Skip to contentSkip to conversation

What's cooking in git.git (Jan 2025, #05; Fri, 17)

PATCH New Junio C Hamano 7 participants 23 messages Jan 18, 2025 – Jan 24, 2025 Documentation

lore
Share & tools
lore
#0 lore raw

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 are candidate to be in a future release). Commits prefixed with '-' are only in 'seen', and aren't considered "accepted" at all and may be annotated with an URL to a message that raises issues but they are no means exhaustive. A topic without enough support may be discarded after a long period of no activity (of course they can be resubmit when new interests arise).

There are quite a few topics that are listed here but without much review activities. I'll review the notes below with list archive myself to see which ones are truly stale and discard them, maybe later next week.

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']

* as/long-option-help-i18n (2024-12-30) 1 commit
(merged to 'next' on 2024-12-30 at 900c79808f)
+ parse-options: localize mark-up of placeholder text in the short help

Tweak the help text used for the option value placeholders by parse-options API so that translations can customize the "<>" placeholder signal (e.g. "--option=<value>").
source: <[email protected]>

* mb/t7110-use-test-path-helper (2025-01-03) 1 commit (merged to 'next' on 2025-01-06 at cd96b0ac82)
+ t7110: replace test -f with test_path_is_* helpers

Test modernization.
source: <[email protected]>

* ps/meson-weak-sha1-build (2024-12-30) 8 commits
(merged to 'next' on 2025-01-01 at e01db872e4)
+ meson: provide a summary of configured backends
+ meson: wire up unsafe SHA1 backend
+ meson: add missing dots for build options
+ meson: simplify conditions for HTTPS and SHA1 dependencies
+ meson: require SecurityFramework when it's used as SHA1 backend
+ meson: deduplicate access to SHA1/SHA256 backend options
+ meson: consistenlty spell 'CommonCrypto'
+ Merge branch 'ps/weak-sha1-for-tail-sum-fix' into ps/meson-weak-sha1-build (this branch is used by ps/build-meson-fixes and ps/zlib-ng.)

meson-based build now supports the unsafe-sha1 build knob.
source: <[email protected]>

* ps/more-sign-compare (2024-12-27) 10 commits
(merged to 'next' on 2025-01-01 at 41c78cf690)
+ sign-compare: avoid comparing ptrdiff with an int/unsigned
+ commit-reach: use size_t to track indices when computing merge bases
+ shallow: fix -Wsign-compare warnings
+ builtin/log: fix remaining -Wsign-compare warnings
+ builtin/log: use size_t to track indices
+ commit-reach: use size_t to track indices in get_reachable_subset()
+ commit-reach: use size_t to track indices in remove_redundant()
+ commit-reach: fix type of min_commit_date
+ commit-reach: fix index used to loop through unsigned integer
+ prio-queue: fix type of insertion_ctr

More -Wsign-compare fixes.
cf. https://staticthinking.wordpress.com/2023/07/25/wsign-compare-is-garbage/
source: <20241227-b4-pks-commit-reach-sign-compare-v1-0-07c59c2aa632@pks.im>

* ps/object-collision-check (2025-01-06) 4 commits
(merged to 'next' on 2025-01-06 at 540e2bae11)
+ object-file: retry linking file into place when occluding file vanishes
+ object-file: don't special-case missing source file in collision check
+ object-file: rename variables in check_collision() (merged to 'next' on 2024-12-30 at e083ea3154)
+ object-file: fix race in object collision check

CI jobs gave sporadic failures, which turns out that that the
object finalization code was giving an error when it did not have
to.
source: <20250106-b4-pks-object-file-racy-collision-check-v2-0-8b3984ecbb18@pks.im>

* re/submodule-parse-opt (2024-12-11) 7 commits (merged to 'next' on 2024-12-21 at 9e65a56a63)
+ git-submodule.sh: rename some variables
+ git-submodule.sh: improve variables readability
+ git-submodule.sh: add some comments
+ git-submodule.sh: get rid of unused variable
+ git-submodule.sh: get rid of isnumber
+ git-submodule.sh: improve parsing of short options
+ git-submodule.sh: improve parsing of some long options

"git submodule" learned various ways to spell the same option, e.g. "--branch=B" can be spelled "--branch B" or "-bB".
source: <[email protected]>

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

* sj/meson-doc-technical-dependency-fix (2025-01-14) 1 commit (merged to 'next' on 2025-01-16 at 3ec55e0703)
+ meson: fix missing deps for technical articles

The meson build procedure for Documentation/technical/ hiearchy was missing necessary dependencies, which has been corrected.

Will merge to 'master'.
source: <5114dc9a00377826a55f6bab007d2ad1a4de8bc5.1736866030.git.sam@gentoo.org>

* tc/meson-use-our-version-def-h (2025-01-14) 1 commit (merged to 'next' on 2025-01-16 at 76e9e81736)
+ meson: ensure correct version-def.h is used

The meson build procedure looked for the 'version-def.h' file in a wrong directory, which has been corrected.

Will merge to 'master'.
source: <[email protected]>

* js/libgit-rust (2025-01-16) 6 commits
- fixup! common-main: split init and exit code into new files
- Makefile: add option to build and test libgit-rs and libgit-rs-sys
- libgit: add higher-level libgit crate
- libgit-sys: also export some config_set functions
- libgit-sys: introduce Rust wrapper for libgit.a
- common-main: split init and exit code into new files

Foreign language interface for Rust into our code base has been added.

Needs to be aware of meson build?
cf. <[email protected]>
source: <[email protected]>

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)

"git refs migrate" for migrating reflog data was broken.

Will merge to 'master'.
cf. <[email protected]>
source: <CAOLa=ZTL9n_DPhNr49XAd6bT838kc09oVx_AH7Pb4o8VK_xQ9w@mail.gmail.com>

* jc/show-usage-help (2025-01-17) 6 commits
- builtin: send usage() help text to standard output
- oddballs: send usage() help text to standard output
- builtins: send usage_with_options() help text to standard output
- usage: add show_usage_if_asked()
- parse-options: add show_usage_with_options_if_asked()
- t0012: optionally check that "-h" output goes to stdout

The help text from "git $cmd -h" appear on the standard output for some $cmd and the standard error for others. The built-in commands have been fixed to show them on the standard output consistently.

