio5@pks.im>
 cf. <aqJWX0INerT8F687@pks.im>
 source: <20260901-toon-speed-up-last-modified-v4-0-a09949800404@iotcl.com>

* ds/trace2-tolerate-failed-timestamp (2026-08-31) 7 commits
 - trace2: remove use of xcalloc()
 - trace2: remove use of ALLOC_GROW()
 - trace2: remove use of xstrfmt()
 - trace2: remove use of ALLOC_ARRAY()
 - trace2: remove use of xstrdup()
 - trace2: tolerate failed timestamp formatting
 - banned-die: create header for banning of functions

 Functions like `xstrfmt()` and `xcalloc()` have been banned from use
 in the trace2 API codebase to prevent calls to `die()` which lead to
 unwanted process exits and recursion when memory allocation fails.

 Needs review.
 source: <pull.2178.v3.git.1788197143.gitgitgadget@gmail.com>

* pz/fetch-submodule-errors-config (2026-07-16) 2 commits
 - fetch: add fetch.submoduleErrors to make submodule fetch errors non-fatal
 - submodule: fix premature failure in recursive submodule fetch

 The 'git fetch' command can now configure how submodule fetch errors
 are handled via 'fetch.submoduleErrors' and '--submodule-errors',
 making them non-fatal.  A premature failure during recursive submodule
 fetches has been fixed by deferring the error until the OID-based
 retry phase fails.

 Needs review.
 source: <20260716140956.1023740-1-paulius.zaleckas@gmail.com>

* fz/rebase-autosquash-empty (2026-08-27) 1 commit
 - sequencer: honor --empty when a fixup!/squash! empties its target

 A commit that is emptied by melding a 'fixup!' or 'squash!' commit
 during 'git rebase --autosquash' is now handled according to the
 '--empty' option, allowing it to be dropped, kept, or to halt the
 rebase.

 Waiting for response.
 cf. <511300fe-112d-4f20-bd3f-e401e68c4a27@gmail.com>
 source: <20260827-fz-autosquash-empty-v4-1-f98ffd575780@gmail.com>

* ij/subtree-reject-v2-config (2026-07-06) 2 commits
 - git-subtree: Bail out if we find output from Rust rewrite (test)
 - git-subtree: Bail out if we find output from Rust rewrite

 The shell script implementation of 'git subtree' has been updated to
 check for the presence of the configuration file of the new Rust
 implementation, preventing users from accidentally running the old
 script on repositories already managed by the new tool.

 Expecting a reroll.
 cf. <27219.20156.438730.881821@chiark.greenend.org.uk>
 source: <20260706115816.20267-1-ijackson@chiark.greenend.org.uk>

* mm/line-log-limited-ops (2026-09-02) 7 commits
 - diffcore-pickaxe: limit -G to the -L tracked range
 - diff: support --check with -L line ranges
 - diff: support stat formats with -L
 - diff: extract a line-range diff helper for reuse
 - diff: emit -L hunk headers via xdiff's formatter
 - diff: simplify the line-range filter by classifying removals immediately
 - diff: rename line-range filter struct and clarify fields
 (this branch is used by mm/diff-process-hunks.)

 The 'git log -L<range>:<path>' command has been taught to limit
 various 'diff' operations, such as '--stat', '--check', and '-G', to
 the specified range and path.

 Needs review.
 source: <pull.2152.v3.git.1788411919.gitgitgadget@gmail.com>

* hn/history-squash (2026-09-08) 8 commits
  (merged to 'next' on 2026-09-15 at 28dd773b8b)
 + history: support editing squashed commit messages
 + history: create squashed commits without editing
 + history: protect branches when squashing a range
 + history: validate squash revision ranges
 + history: add skeleton for squash subcommand
 + sequencer: share the squash message marker helpers and flags
 + history: give commit_tree_ext a message template
 + history: extract helper for a commit's parent tree

 The experimental 'git history' command has been taught a new 'squash'
 subcommand to fold a range of commits into a single commit, with any
 descendants replayed on top.

 Will cook in 'next'.
 cf. <xmqqik4fv4v1.fsf@gitster.g>
 source: <pull.2337.v15.git.git.1788900119.gitgitgadget@gmail.com>

* tb/midx-incremental-custom-base (2026-06-12) 3 commits
 - midx-write: include packs above custom incremental base
 - midx: pass custom '--base' through incremental writes
 - t5334: expose shared `nth_line()` helper

 The 'git multi-pack-index write --incremental' command has been
 corrected to properly honor the '--base' option.  Previously, the
 custom base was ignored by the normal write path; packs from layers
 above the selected base were incorrectly skipped by the pack exclusion
 logic, and reachability closure for bitmaps was broken.

 Expecting a reroll.
 cf. <an4uIQA09rDCwwBp@com-79390>
 cf. <apUbQ4S-zJGtBeu2@pks.im>
 source: <cover.1781294771.git.me@ttaylorr.com>

* mm/diff-process-hunks (2026-08-13) 11 commits
 - fixup! diff: consult oid-only hunk providers via diff.<driver>.process
 - diff: consult oid-only hunk providers via diff.<driver>.process
 - userdiff: add diff.<driver>.process config
 - sub-process: add a gentle status read
 - sub-process: separate process lifecycle from hashmap management
 - blame: read precomputed hunks
 - diff: read precomputed hunks for stat output
 - diff: record precomputed hunks during stat output
 - diff-hunks: add the store format, library, and command
 - diff: introduce a hunk provider interface
 - gitattributes: document how external diff drivers relate to diff features
 - Merge branch 'mm/line-log-limited-ops' into mm/diff-process-hunks
 (this branch uses mm/line-log-limited-ops.)

 A new 'diff.<driver>.process' configuration has been introduced to
 allow a long-running external process to act as a hunk provider,
 enabling external tools to control which lines Git considers changed
 while leaving all output formatting (word diff, color, blame, etc.) to
 Git's standard pipeline.

 Expecting a reroll.
 cf. <CAC2Qwm+kzT_3_GKrpay=JLGYsxS10oWCg2MJPHrCVogFHA0OdA@mail.gmail.com>
 source: <20260801174156.2998808-1-mmontalbo@gmail.com>

* kh/format-patch-range-diff-notes (2026-08-24) 3 commits
 . format-patch: learn --[no-]range-diff-notes
 . revision.h: rename struct member to reflect notes role
 . format-patch: simplify get_notes_arg parameters

 The 'format-patch' command has been updated with options to
 configure notes specifically for range-diff output, allowing them to
 differ from the notes displayed on the patches themselves.

 Expecting a reroll.
 cf. <8f0a076b-4822-44e2-a842-cc1e39ae1c1d@app.fastmail.com>
 source: <CV_format-patch_learn_--range-diff-notes.c57@msgid.xyz>

On Tue, Sep 22, 2026, at 02:11, Junio C Hamano wrote:

I’ll comment since it’s almost been a month. It’s a straightforward reroll but I haven’t been able to do the about hour’s worth of work. I should get some time at the latest on the weekend.

Hi Junio,

On Mon, 21 Sep 2026, Junio C Hamano wrote:

Would you say that the following is an accurate characterization of the timeline, so that people who need to plan dependent projects can rely on it?

The participants of the Git Contributor Summit agreed on a certain release shape, rather than timing: release v2.99.1 and v3.0 simultaneously, differing only in the breaking-change defaults. The timeline discussed was v2.56 in September, v2.98 in December, then v2.99/v3.0 around March 2027; April also came up. Dropping v2.57 in favor of v2.98 was an explicit decision. This would make v3.0 principally a deliberate compatibility transition, rather than a separate batch of features.

Rust was treated as mandatory for v3.0, and the explicit check for objections drew none from anyone in the room. Compared with the contentious portability discussion on the Git mailing list, that is a notable signal. It does not establish that the portability problems themselves are solved.

One of Git v3.0's bigger-impact challenges is SHA-256 interoperability, which is purportedly done, but not on the Git mailing list yet, and there was affirmation that it will handle historical tags too. While GitLab already has support for SHA-256, GitHub has it only in private preview with general availablility likely before the end of the year. JGit does _not_ have SHA-256 support, and no participant knew of anyone funding it. The ecosystem transition therefore remains uneven.

Is this a fair summary of the "Git v3.0" breakout session at the Git Contributor Summit, from your point of view?

Ciao, Johannes

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

My outline was deliberately limited up to end of this year as I am hesitant to say beyond that point before the meeting notes are made public. I am not sure who will be releasing it to the public and when, though with the open nature of this community I believe it will happen soon. I do not recall anything controversial in the 3.0 section of the meeting notes.

It signals that the room was smaller than the list. Or enough time passed since we discussed some minority platforms having trouble with Rust the last time to change the situation. Or little bit of both ;-)

Thanks.

Hi Junio,

On Tue, 22 Sep 2026, Junio C Hamano wrote:

Fair enough: I did ask you to confirm my summit summary. Let me separate that from what I need for planning: a proposal from you as release maintainer can be discussed on the list on its own merits, without waiting for publication of the meeting record or claiming summit consensus.

Could you use the next What's cooking to outline your working release plan: whether paired 2.99.1/3.0 releases in March or April 2027 would be your proposed target, what conditions could move it, and when we should review that target?

If you cannot yet choose a target, could that update identify what needs resolving and when you expect to revisit the choice?

Your assessment would give dependent projects a common baseline to coordinate around. A provisional planning assumption would be useful; it need not be a guarantee.