Will merge to 'next'.
cf. <[email protected]>

* kn/pack-write-with-reduced-globals (2025-01-17) 5 commits
- pack-write: pass hash_algo to internal functions
- pack-write: pass hash_algo to write_rev_file()
- pack-write: pass hash_algo to write_idx_file()
- pack-write: pass repository to index_pack_lockfile()
- pack-write: pass hash_algo to fixup_pack_header_footer()

Code clean-up.

Well merge to 'next'?
cf. <[email protected]>
source: <[email protected]>

* ps/reftable-sign-compare (2025-01-16) 10 commits
- reftable: address trivial -Wsign-compare warnings
- reftable/blocksource: adjust read_block() to return ssize_t
- reftable/blocksource: adjust type of the block length
- reftable/block: adjust type of the restart length
- reftable/block: adapt header and footer size to return a size_t
- reftable/basics: adjust hash_size() to return uint32_t
- reftable/basics: adjust common_prefix_size() to return size_t
- reftable/record: handle overflows when decoding varints
- reftable/record: drop unused print function pointer
- meson: stop disabling -Wsign-compare

THe reftable/ library code has been made -Wsign-compare clean.

Will merge to 'next'?
source: <[email protected]>

* sk/unit-tests (2025-01-17) 4 commits
- t/unit-tests: convert reftable tree test to use clar test framework
- t/unit-tests: adapt priority queue test to use clar test framework
- t/unit-tests: convert mem-pool test to use clar test framework
- t/unit-tests: handle dashes in test suite filenames

Move a few more unit tests to the clar test framework.

Will merge to 'next'.
source: <[email protected]>

* zh/gc-expire-to (2025-01-16) 1 commit
- gc: add --expire-to option

"git gc" learned the "--expire-to" option and passes it down to underlying "git repack".

Needs review.
source: <[email protected]>

* jc/cli-doc-option-and-config (2025-01-17) 1 commit (merged to 'next' on 2025-01-17 at 71f41b00d8)
+ gitcli: document that command line trumps config and env

Doc update.

Will merge to 'master'.
source: <[email protected]>

* jk/pack-header-parse-alignment-fix (2025-01-17) 3 commits
- index-pack, unpack-objects: use skip_prefix to avoid magic number
- parse_pack_header_option(): avoid unaligned memory writes
- packfile: factor out --pack_header argument parsing

It was possible for "git unpack-objects" and "git index-pack" to make an unaligned access, which has been corrected.

Will merge to 'next'.
source: <[email protected]>

* kn/reflog-migration-fix-followup (2025-01-17) 4 commits
- reftable: prevent 'update_index' changes after header write
- refs: use 'uint64_t' for 'ref_update.index'
- refs: mark ref_transaction_update_reflog() as static
- Merge branch 'kn/reflog-migration-fix' into kn/reflog-migration-fix-followup (this branch uses kn/reflog-migration-fix.)

Code clean-up.

source: <20250117-461-corrupted-reftable-followup-v1-0-70ee605ae3fe@gmail.com>

* mh/connect-sign-compare (2025-01-17) 1 commit
- connect: address -Wsign-compare warnings

The code in connect.c has been updated to work around complaints from -Wsign-compare.

Will merge to 'next'.
source: <[email protected]>

* ps/build-meson-subtree (2025-01-17) 3 commits
- meson: wire up the git-subtree(1) command
- meson: introduce build option for contrib
- contrib/subtree: fix building docs

THe meson-driven build is now aware of "git-subtree" housed in contrib/subtree hierarchy.

Will merge to 'next'.
source: <[email protected]>

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

* ak/instaweb-python-port-binding-fix (2025-01-10) 1 commit (merged to 'next' on 2025-01-17 at bcb5e21e0b)
+ instaweb: fix ip binding for the python http.server

The "instaweb" bound only to local IP address without "--local" and to all addresses with "--local", which was the other way around, when using Python's http.server class, which has been corrected.

Will merge to 'master'.
source: <[email protected]>

* bf/fetch-set-head-fix (2025-01-13) 1 commit
- fetch set_head: fix non-mirror remotes in bare repositories

Fetching into a bare repository incorrectly assumed it always used a mirror layout when deciding to update remote-tracking HEAD, which has been corrected.

Needs review.
source: <[email protected]>

* mh/doc-credential-helpers-with-pat (2025-01-10) 2 commits (merged to 'next' on 2025-01-17 at a70beabaf5)
+ docs: discuss caching personal access tokens
+ docs: list popular credential helpers

Document that it is insecure to use Personal Access Tokens, which some hosting providers take as username/password, embedded in URLs.

Will merge to 'master'.
source: <[email protected]>

* ps/build-meson-fixes (2025-01-14) 12 commits
- ci: wire up Visual Studio build with Meson
- ci: raise error when Meson generates warnings
- meson: fix compilation with Visual Studio
- meson: make the CSPRNG backend configurable
- meson: wire up fuzzers
- meson: wire up generation of distribution archive
- meson: wire up development environments
- meson: fix dependencies for generated headers
- meson: populate project version via GIT-VERSION-GEN
- GIT-VERSION-GEN: allow running without input and output files
- GIT-VERSION-GEN: simplify computing the dirty marker
- Merge branch 'ps/meson-weak-sha1-build' into ps/build-meson-fixes (this branch is used by ps/zlib-ng.)

More build fixes and enhancements on meson based build procedure.

Needs review.
source: <[email protected]>

* ps/zlib-ng (2025-01-16) 12 commits
- ci: make "linux-musl" job use zlib-ng
- ci: switch linux-musl to use Meson
- compat/zlib: allow use of zlib-ng as backend
- git-zlib: cast away potential constness of next_in pointer
- compat/zlib: provide stubs for deflateSetHeader()
- compat/zlib: provide deflateBound() shim centrally
- git-compat-util: move include of "compat/zlib.h" into "git-zlib.h"
- compat: introduce new "zlib.h" header
- git-compat-util: drop z_const define
- compat: drop uncompress2() compatibility shim
- Merge branch 'ps/build-meson-fixes' into ps/zlib-ng
- Merge branch 'ps/meson-weak-sha1-build' into ps/zlib-ng (this branch uses ps/build-meson-fixes.)

The code paths to interact with zlib has been cleaned up in preparation for building with zlib-ng.

Needs review.
source: <[email protected]>

* rs/ref-fitler-used-atoms-value-fix (2025-01-13) 1 commit
- ref-filter: share bases and is_base_tips between formatting and sorting

"git branch --sort=..." and "git for-each-ref --format=... --sort=..." did not work as expected with some atoms, which has been corrected.

cf. https://lore.kernel.org/git/[email protected]/
source: <[email protected]>

* tb/unsafe-hash-cleanup (2025-01-17) 8 commits
- hash.h: drop unsafe_ function variants
- csum-file: introduce hashfile_checkpoint_init()
- t/helper/test-hash.c: use unsafe_hash_algo()
- csum-file.c: use unsafe_hash_algo()
- hash.h: introduce unsafe_hash_algo()
- csum-file.c: extract algop from hashfile_checksum_valid()
- csum-file: store the hash algorithm as a struct field
- t/helper/test-tool: implement sha1-unsafe helper

The API around choosing to use unsafe variant of SHA-1 implementation has been updated in an attempt to make it harder to abuse.

Will merge to 'next'?
source: <[email protected]>

* dk/zsh-config-completion-fix (2025-01-06) 1 commit (merged to 'next' on 2025-01-10 at efba7d534c)
+ completion: repair config completion for Zsh

Completion script updates for zsh

Will merge to 'master'.
source: <[email protected]>

* en/object-name-with-funny-refname-fix (2025-01-13) 2 commits (merged to 'next' on 2025-01-16 at 89cd7778c9)
+ object-name: be more strict in parsing describe-like output
+ object-name: fix resolution of object names containing curly braces

Extended SHA-1 expression parser did not work well when a branch with an unusual name (e.g. "foo{bar") is involved.

Will merge to 'master'.
source: <[email protected]>

* sj/ref-consistency-checks-more (2025-01-06) 10 commits
- builtin/fsck: add git refs verify child process
- packed-backend: check whether the "packed-refs" is sorted
- packed-backend: add check for object consistency
- packed-backend: create "fsck_packed_ref_entry" to store parsing info
- packed-backend: add "packed-refs" entry consistency check
- packed-backend: check whether the refname contains NULL binaries
- packed-backend: add "packed-refs" header consistency check
- packed-backend: check whether the "packed-refs" is regular
- builtin/refs.h: get worktrees without reading head info
- files-backend: add object check for regular ref

"git fsck" becomes more careful when checking the refs.
source: <Z3qNUizvHJLgMx1y@ArchLinux>

* jk/lsan-race-ignore-false-positive (2025-01-07) 3 commits (merged to 'next' on 2025-01-09 at 3d7cd910b5)
+ test-lib: add a few comments to LSan log checking
+ test-lib: simplify lsan results check
+ test-lib: invert return value of check_test_results_san_file_empty

The code to check LSan results has been simplified and made more robust.

Will merge to 'master'.
source: <[email protected]>

* jk/t7407-use-test-grep (2025-01-07) 1 commit
(merged to 'next' on 2025-01-09 at 1d584ee42d)
+ t7407: use test_grep

Test clean-up.

Will merge to 'master'.
source: <[email protected]>

* jt/fsck-skiplist-parse-fix (2025-01-07) 1 commit (merged to 'next' on 2025-01-09 at d08b30fd78)
+ fsck: reject misconfigured fsck.skipList

A misconfigured "fsck.skiplist" configuration variable was not diagnosed as an error, which has been corrected.

Will merge to 'master'.
source: <[email protected]>

* ps/reftable-get-random-fix (2025-01-07) 2 commits (merged to 'next' on 2025-01-09 at bc024b7a45)
+ reftable/stack: accept insecure random bytes
+ wrapper: allow generating insecure random bytes

The code to compute "unique" name used git_rand() which can fail or get stuck; the callsite does not require cryptographic security. Introduce the "insecure" mode and use it appropriately.

Will merge to 'master'.
source: <[email protected]>

* mh/credential-cache-authtype-request-fix (2025-01-09) 1 commit
- credential-cache: respect authtype capability

The "cache" credential back-end did not handle authtype correctly, which has been corrected.
source: <[email protected]>

* mh/gitattr-doc-markup-fix (2025-01-07) 1 commit (merged to 'next' on 2025-01-10 at 9b8f84ebe2)
+ docs: fix typesetting of merge driver placeholders

Doc markup fix.

Will merge to 'master'.
source: <20250107212421.7yyvuzw4uqxnqv7t@archP14s>

* sk/unit-test-hash (2025-01-09) 1 commit (merged to 'next' on 2025-01-13 at 865b121824)
+ t/unit-tests: convert hash to use clar test framework

Test update.

Will merge to 'master'.
source: <[email protected]>

* aj/difftool-config-doc-fix (2025-01-09) 1 commit (merged to 'next' on 2025-01-10 at b8902a53d1)
+ difftool docs: restore correct position of tool list

Docfix.

Will merge to 'master'.
source: <[email protected]>

* jk/combine-diff-cleanup (2025-01-09) 14 commits
- tree-diff: make list tail-passing more explicit
- tree-diff: simplify emit_path() list management
- tree-diff: use the name "tail" to refer to list tail
- tree-diff: drop list-tail argument to diff_tree_paths()
- combine-diff: drop public declaration of combine_diff_path_size()
- tree-diff: inline path_appendnew()
- tree-diff: pass whole path string to path_appendnew()
- tree-diff: drop path_appendnew() alloc optimization
- run_diff_files(): de-mystify the size of combine_diff_path struct
- diff: add a comment about combine_diff_path.parent.path
- combine-diff: use pointer for parent paths
- tree-diff: clear parent array in path_appendnew()
- combine-diff: add combine_diff_path_new()
- run_diff_files(): delay allocation of combine_diff_path

Code clean-up for code paths around combined diff.
source: <[email protected]>

* sc/help-autocorrect-one (2025-01-13) 1 commit
- help: interpret boolean string values for help.autocorrect

"[help] autocorrect = 1" used to be a way to say "please wait for 0.1 second after suggesting a typofix of the command name before running that command"; now it means "yes, if there is a plausible typofix for the command name, please run it immediately".

Looking good except for "should 0 and false be 'tell it without doing it'?".
source: <[email protected]>

* ja/doc-notes-markup-updates (2025-01-10) 1 commit
- doc: convert git-notes to new documentation format

Doc mark-up updates.
source: <[email protected]>

* ja/doc-restore-markup-update (2025-01-10) 1 commit
- doc: convert git-restore to new style format

Doc mark-up updates.
source: <[email protected]>

* ua/os-version-capability (2025-01-17) 6 commits
. version: introduce osversion.command config for os-version output . connect: advertise OS version
. t5701: add setup test to remove side-effect dependency . version: extend get_uname_info() to hide system details . version: refactor get_uname_info()
. version: refactor redact_non_printables()

The value of "uname -s" is by default sent over the wire as a new capability, with an opt-out for privacy-concious folks.

source: <[email protected]>

* ja/doc-commit-markup-updates (2025-01-15) 5 commits
- doc: migrate git-commit manpage secondary files to new format
- doc: convert git commit config to new format
- doc: make more direct explanations in git commit options
- doc: the mode param of -u of git commit is optional
- doc: apply new documentation guidelines to git commit

Doc updates.