Thanks, Johannes

Hi Junio,

On Tue, 22 Sep 2026, Junio C Hamano wrote:

I originally wrote this summary of the Git Contributor Summit only for my own records, but since you hinted at wanting some meeting notes, figured it might be interesting to other people, too. So here goes my distillation of the breakout-sessions from this year's Contributors' Summit. Many of these ideas still need discussion on the list; proposals below are not project-wide decisions.

Security mailing list and process

The security list is seeing a large influx of outside reports, apparently often AI-assisted or generated. Duplicates and reports outside Git's security model consume triage time, while some patches have waited for months. Waiting for an empty queue is not a workable release criterion: we need to release the fixes we have, rather than hold them indefinitely while more reports arrive. No fixed release cadence or guaranteed response time was settled.

Documenting Git's security model would save repeated explanations of what does and does not constitute a vulnerability. (Personal note, not discussed at the Summit: Stolee had tried to start a conversation about Git's security boundary a long time ago, but nobody replied. Maybe the AI onslaught will provide enough motivation to get that discussion going.) Moving non-security bugs to the public list more readily was also encouraged. A timeout after which public discussion would be presumed OK was proposed, but not agreed.

More company staffing would help, without turning this into an obligation for volunteers. Paying for dedicated help was discussed, with onboarding costs a concern. Using AI for triage or fixes raises confidentiality and DCO questions of its own. (Personal note: There seemed to be some sentiment in the room that contradicted the earlier agreed-on finding on the mailing list that using AI for triaging and for investigating wasn't a copyright concern and should therefore be considered permissible.)

The release bottleneck is not really tag automation. Merging and backporting fixes, preparing advisories, and handling CVEs take work, and too much of that knowledge lives in people's heads. Peff volunteered to start a public discussion of the process and dig up existing resources. (Personal note: I have done that merging, backporting, etc plenty of times, and I don't think that it is the bottleneck, and I was rather surprised to hear that it is complicated. Sure, there are the expected merge conflicts when merging `maint-*` branches, and running -- and fixing! -- CI on all of the tags in a private repository should go without saying, but that's all craft of the trade. Rather, the indecision and lack of engagement on the git-security mailing list is what I see as the blocker. I'd happily volunteer to juggle those branch thickets if that was truly the make-or-break issue here.)

Microsoft's release process reportedly needs about seven weeks. There were no objections in the room to proceeding without waiting for that schedule. (Personal note: That release process was misrepresented, which is surprising, as I coordinated two or three Git for Windows security bugfix releases _on the git-security list_ since the most recent Git security bugfix release, it's always the same thing: release on a second Tuesday of the month, three weeks before that the patches need to have settled, everybody goes home with a dependable timeline. It's not really that big of a deal. Testing patches, constructive feedback, these are the things that are missing and therefore blocking the process. I sensed a lot of finger-pointing in this discussion. I mean, I don't blame anybody for avoiding security work: it is stressful and intense. The responsibility for getting things wrong is enormous. I know that because I've done my share of that, probably more than most in the Git project, and I will do even more in the future. But when I don't have the time, or the energy, I am aware that I, myself, am the bottleneck; I don't need to blame others.)

Git v3.0

The proposed target is spring 2027: v2.56 in September 2026, v2.98 in December, then v2.99 and v3.0 next spring. The jump in version numbers is intended to signal the approaching breaking changes. March and April were both mentioned; the precise timing is not settled.

There was agreement on releasing v2.99.1 and v3.0 together, differing only in the BREAKING_CHANGES (v3.0 switches them on). That keeps the transition separate from another round of feature development. Nobody in the room objected to Rust becoming mandatory in v3.0. (Personal note: probably because Randall wasn't there, to say that NonStop support would be a blocker and that Git please wait.) A v2.x LTS remains an open question, with Gentoo mentioned as an interested party. (Personal note: I am still advocating for an in-tree Long Term Support branch, and since Junio indicated that he's less than eager to take care of that, I would love for Patrick Steinhardt to be the "LTS lieutenant", I vaguely remember that he said he'd do it if asked, and I trust his judgement, so I'd ask.)

SHA-256 support across the ecosystem had been a blocker. GitHub reported experimental support, with general availability expected around November. GitLab already had public, non-experimental support, and libgit2 supports it, too. JGit remains a gap; Google was not planning to fund that work.

The SHA-1/SHA-256 interoperability work, including historical tags, was reported to be implemented but not yet sent to the list. (Personal note: I think that the room seriously "mis-underestimated" the real-world impact of this. The code is not even on the Git mailing list, and the ramifications of not having a robust plan how to deal with partial clones or submodules or even signed tags strikes me as a dealbreaker. I would not be surprised if the decision to enforce SHA-256 as default would have to be revisited before v3.0, and possibly overturned.)