Will merge to 'next'?
source: <[email protected]>

* ps/ci-misc-updates (2025-01-10) 10 commits
- ci: remove stale code for Azure Pipelines
- ci: use latest Ubuntu release
- ci: stop special-casing for Ubuntu 16.04
- gitlab-ci: add linux32 job testing against i386
- gitlab-ci: remove the "linux-old" job
- github: simplify computation of the job's distro
- github: convert all Linux jobs to be containerized
- github: adapt containerized jobs to be rootless
- t7422: fix flaky test caused by buffered stdout
- t0060: fix EBUSY in MinGW when setting up runtime prefix

CI updates (containerization, dropping stale ones, etc.).
source: <[email protected]>

* sk/strlen-returns-size_t (2024-12-26) 1 commit
- date.c: Fix type missmatch warings from msvc

Code clean-up.

The remainder needs to be reviewed.
source: <[email protected]>

* sk/maintenance-remote-prune (2025-01-03) 1 commit
- maintenance: add prune-remote-refs task

A new periodic maintenance task to run "git remote prune" has been introduced.

Expecting a reroll.
source: <[email protected]>

* jc/show-index-h-update (2024-12-20) 1 commit
- show-index: the short help should say the command reads from its input

Doc and short-help text for "show-index" has been clarified to stress that the command reads its data from the standard input.

Comments?
source: <[email protected]>

* ps/the-repository (2024-12-18) 15 commits
(merged to 'next' on 2025-01-09 at 1de40edade)
+ match-trees: stop using the_repository
+ graph: stop using the_repository
+ add-interactive: stop using the_repository
+ tmp-objdir: stop using the_repository
+ resolve-undo: stop using the_repository
+ credential: stop using the_repository
+ mailinfo: stop using the_repository
+ diagnose: stop using the_repository
+ server-info: stop using the_repository
+ send-pack: stop using the_repository
+ serve: stop using the_repository
+ trace: stop using the_repository
+ pager: stop using the_repository
+ progress: stop using the_repository
+ Merge branch 'ps/build-sign-compare' into ps/the-repository

More code paths have a repository passed through the callchain, instead of assuming the primary the_repository object.

Will merge to 'master'.
source: <20241217-pks-use-the-repository-conversion-v1-0-0dba48bcc239@pks.im>

* jc/doc-attr-tree (2024-12-14) 1 commit
- doc: give attr.tree a bit more visibility

Make sure that "git --attr-source=X", GIT_ATTR_SOURCE, and attr.tree configuration variables appear at the same places in the documentation.

On hold.
cf. <[email protected]>
source: <[email protected]>

* ps/3.0-remote-deprecation (2025-01-06) 6 commits
- remote: announce removal of "branches/" and "remotes/"
- builtin/pack-redundant: remove subcommand with breaking changes
- ci: repurpose "linux-gcc" job for deprecations
- ci: merge linux-gcc-default into linux-gcc
- Makefile: wire up build option for deprecated features
- Merge branch 'ps/build' into ps/3.0-remote-deprecation

Following the procedure we established to introduce breaking changes for Git 3.0, allow an early opt-in for removing support of $GIT_DIR/branches/ and $GIT_DIR/remotes/ directories to configure remotes.
source: <20250106-pks-remote-branches-deprecation-v2-0-2ce87c053536@pks.im>

* cc/lop-remote (2024-12-07) 5 commits
. doc: add technical design doc for large object promisors . promisor-remote: check advertised name or URL
. Add 'promisor-remote' capability to protocol v2 . strbuf: refactor strbuf_trim_trailing_ch()
. version: refactor strbuf_sanitize()

Expecting a reroll.
cf. <CAP8UFD3bdEo1_bg+aX52xSGxmg9KfNrpiX+2LwUM-yDqjvfZbQ@mail.gmail.com>
source: <[email protected]>

* ds/backfill (2024-12-20) 6 commits
- backfill: assume --sparse when sparse-checkout is enabled
- backfill: add --sparse option
- backfill: add --min-batch-size=<n> option
- backfill: basic functionality and tests
- backfill: add builtin boilerplate
- Merge branch 'ds/path-walk-1' into ds/backfill (this branch uses ds/path-walk-1.)

Lazy-loading missing files in a blobless clone on demand is costly as it tends to be one-blob-at-a-time. "git backfill" is introduced to help bulk-download necessary files beforehand.

Expecting a reroll.
cf. <[email protected]>
source: <[email protected]>

* tb/incremental-midx-part-2 (2024-11-20) 15 commits
- midx: implement writing incremental MIDX bitmaps
- pack-bitmap.c: use ewah_or_iterator for type bitmap iterators
- pack-bitmap.c: keep track of each layer's type bitmaps
- ewah: implement struct ewah_or_iterator
- pack-bitmap.c: apply pseudo-merge commits with incremental MIDXs
- pack-bitmap.c: compute disk-usage with incremental MIDXs
- pack-bitmap.c: teach rev-list --test-bitmap about incremental MIDXs
- pack-bitmap.c: support bitmap pack-reuse with incremental MIDXs
- pack-bitmap.c: teach show_objects_for_type() about incremental MIDXs
- pack-bitmap.c: teach bitmap_for_commit() about incremental MIDXs
- pack-bitmap.c: open and store incremental bitmap layers
- pack-revindex: prepare for incremental MIDX bitmaps
- Documentation: describe incremental MIDX bitmaps
- Merge branch 'tb/pseudo-merge-bitmap-fixes' into tb/incremental-midx-part-2
- Merge branch 'tb/incremental-midx-part-1' into tb/incremental-midx-part-2

Incrementally updating multi-pack index files.

Needs review.
source: <[email protected]>

* ps/send-pack-unhide-error-in-atomic-push (2024-11-14) 2 commits
- transport: don't ignore git-receive-pack(1) exit code on atomic push
- t5504: modernize test by moving heredocs into test bodies

"git push --atomic --porcelain" used to ignore failures from the other side, losing the error status from the child process, which has been corrected.

Needs to see if competing parallel topic needs to replace this one.
source: <20241113-pks-push-atomic-respect-exit-code-v1-0-7965f01e7f4e@pks.im>

* ds/name-hash-tweaks (2024-12-20) 8 commits
- pack-objects: add third name hash version
- pack-objects: prevent name hash version change
- test-tool: add helper for name-hash values
- p5313: add size comparison test
- pack-objects: add GIT_TEST_NAME_HASH_VERSION
- repack: add --name-hash-version option
- pack-objects: add --name-hash-version option
- pack-objects: create new name-hash function version

"git pack-objects" and its wrapper "git repack" learned an option to use an alternative path-hash function to improve delta-base selection to produce a packfile with deeper history than window size.

Comments?
source: <[email protected]>

* ds/path-walk-1 (2024-12-20) 7 commits
- path-walk: reorder object visits
- path-walk: mark trees and blobs as UNINTERESTING
- path-walk: visit tags and cached objects
- path-walk: allow consumer to specify object types
- t6601: add helper for testing path-walk API
- test-lib-functions: add test_cmp_sorted
- path-walk: introduce an object walk by path (this branch is used by ds/backfill.)

Introduce a new API to visit objects in batches based on a common path, or by type.

Will merge to 'next'?
cf. <[email protected]>
source: <[email protected]>

* ej/cat-file-remote-object-info (2025-01-14) 8 commits
- cat-file: add remote-object-info to batch-command
- transport: add client support for object-info
- serve: advertise object-info feature
- fetch-pack: move fetch initialization
- fetch-pack: refactor packet writing
- t1006: split test utility functions into new "lib-cat-file.sh"
- cat-file: add declaration of variable i inside its for loop
- git-compat-util: add strtoul_ul() with error handling

"git cat-file --batch" and friends can optionally ask a remote server about objects it does not have.

Comments?
source: <[email protected]>

* jc/move-is-bare-repository-cfg-variable-to-repo (2024-11-07) 3 commits . repository: BUG when is_bare_cfg is not initialized
. setup: initialize is_bare_cfg
. git: remove is_bare_repository_cfg global variable

Code rewrite to turn the is_bare_repository_cfg global variable into a member in the the_repo singleton repository object.

Will discard.
Has been in "Waiting for response to reviews" state for too long. cf. <[email protected]>
Seems to break t0021-conversion on Windows.
cf. https://lore.kernel.org/git/[email protected]/
source: <[email protected]>

#1 lore raw
Quoted 10 lines · Junio C Hamano

* jk/pack-header-parse-alignment-fix (2025-01-17) 3 commits
- index-pack, unpack-objects: use skip_prefix to avoid magic number
- parse_pack_header_option(): avoid unaligned memory writes
- packfile: factor out --pack_header argument parsing

It was possible for "git unpack-objects" and "git index-pack" to make an unaligned access, which has been corrected.

Will merge to 'next'.
source: <[email protected]>

I was planning to re-roll this with your sparse fix included, and adding another patch to do get_be32() on the reading side. So maybe hold off for a moment.

(I'd also be interested in any comments on the "maybe we should just align these buffers" approach; I'm undecided on it).

-Peff

#2 lore raw
Quoted 10 lines · Jeff King
Quoted 4 lines · Junio C Hamano

* jk/pack-header-parse-alignment-fix (2025-01-17) 3 commits
...
Will merge to 'next'.
source: <[email protected]>

I was planning to re-roll this with your sparse fix included, and adding another patch to do get_be32() on the reading side. So maybe hold off for a moment.

Thanks.

Quoted 2 lines

(I'd also be interested in any comments on the "maybe we should just align these buffers" approach; I'm undecided on it).

Unless we have the buffer _inside_ the helper function that may perform the possibly-unaligned access, I am not sure how it helps.

I guess that we can align buffers used by two existing callers, document that the helper function takes an aligned buffer and that it is a fault of the caller if somebody passes an unaligned buffer, but I am not sure if that is where we want to go.

#3 lore raw
Quoted 5 lines · Junio C Hamano
Quoted 2 lines

(I'd also be interested in any comments on the "maybe we should just align these buffers" approach; I'm undecided on it).

Unless we have the buffer _inside_ the helper function that may perform the possibly-unaligned access, I am not sure how it helps.

We sort-of do. The offending code is all static local to unpack-objects.c, and always operates on the same buffer (directly for writing, and for reading through the static fill() macro which returns it directly). And likewise in index-pack.c.

I think these two are oddballs in that they read parts of a pack into a buffer. Whereas all of the more generic pack code will mmap() it, and presumably that ends up with suitable alignment. I guess platforms with NO_MMAP would read into a malloc'd buffer, but that should likewise be prepared for any alignment. (I suppose another way of achieving alignment would simply be to turn "buffer" into a pointer and malloc it at the program start, but that still leaves the need to fix sizeof() calls).

Quoted 4 lines

I guess that we can align buffers used by two existing callers, document that the helper function takes an aligned buffer and that it is a fault of the caller if somebody passes an unaligned buffer, but I am not sure if that is where we want to go.

The functions themselves aren't really reusable, so any new code which wants to do the same thing would end up rewriting it and potentially creating the same problem. But that's probably an argument for switching away from the cast and to put/get_be32(). It provides a more obviously better example for people to copy from.

I'll post a re-roll in a bit.

-Peff

#4 lore raw
Quoted 7 lines · Jeff King
Quoted 2 lines

Unless we have the buffer _inside_ the helper function that may perform the possibly-unaligned access, I am not sure how it helps.

We sort-of do. The offending code is all static local to unpack-objects.c, and always operates on the same buffer (directly for writing, and for reading through the static fill() macro which returns it directly). And likewise in index-pack.c.

Oh, reading this again, I guess you were thinking of the helper that I had factored out. Yes, if that requires aligned memory that is an awful interface. :-/

I was thinking to just leave the offending code untouched in the individual commands if we went this route.

But anyway, I'll prepare a version going the other get_be32() direction.

-Peff

#5 lore raw
Quoted 2 lines · Jeff King

I was thinking to just leave the offending code untouched in the individual commands if we went this route.

Ah, I see. Yeah, then we have a subtle and possibly brittle code paths that are well contained inside two functions.

Quoted 1 line

But anyway, I'll prepare a version going the other get_be32() direction.