Documentation

Julia's work highlights the gap between documentation written by people who know Git inside out and users who do not yet know what objects, the index, or upstream mean. We need approachable learning material as well as reference documentation. Both need work; keeping manpages concise does not mean they cannot have better explanations and examples. (Personal note: I am beyond excited that Julia, whose work I have always admired, got interested in improving Git's documentation, which is in dear need of being improved, mainly because it does not cater to the majority of Git users out there who are unlikely to wander onto the Git mailing list, ever. I just hope that old-timers who really do not need the documentation nor understand the need of those who do need it show enough appreciation for the fresh views and for Julia's understanding of the target audience.)

Discoverability matters, too. The website (https://git-scm.com/) needs clearer entry points for learning Git, and existing guides are harder to find than manpages. Missing subsection links are another improvement we could make incrementally. (Personal note: Judging by the history of that site, I do wonder whether the core Git contributors are interested in helping this effort at all. For example, there are a growing number of PRs suggesting to add new UIs to the growing list, but I gave up reviewing them because I was the only one doing so.)

There was support for replacing outdated material and for merging useful improvements, then iterating, rather than trying to perfect everything before it lands. Bringing user feedback to the list without flooding it remains a challenge.

The current funding covers only 100 hours split between two people. Additional project and company funding was encouraged; brian, Emily, and Mark offered to explore company support. (Personal note: I had tried, back when GitHub still funded my team, to start something like that, without any success. To the contrary, even Git for Windows and Git Credential Manager got defunded.)

On the tooling side, using only Asciidoctor instead of maintaining both AsciiDoc and Asciidoctor support was proposed as a possible Git v3.0 change. Distribution support and rendering differences need checking, with doc-diff suggested for comparing the outputs. Patrick filed an issue during the discussion. (Personal note: AFAIU the AsciiDoc spec is now maintained by Asciidoctor, and I am aware already of one change that was made to the spec without adapting AsciiDoc accordingly. So the entire discussion might be quite moot already.)

Other ideas included richer diagrams for HTML while retaining text versions for manpages, and privacy-respecting traffic measurements to help prioritize documentation work. No diagram format was chosen, and caching and AI scraping complicate getting useful traffic data. Mermaid was proposed, and even GraphViz. (Personal note: I added support for Mermaid diagrams to https://git-scm.com/, but it turned out to be too limited, so I added GraphViz support. The support code for this is a bit of a beast, having a wasm version of GraphViz for development, pre-rendering the diagrams as SVG and as PDF during deployment of the site; it was quite a bit of fun to implement all that.)

Outreachy sponsoring

The goal is to support three interns in the round starting in early December, at $10,000 each. The corporate sponsorship previously provided by GitLab and GitHub has dried up, leaving Git itself to pay. There are company contacts to follow up with; Emily offered to ask Google's OSPO, without high expectations. No new sponsorship commitments were made. (Personal note: I don't think that these internships provide enough publicity to give companies much of an incentive to fund this. Which I find a bit of a shame, Outreachy in particular does a lot of important, good work, and if I wasn't so constantly overworked, I would want to mentor again; I always found it rewarding, even if I hold myself to a quite high bar which is quite draining.)

A related point for Git Merge 2027: announcing the location early would help Outreachy and GSoC interns plan attendance. No location was selected.

Pluggable object database

Patrick's pluggable object database is working, but it is not complete: commit-graph and multi-pack-index integration are still outstanding, and a repository extension is planned. (Personal note: It might be interesting to see whether implementing a storage backend is easier in core Git or in another Git-compatible implementation. JGit should be a natural target, having originated within BigTable-sized constraints, i.e. a different storage system, but funding seems to have dried up, there's not even SHA-256 support, so JGit might not be as hackable as it once was.)

Content-defined chunking prompted an important distinction between changing how objects are stored and changing the logical object model. Starting at the storage layer would let us preserve existing blob OIDs rather than require ecosystem-wide changes. A new pack/index format could provide another representation of the same object, much as deltas do today. No particular representation was agreed. (Personal note: It is curious to me why nobody tought about inventing a "meta blob", i.e. an object much like a tree object, except that it stitches together a larger blob. This would allow for the content-defined chunking that `rsync` already championed, way before Git was born! It would have allowed a Git native large file support worth writing home about, and could have replaced Git LFS. Xet (https://huggingface.co/docs/hub/xet/index) would not have had to be invented, and it would have allowed game development to move to Git. I can only imagine that the time it would cost to get even the first patches of this into core Git would be seen as prohibitive by any company who may have considered the effort.)

There is also an API question: does a backend seeing only object content have enough context to make good storage and delta choices, or should it receive richer information? More searchable tree storage and a Git "commit cloud" were other possibilities raised, not committed plans. (Personal note: At a previous GitMerge, Facebook presented their work, see e.g. https://github.com/facebook/sapling/blob/main/eden/mononoke/blobstore/packblob/README.md, which includes separating actual storage from transport. That is, already at push time, derived metadata is computed in async jobs which provide several potential deltas ready-to-go when a client clones or fetches. They reported clones with regular Git clients that are twice as fast, just because the server doesn't need to spend much compute on the data it sends. So there is a lot to be learned out there already.)