Thanks. That probably results in better code with fewer magic.

THanks.

#6 lore raw
In reply to Junio C Hamano
Quoted 10 lines · Junio C Hamano

* sc/help-autocorrect-one (2025-01-13) 1 commit
- help: interpret boolean string values for help.autocorrect

"[help] autocorrect = 1" used to be a way to say "please wait for 0.1 second after suggesting a typofix of the command name before running that command"; now it means "yes, if there is a plausible typofix for the command name, please run it immediately".

Looking good except for "should 0 and false be 'tell it without doing it'?".
source: <[email protected]>

(The text below is from the original thread; sorry I don't have it handy so I just replied here instead)

Quoted 8 lines

... but would it be simpler if we made it an extended boolean, i.e.

    true, yes, on, 1  -> same as "immediate"
    false, no, off, 0 -> same as "never"
    immediate         -> same as what we currently do
    never             -> same as what we currently do
    prompt            -> same as what we currently do
    number            -> same as what we currently do

I do think that, "0 -> same as never," makes a lot of sense from a usability perspective.

On the other hand, I don't think that, "1 -> same as immediate," is a very safe thing to do. One reason is that we should try to do the least surprising thing possible, especially when the command may be something that the user did not intend to run.

Recall that this topic was spun up because a value of "1" was interpreted as 0.1 seconds, which is effectively the same as "immediate." The original interpretation is arguably a usability issue that is not improved by these changes.

I would instead recommend that, "1 -> same as prompt," would be a safer and less surprising behavior. If the user wants "immediate" they can be explicit about it. "immediate" is the most dangerous of all of these options so adding ambiguous routes to it seems like a step backwards.

I don't really think backwards-compatibility is much of a concern here at all. It *would* be a concern if we were moving from a safe behavior to a less-safe behavior (like this patch currently does) but not so in the other direction like I'm proposing by making "1" mean "prompt".

Do we think this is a valid assessment and worth changing? If so I can try to whip up a quick patch over the existing state of "seen" to change the behavior to "prompt" when "1" is seen and "never" when "0" is seen.

Signature · 1 line
David
#7 lore raw

PATCH 1/1 help: make help.autocorrect = 1 the same as "prompt"

Choose the least surprising behavior by prompting the user when when help.autocorrect is set to a boolean "true" value.

Make "0" and other "false" boolean values the same as "never".

Make "help.autocorrect = show" the same as the default behavior so that users can override it in a repository-local .git/config.

---
Here's what the proposed changes to make "true" = "prompt" might look like.

 Documentation/config/help.txt | 26 ++++++++++++++++--------
 help.c                        | 14 +++++++++----
 t/t9003-help-autocorrect.sh   | 38 +++++++++++++++++------------------
 3 files changed, 47 insertions(+), 31 deletions(-)
Documentation/config/help.txt +18 -8
index a4c6079af8..344b4b13f7 100644
@@ -11,14 +11,24 @@ help.autoCorrect::
 	If git detects typos and can identify exactly one valid command similar
 	to the error, git will try to suggest the correct command or even
 	run the suggestion automatically. Possible config values are:
-	 - 0: show the suggested command (default).
-	 - 1, "true", "on", "yes", "immediate": run the suggested command
-immediately.
-	 - positive number > 1: run the suggested command after specified
-deciseconds (0.1 sec).
-	 - "false", "off", "no", "never": don't run or show any suggested command.
-	 - "prompt": show the suggestion and prompt for confirmation to run
-the command.
++
+--
+show;;
+	Show the suggested command but do not run it. This is the default.
+
+immediate;;
+	Run the suggested command immediately.
+
+never, false, off, no, 0;;
+	Do not run or show the suggested command.
+
+prompt, true, on, yes, 1;;
+	Show the suggestion and prompt for confirmation to run the command.
+
+positive number greater than 1;;
+	Run the suggested command after specified deciseconds (0.1 sec).
+--
++
 
 help.htmlPath::
 	Specify the path where the HTML documentation resides. File system paths
help.c +10 -4
index 7148963e46..447e0c9423 100644
@@ -552,6 +552,7 @@ struct help_unknown_cmd_config {
 	struct cmdnames aliases;
 };
 
+#define AUTOCORRECT_SHOW (-4)
 #define AUTOCORRECT_PROMPT (-3)
 #define AUTOCORRECT_NEVER (-2)
 #define AUTOCORRECT_IMMEDIATELY (-1)
@@ -560,7 +561,7 @@ static int parse_autocorrect(const char *value)
 {
 	switch (git_parse_maybe_bool_text(value)) {
 		case 1:
-			return AUTOCORRECT_IMMEDIATELY;
+			return AUTOCORRECT_PROMPT;
 		case 0:
 			return AUTOCORRECT_NEVER;
 		default: /* other random text */
@@ -573,6 +574,8 @@ static int parse_autocorrect(const char *value)
 		return AUTOCORRECT_NEVER;
 	if (!strcmp(value, "immediate"))
 		return AUTOCORRECT_IMMEDIATELY;
+	if (!strcmp(value, "show"))
+		return AUTOCORRECT_SHOW;
 
 	return 0;
 }
@@ -589,8 +592,10 @@ static int git_unknown_cmd_config(const char *var, const char *value,
 
 		if (!v) {
 			v = git_config_int(var, value, ctx->kvi);
-			if (v < 0 || v == 1)
-				v = AUTOCORRECT_IMMEDIATELY;
+			if (v == 1)
+				v = AUTOCORRECT_PROMPT;
+			else if (v <= 0)
+				v = AUTOCORRECT_NEVER;
 		}
 
 		cfg->autocorrect = v;
@@ -713,7 +718,8 @@ char *help_unknown_cmd(const char *cmd)
 		     n++)
 			; /* still counting */
 	}
-	if (cfg.autocorrect && n == 1 && SIMILAR_ENOUGH(best_similarity)) {
+	if (cfg.autocorrect && cfg.autocorrect != AUTOCORRECT_SHOW &&
+	    n == 1 && SIMILAR_ENOUGH(best_similarity)) {
 		char *assumed = xstrdup(main_cmds.names[0]->name);
 
 		fprintf_ln(stderr,
t/t9003-help-autocorrect.sh +19 -19
index 85a5074b5e..ea20526248 100755
@@ -29,7 +29,7 @@ test_expect_success 'setup' '
 '
 
 test_expect_success 'autocorrect showing candidates' '
-	git config help.autocorrect 0 &&
+	git config help.autocorrect show &&
 
 	test_must_fail git lfg 2>actual &&
 	grep "^	lgf" actual &&
@@ -38,28 +38,28 @@ test_expect_success 'autocorrect showing candidates' '
 	grep "^	distimdistim" actual
 '
 
-for immediate in -1 immediate
-do
-	test_expect_success 'autocorrect running commands' '
-		git config help.autocorrect $immediate &&
+test_expect_success 'autocorrect running commands' '
+	git config help.autocorrect immediate &&
 
-		git lfg >actual &&
-		echo "a single log entry" >expect &&
-		test_cmp expect actual &&
+	git lfg >actual &&
+	echo "a single log entry" >expect &&
+	test_cmp expect actual &&
 
-		git distimdist >actual &&
-		echo "distimdistim was called" >expect &&
-		test_cmp expect actual
-	'
-done
+	git distimdist >actual &&
+	echo "distimdistim was called" >expect &&
+	test_cmp expect actual
+'
 
-test_expect_success 'autocorrect can be declined altogether' '
-	git config help.autocorrect never &&
+for never in -1 0 no false never
+do
+	test_expect_success 'autocorrect can be declined altogether' '
+		git config help.autocorrect $never &&
 
-	test_must_fail git lfg 2>actual &&
-	grep "is not a git command" actual &&
-	test_line_count = 1 actual
-'
+		test_must_fail git lfg 2>actual &&
+		grep "is not a git command" actual &&
+		test_line_count = 1 actual
+	'
+done
 
 test_expect_success 'autocorrect works in work tree created from bare repo' '
 	git clone --bare . bare.git &&
Signature · 1 line
2.48.0.rc2.33.gfa4adeb460
#8 lore raw
In reply to David Aguilar
Quoted 12 lines · David Aguilar

(The text below is from the original thread; sorry I don't have it handy so I just replied here instead)

Quoted 8 lines

... but would it be simpler if we made it an extended boolean, i.e.

    true, yes, on, 1  -> same as "immediate"
    false, no, off, 0 -> same as "never"
    immediate         -> same as what we currently do
    never             -> same as what we currently do
    prompt            -> same as what we currently do
    number            -> same as what we currently do
Quoted 2 lines

I do think that, "0 -> same as never," makes a lot of sense from a usability perspective.

I obviously do not agree. "Suggest the right spelling and let the user decide without time-bomb" is a very useful and safe UI, and the above summary was done by mistake.

Quoted 4 lines

I would instead recommend that, "1 -> same as prompt," would be a safer and less surprising behavior. If the user wants "immediate" they can be explicit about it. "immediate" is the most dangerous of all of these options so adding ambiguous routes to it seems like a step backwards.

Thanks for raising your concern.

As somebody who does *not* use the time-bomb UI that makes me wait when the heuristics guessed correctly and forces me to scramble to hit \C-c when it didn't, I am not qualified to comment in favor of such a huge behaviour change, so I won't, and let others discuss.

Quoted 4 lines

I don't really think backwards-compatibility is much of a concern here at all. It *would* be a concern if we were moving from a safe behavior to a less-safe behavior (like this patch currently does) but not so in the other direction like I'm proposing by making "1" mean "prompt".

#9 lore raw
In reply to Junio C Hamano
Quoted 10 lines · Junio C Hamano

* ds/name-hash-tweaks (2024-12-20) 8 commits
- pack-objects: add third name hash version
- pack-objects: prevent name hash version change
- test-tool: add helper for name-hash values
- p5313: add size comparison test
- pack-objects: add GIT_TEST_NAME_HASH_VERSION
- repack: add --name-hash-version option
- pack-objects: add --name-hash-version option
- pack-objects: create new name-hash function version

Quoted 7 lines

"git pack-objects" and its wrapper "git repack" learned an option to use an alternative path-hash function to improve delta-base selection to produce a packfile with deeper history than window size.

Comments?
source: <[email protected]>

I'll poke the thread, too, but this seems to be the most promising topic in the area of better delta compression. The latest version does not have any comments.

The only decision point I think remains is whether or not to include the last patch (--name-hash-version=3) which I would be happy either way.

Thanks,
-Stolee

#10 lore raw
Quoted 18 lines · Derrick Stolee
Quoted 7 lines · Junio C Hamano

* ds/name-hash-tweaks (2024-12-20) 8 commits
- pack-objects: add third name hash version
- pack-objects: prevent name hash version change
- test-tool: add helper for name-hash values
- p5313: add size comparison test
- pack-objects: add GIT_TEST_NAME_HASH_VERSION
- repack: add --name-hash-version option

Quoted 8 lines

- pack-objects: add --name-hash-version option
- pack-objects: create new name-hash function version "git pack-objects" and its wrapper "git repack" learned an option to use an alternative path-hash function to improve delta-base selection to produce a packfile with deeper history than window size.
Comments?
source: <[email protected]>

Quoted 7 lines

I'll poke the thread, too, but this seems to be the most promising topic in the area of better delta compression. The latest version does not have any comments.

The only decision point I think remains is whether or not to include the last patch (--name-hash-version=3) which I would be happy either way.

I am happy with the updated function that gives us better of both worlds, without losing too much from the "renamed from other directory" while making sure we do not lose too many bits in deeper trees.

Thanks.

#11 lore raw
Quoted 24 lines · Junio C Hamano
Quoted 22 lines · Derrick Stolee
Quoted 7 lines · Junio C Hamano

* ds/name-hash-tweaks (2024-12-20) 8 commits
- pack-objects: add third name hash version
- pack-objects: prevent name hash version change
- test-tool: add helper for name-hash values
- p5313: add size comparison test
- pack-objects: add GIT_TEST_NAME_HASH_VERSION
- repack: add --name-hash-version option

Quoted 8 lines

- pack-objects: add --name-hash-version option
- pack-objects: create new name-hash function version "git pack-objects" and its wrapper "git repack" learned an option to use an alternative path-hash function to improve delta-base selection to produce a packfile with deeper history than window size.
Comments?
source: <[email protected]>

I'll poke the thread, too, but this seems to be the most promising topic in the area of better delta compression. The latest version does not have any comments.

Quoted 8 lines
Quoted 3 lines

The only decision point I think remains is whether or not to include the last patch (--name-hash-version=3) which I would be happy either way.

I am happy with the updated function that gives us better of both worlds, without losing too much from the "renamed from other directory" while making sure we do not lose too many bits in deeper trees.