AI

The current SubmittingPatches policy is rooted in DCO certification and advice from SFC lawyers. The unresolved question is whether, and to what extent, contributors can certify AI-generated code. There was substantial disagreement about acceptable use, provenance and legal risks, community trust, review burden, and whether the current caution excludes useful tools. No policy change was agreed. (Personal note: You'd think that the opinion of lawyers is taken at face value, but no, it seems that some core Git contributors seem to disagree with the lawyers in favor of their own opinion...)

Several participants found language and proofreading assistance useful. A particular concern was submissions where the human does little more than relay agent output, leaving reviewers to deal with the consequences. The influx of poor GSoC contributions was one example. Attribution such as Assisted-by was suggested to make tool use clearer, but attribution alone does not answer the quality or DCO questions. (Personal note: my precedent of "Assisted-by" was called out as helpful, and I do think it is. I make a difference between AI-generated and AI-assisted. I'm not a fast typer, so I benefit a lot from being able to tell an LLM to please refactor out these four lines with the appropriate signature. There's not much creativity in there. I also like to let AI present me the call graphs for certain code locations, because due to the choice of C, which thanks to the C preprocessor is not easy to analyze statically, there are no competent tools other than LLMs that I can use for the task. I was highly surprised, though, to see how much enmity against AI in general was voiced, not by many, but many, many times, and how that contrasts with the Linux project which I hitherto had not considered to be as particularly open to modern practices.)

brian and Taylor agreed to put differing policy proposals on the list. The suggested process is to have alternatives examined by SFC counsel, make the risks clear, and then consider a vote. Emily volunteered to organize the voting procedure. Eligibility and the details remain open; Junio's authority as maintainer remains central. (Personal note: Since Taylor works for OpenAI now, I was not surprised by his stance, but brian works at GitHub, home of GitHub Copilot, and I am not sure how favorable their employer would look at their semi-public utterings about AI...)

Protocol v2 for pushes

There are concrete use cases now: repositories with millions of refs, including a reported 896 MB ref advertisement. Reftable improves ref update throughput, but does not by itself solve the advertisement problem. Nobody objected to push protocol v2, and the fetch-v2 infrastructure already provides much of the foundation.

The discussion covered advertising fewer refs, letting clients identify useful branches, and replacing large advertisements with a few rounds of push negotiation. Negotiation results could also help optimize the server's connectivity checks. Some improvements might be possible in the existing protocol before introducing v2.

We need to measure the tradeoffs rather than assume fewer bytes means faster pushes. One example involved a shallow push taking 35 seconds instead of two because of work to minimize the transfer. Shallow boundaries and unrelated histories complicate the proposed heuristics.

SHA-256 interoperability is another reason to want push v2: the current push protocol requires using the server's primary hash algorithm. Negotiation could make that more flexible.

Forge replication to thousands of mirrors would benefit from finding out cheaply whether refs have changed, rather than downloading full advertisements from every target. Checksums, ETag-like values, and reftable generation numbers were discussed, with concurrent updates complicating the picture.

Compact or compressed advertisements are also worth exploring and benchmarking, possibly reusing reftable's format. That does not mean sending the server's actual reftable, including hidden refs. There are several promising directions here, but no final design yet.

Some breakout sessions were planned, but apparently had to be cut.

Personal notes: I wasn't present for all of the sessions, in the afternoon I had other commitments; Therefore these notes (which AI assisted me in distilling) came partially from what I dictated and partially from the shared Google Document in which a few volunteers gracefully wrote notes. I found it challenging to connect as a remote participant. The link to the Google Meet, as well as control over the lobby thereof, seems to have been restricted to at most a few people, which might have contributed to the long waiting time before I was allowed in, and it definitely contributed to my comments not reaching the discussion in time to have an impact. I would have loved for Junio or the other two brave souls who also participated remotedly to have had more "air time". I am still a fan of the idea to have more frequent, smaller, virtual Contributor Summits, organized by a rotating cast. (Maybe I can get Emily to host the next one.) I was very happy that Junio was participating, as he _is_ the project lead, and in past Contributor Summits decisions were taken without him, which I found odd. Timing was really challenging for him, though, it was way past midnight for him. I'm all the more grateful that he did participate.

Final remark: This summary is obviously biased. I lightly edited it to separate better between my personal views and a hopefully unbiased account of what was discussed, and how, and by who. Nevertheless, I am but human. As a consequence, I would be delighted if other participants would share their summaries, so that my bias can be balanced out.

Ciao, Johannes

Hi Johannes,

Thanks so much for this summary, I didn't participate at the contributor summit as I was leading one of the breakout sessions in the morning and had a plane to catch in the afternoon, so I'm very grateful of your summary.

On 22/09/2026 19:44, Johannes Schindelin wrote:

Pluggable backend implementations for JGit have been possible for quite some time, although, admittedly, I don't think any made it to production. Maybe at the time when this was introduced(16 years ago!!) by Shawn[1] in JGit it wasn't fashionable yet and so the project was never carried forward. I know Luca submitted a talk for the Gerrit User Summit to present a Cassandra back-end, for which I can see conversation started 10 years ago[2].

Regarding JGit support's for SHA256, I know some corporations have had interest in sponsoring this work and have discussed potentially implementing together it with GerritForge, but, as far as I know, work isn't ongoing yet. JGit is still very much developed and kept up to date with great effort from the community, so I believe it to still be as hackable as it was, there just hasn't been enough interest for SHA-256 yet, which I agree is a shame, hopefully in the near future this gets remediated.

[1] https://github.com/spearce/jgit_cassandra [2] https://groups.google.com/g/repo-discuss/c/IekVPmow0yE

On Tue, Sep 22, 2026 at 2:45 PM Johannes Schindelin <Johannes.Schindelin@gmx.de> wrote:

Thanks for this!

If you are able/willing to point me towards the [relevant] Gentoo folks, I'd love to chat with them. It's been my distribution of choice for the last year and, while I've helped some on the "packaging Git" side, I have to imagine the 2.x request comes more from thinking about the surrounding things Git is used for (both developing the system and running it, such as the 3rd-party repository syncs or live package installs that use Git).

Anyway, I'd like to help them out :)

Don't forget that Julia received a grant to continue working on docs!

https://www.sovereign.tech/news/meet-the-2026-sovereign-tech-fellows

Did I understand correctly from a recent discussion that v2 also requires more "setup" and is thus not easy to use for smaller "local" hosts? If so, making it more accessible might also be nice.

In other projects I've participated in, more frequent smaller check-ins have been helpful. It also feels less icky to miss one for whatever reason, since the next one is not too far off!

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

(I decided to split up this topic into a separate mail thread)

I think we agreed on that. Because we realize the influx will not stop any time soon.

This is something I was planning to drive after the summit, so I agree.

I was not aware of that, maybe that happened before I started participating there. But I'm happy to get that conversation going again. I'll try to digg up that discussion.

Yes.

That is true, but then still someone has to do the work. We're already lacking people to work on /real/ reports, so having people to fix bugs reported is also a staffing issue.

I'm against that, but I don't think we need to agree on that already. We can think about this after settling the other stuff.

All I can say for now, we at GitLab plan to invest more people power into dealing with security reports.

Yeah, no consensus here. The concern wasn't as much copyright, but passing vulnerabilities to a model, making it possibly train on that exposing that information to who-knows-where.

As I pointed out during the conversation, I just don't know how to do this.

It's complicated, because we (or I at least) doesn't know how.

I'd love to learn more about this.

I'm very grateful for that. As I mentioned (not sure that made it into the notes), I'd love to shadow someone doing this to learn from. This can help me get this going myself and that would distribute the load and knowledge among more people, for which seems to be a real need.

I wouldn't say this is surprising, I think most people just don't know the details.

Well, thanks, that's clear.

Personally I think we can try to follow that schedule, if possible. We currently have a few patches waiting for months to be released, and that's not because of the Git for Windows release schedule, but more because the lack of call to action. Those easily can get out with the GfW cadence.

But some other cases, like the 0day Elijah has been working on this month, I'm not sure that can wait for GfW?

I wouldn't say so, I would blame it on the lack knowledge about the process. Or maybe that's only me speaking.

That's why I'd like to learn from you.

You are not the bottleneck, or at least, you don't need to be the bottleneck, we should set up things to spread out the load.

So I would like to suggest you and me collaborate closely to get a next security release out and then we can go from there.

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

I've expressed myself multiple times as well how excited I am to have Julia working on this, but it cannot be expressed enough.

I think the old-timers do, but it takes skills to have a very deep understanding and still being able to explain things to newbies.

I share the blame here. Some time ago I volunteered to step in to do maintainance work on git-scm.com, but I haven't been devoting as much time as I would like.

Talking about the UI list, that's a problem which I'm not sure worth discussing here, but to folks interested, there is some context in the PR[1] you created.

[1]: https://github.com/git/git-scm.com/pull/2179

Iteration will be key here, and I would say some steps have been taken already. Very tiny steps though.

Finding a medium to gather user feedback is the problem. I think Discord is a better place than the mailing list (assuming that's what you mean by "list"?).

This feels feasible. I think brian suggested to use Open Blocks[2] and have a man-page ASCII version next to /something else/.

[2]: https://docs.asciidoctor.org/asciidoc/latest/blocks/open-blocks/

For the record, we have been talking about this[3] in the past.

[3]: https://github.com/git/git-scm.com/issues/2054

Yeah, that's the issue.

Thanks for that! They don't look bad on the cheat sheet[4].

[4]: https://git-scm.com/cheat-sheet#combine-diverged-branches

Toon Claes <toon@iotcl.com> writes:

;-)

Not just these two skills, but you need to know what are the points new people find hard to get when they are starting, and that is the weakest spot for our old-timers.

IIRC, the view in the room was in favor of keeping and incrementally improving the manual pages and reference material while rewriting tutorials and material on concepts from scratch. I very much agree with the "rewrite from scratch" part, because the concepts have evolved and the world view have been updated even if the concepts at the very core of the system may have not changed.

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

As I wrote, after the current cycle ends at the end of this month, a 10-12 week cycle including the end-of-year slowness would mean the next cycle 2.98 will end at the end of this year. Expolation from there, 2.99 will be March 2027.

The consensus in the room was that we want to use 2.99 as a signal that something big is coming, so between 2.99 and 3.0 needs to be some lead time for "advertisement". This lead time between 2.99 and 3.0 does not have to be the usual 8-to-12-weeks full release cycle.

I do not think there was a firm agreement on the date for 2.99.1 and 3.0. Potential factors mentioned in the room included that we may want to match the LTS release schedule of major distros. My preference would be to give a month after 2.99 to apply only accumulated bugfixes and nothing else and tag it as 2.99.1, which means 2.99.1 would be April 2027.

The contents of 3.0 should be identical to 2.99.1 except for the breaking changes are enabled in 3.0 while they are disabled in 2.99.1. Volunteers can run 2.99.X series indefinitely to help LTS distributions.

At the release engineering level, I am very tempted to keep the WITH_BREAKING_CHANGES Makefile knob in 3.0 release in order to keep the differences between 2.99.1 and 3.0 to absolute minimum, and then remove the "dead code" that is used when WITH_BREAKING_CHANGES is not enabled from 3.X at our leasure.

So the above is what I have in mind, shaped mostly around the concensus at Contributor's summit (or at least how I understand what the concensus was), with my preference filling in what was not firmly decided in the room.

Good enough?

Junio C Hamano <gitster@pobox.com> writes:

It was so full of typoes and grammos because I didn't pass it thru spell checker as usual. Sorry about that. Here is a replacement.

As I wrote, after the current cycle ends at the end of this month, a 10-to-12-week cycle including the end-of-year slowness would mean the next cycle, 2.98, will end at the end of this year. Extrapolating from there, 2.99 will be March 2027.

The consensus in the room was that we want to use 2.99 as a signal that something big is coming, so there needs to be some lead time between 2.99 and 3.0 for "advertisement". This lead time between 2.99 and 3.0 does not have to be the usual 8-to-12-week full release cycle.

I do not think there was a firm agreement on the date for 2.99.1 and 3.0. Potential factors mentioned in the room included that we may want to match the LTS release schedule of major distributions. My preference would be to give a month after 2.99 to apply only accumulated bugfixes and nothing else, and tag it as 2.99.1, which means 2.99.1 would be April 2027.

The contents of 3.0 should be identical to 2.99.1 except that breaking changes are enabled in 3.0 while they are disabled in 2.99.1. Volunteers can run the 2.99.x series indefinitely to help LTS distributions.

At the release engineering level, I am very tempted to keep the WITH_BREAKING_CHANGES Makefile knob in the 3.0 release in order to keep the differences between 2.99.1 and 3.0 to an absolute minimum, and then remove the "dead code" that is used when WITH_BREAKING_CHANGES is not enabled from the 3.x series at our leisure.

So the above is what I have in mind, shaped mostly around the consensus at the Contributors' Summit (or at least how I understand what the consensus was), with my preference filling in what was not firmly decided in the room.

Hi Junio,

On Wed, 23 Sep 2026, Junio C Hamano wrote:

Thank you so much! This will make it much easier for me to plan out the Git for Windows roadmap, such as switching to Rust-based builds and integrating Git Credential Manager v3.0.

Ciao, Johannes

Hi Daniele,

On Wed, 23 Sep 2026, Daniele Sassoli wrote:

I'm glad it was helpful!

I'm glad to hear that JGit isn't stalled, even though I doubt that the SHA-256 support could be done without any corporate support.

Ciao, Johannes

Hi Toon,

On Wed, 23 Sep 2026, Toon Claes wrote:

I'm glad that you agree that Patch Tuesday isn't all _that_ much of a friction point, this will allow some of my colleagues to relax visibly.

Excellent! Let's continue this conversation in a more private setting ;-)

Ciao, Johannes

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

Note that I consider it risky to treat the timeline as already set in stone.

In particular, if we find that 2.99 needs a longer stabilization effort, we may need to slip 2.99.1 by a month or follow it with 2.99.2 or even 2.99.3, and 3.0 may have to coincide with one of these later maintenance releases. Anything beyond the end of this year is in "we do not yet know and we will play it by ear when the time comes" territory.

Thanks.

On 24/09/2026 5:30 am, Junio C Hamano wrote:

[snip]

Back in January, on the cygwin-announce list[1], an experimental rust package was announced. This package was marked experimental and unmaintained in the cygwin setup program. I was hoping for a more 'official' package to emerge before trying it out on git. (the package was version 1.91.0 of rust built from a source tarball). However, there has been no sign of a formal supported (test or production) package since then (there is still time, of course). ;)

Anyway, this thread prompted me to try the experimental package:

  $ vim config.mak # comment out NO_RUST
  $ cat config.mak
  DEFAULT_TEST_TARGET=prove
  GIT_PROVE_OPTS=--timer -j8
  #NO_RUST=1
  NO_DC_SHA1_SUBMODULE=NoThanks
  DEVELOPER=1
  $ 

Having fetched today, the 'master' branch @0f8e75abeb is v2.56.0-rc2 with
the branch 'en/no-amend-during-conflicts' reverted.
  
  $ make >out1 2>&1
  $ ./git version
  git version 2.56.0.rc2.1.g0f8e75abeb
  $ git describe
  v2.56.0-rc2-1-g0f8e75abeb
  $ diff out out1
  1c1
  < GIT_VERSION=2.56.0.rc2
  ---
  > GIT_VERSION=2.56.0.rc2.1.g0f8e75abeb
  277d276
  <     CC varint.o
  308a308
  >     CARGO target/release/libgitcore.a
  $ 

The 'out' file is yesterdays build of v2.56.0-rc2. (Similarly, the 'sp-out', 'sc' and 'hcout' files record output for v2.56.0-rc2).

  $ make sparse >sp-out1 2>&1
  $ diff sp-out sp-out1
  268d267
  <     SP varint.c
  $ 

  $ ./static-check.pl >sc1
  $ diff sc sc1
  $ 

  $ make -k hdr-check >hcout1 2>&1
  $ diff hcout hcout1
  $ 

  $ . ../git-test-setup
  $ env | grep TEST
  TEST_NO_MALLOC_CHECK=yes
  GIT_TEST_CHAIN_LINT=0
  $ make test >test-out-2-56-rc2-1 2>&1
  $ tail -n 13 test-out-2-56-rc2-1
  Test Summary Report
  -------------------
  unit-tests/bin/unit-tests.exe                    (Wstat: 256 (exited 1) Tests: 261 Failed: 1)
    Failed test:  254
    Non-zero exit status: 1
  t9904-url-parse.sh                               (Wstat: 256 (exited 1) Tests: 53 Failed: 4)
    Failed tests:  39, 42-43, 47
    Non-zero exit status: 1
  Files=1060, Tests=33592, 3519 wallclock secs (42.61 usr 141.30 sys + 9056.74 cusr 12680.95 csys = 21921.60 CPU)
  Result: FAIL
  make[1]: *** [Makefile:82: prove] Error 1
  make[1]: Leaving directory '/home/ramsay/git/t'
  make: *** [Makefile:3424: test] Error 2
  $

Despite the failure, this shows exactly the same failures as v2.56.0-rc2.

So, this doesn't stress the rust compiler very much, but I guess it is slightly encouraging! I suppose Brian has plenty of rust code in a branch somewhere that could be tested ...

Unfortunately, I am just about (in a few hours) to go into hospital for a surgical procedure, so I will be AWOL for some time yet, ...

ATB, Ramsay Jones

[1] https://sourceware.org/pipermail/cygwin-announce/2026-January/012823.html

Thanks Johannes for the detailed summary,

See below some additions on the storage backend in JGit below.

I'd like to provide some context here, as several pluggable JGit storage implementations are actually running in highly demanding production environments today.

First, the Git servers at Google (including gerrit.googlesource.com) are powered by a custom, pluggable JGit storage backend built on top of Google's internal data-center storage. While the backend implementation itself is proprietary, it relies entirely on the exact same JGit classes and pluggable interfaces available in the open-source project. It has been operating smoothly at a massive scale for over a decade.

Second, there is a highly successful open-source implementation focusing on the ref-db layer of Git storage. Developed by GerritForge over six years ago and donated to the AOSP project (see the GitHub mirror at [3]), this backend is used daily in several large, multi-site Gerrit deployments worldwide. It allows for globally consistent updates of Git repository refs across distributed remote sites.

Third, regarding the implementation mentioned by Dani [2]: while it captured attention 10 years ago, without seeing immediate adoption, I suspect the timing simply preceded the current appetite for AI-generated code. The demand is entirely different now. The architectural foundation Google built into JGit over 16 years ago is perfectly positioned to support these kinds of modern innovations.

It is genuinely exciting to see the core Git project moving in this same direction with Patrick's work. Pluggability has always been a core strength of JGit's design, and seeing the broader ecosystem embrace this abstraction is a huge win for all of us.

[3] (github.com/GerritCodeReview/modules_global-refdb)

Luca