I had a couple of thoughts that I meant to share before the holiday break, and haven't quite had a chance to get to it now that I'm back at my desk.

Let me try and find some time to respond to the latest round of this series, and apologies for holding it up in the meantime.

Thanks,
Taylor

#12 lore raw
Quoted 12 lines · Taylor Blau
Quoted 10 lines · Junio C Hamano
Quoted 3 lines · Derrick Stolee
Quoted 1 line · Junio C Hamano

* ds/name-hash-tweaks (2024-12-20) 8 commits

...
I am happy with the updated function that gives us better of both worlds, without losing too much from the "renamed from other directory" while making sure we do not lose too many bits in deeper trees.

Quoted 6 lines

I had a couple of thoughts that I meant to share before the holiday break, and haven't quite had a chance to get to it now that I'm back at my desk.

Let me try and find some time to respond to the latest round of this series, and apologies for holding it up in the meantime.

The topic has been stalled for unusually long time, so it won't hurt too much for it to wait for a few more days, but it wouldn't be fair to stall a topic further with just a promise to "try and find time" forever. Let's say we'll go ahead by this weekend unless we hear otherwise?

I am not ultra-happy with the last step, as I personally do not see this different algorithm as "version" (in that people would always want to use version N+1 over version N when both are available) but as "variant" (in that there may be prefer to use variant N over variant N+1 depending on the circumstances), but that may be just the matter of terminology. What's important is to make sure we do not mix two algorhtims up while creating a packfile.

Thanks.

#13 lore raw
Quoted 18 lines · Junio C Hamano
Quoted 16 lines · Taylor Blau
Quoted 10 lines · Junio C Hamano
Quoted 3 lines · Derrick Stolee
Quoted 1 line · Junio C Hamano

* ds/name-hash-tweaks (2024-12-20) 8 commits

...
I am happy with the updated function that gives us better of both worlds, without losing too much from the "renamed from other directory" while making sure we do not lose too many bits in deeper trees.

I had a couple of thoughts that I meant to share before the holiday break, and haven't quite had a chance to get to it now that I'm back at my desk.

Quoted 8 lines
Quoted 2 lines

Let me try and find some time to respond to the latest round of this series, and apologies for holding it up in the meantime.

The topic has been stalled for unusually long time, so it won't hurt too much for it to wait for a few more days, but it wouldn't be fair to stall a topic further with just a promise to "try and find time" forever. Let's say we'll go ahead by this weekend unless we hear otherwise?

I agree, and I apologize for the delay. I prioritized this yesterday and left some review which I think you have seen since sending this email.

Quoted 7 lines

I am not ultra-happy with the last step, as I personally do not see this different algorithm as "version" (in that people would always want to use version N+1 over version N when both are available) but as "variant" (in that there may be prefer to use variant N over variant N+1 depending on the circumstances), but that may be just the matter of terminology. What's important is to make sure we do not mix two algorhtims up while creating a packfile.

Yeah, I think "variant" is probably more accurate, but I don't mind the naming. I think having a unique identifier is important, but I am not convinced that we need to introduce v2 and v3 at the same time. I would rather see us unify behind a single approach to present a clearer/smaller set of options to users.

Thanks,
Taylor

#14 lore raw
Quoted 5 lines · Taylor Blau

Yeah, I think "variant" is probably more accurate, but I don't mind the naming. I think having a unique identifier is important, but I am not convinced that we need to introduce v2 and v3 at the same time. I would rather see us unify behind a single approach to present a clearer/smaller set of options to users.

I agree with you that v2 is superiour most of the time over v1 and v3. If we keep v3, then "version" is an awkward phrasing to use. Some people with specialized needs may use "v3" while most people who do nnot have to use "v1" are better off using "v2" not "v3".

If we were to drop v3, then "version" starts to make sense again, as "v1" is kept primarily for backward compatibility, and those who can afford to follow the latest can "upgrade" to "v2".

Perhaps we can first agree to drop the last step from the series, keep calling these "versions", and then later add "v3" when we come up with an algorithm that would perform better than "v2" in almost all cases? I dunno.

Thanks.

#15 lore raw
In reply to Junio C Hamano
Quoted 10 lines · Junio C Hamano

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)

"git refs migrate" for migrating reflog data was broken.

Will merge to 'master'.
cf. <[email protected]>
source: <CAOLa=ZTL9n_DPhNr49XAd6bT838kc09oVx_AH7Pb4o8VK_xQ9w@mail.gmail.com>

This seems to be breaking on 'next'. I tested it locally with

GIT_TEST_DEFAULT_REF_FORMAT=reftable meson test -v --test-args='-i'
t1400-update-ref

my local tests were made on files backend, and it didn't trigger on the CI either for some reason (I shall investigate that soon). But dscho (CC'd) reported that macos builds for reftable were failing [1] for his branch and I could bisect it to this.

I'm yet to understand why this fails and also why the CI didn't notify of the issue. But that is something I shall do next. For now we need to remove it from next.

[1]: https://github.com/dscho/git/actions/runs/12906424058/job/35987723223

#16 lore raw
Quoted 22 lines · Karthik Nayak
Quoted 10 lines · Junio C Hamano

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)

"git refs migrate" for migrating reflog data was broken.

Will merge to 'master'.
cf. <[email protected]>
source: <CAOLa=ZTL9n_DPhNr49XAd6bT838kc09oVx_AH7Pb4o8VK_xQ9w@mail.gmail.com>

This seems to be breaking on 'next'. I tested it locally with

  GIT_TEST_DEFAULT_REF_FORMAT=reftable meson test -v --test-args='-i' t1400-update-ref

my local tests were made on files backend, and it didn't trigger on the CI either for some reason (I shall investigate that soon). But dscho (CC'd) reported that macos builds for reftable were failing [1] for his branch and I could bisect it to this.

Quoted 5 lines

I'm yet to understand why this fails and also why the CI didn't notify of the issue. But that is something I shall do next. For now we need to remove it from next.

[1]: https://github.com/dscho/git/actions/runs/12906424058/job/35987723223

This is reproducible when the leak sanitizier is enabled and tested against reftable:

So setting up meson with:
CC=clang meson setup --reconfigure -Db_sanitize=address,undefined build
and running the test in the build folder with:
GIT_TEST_DEFAULT_REF_FORMAT=reftable meson test -v
--test-args='-ixd' t1400-update-ref

reproduces the issue. I haven't found the root cause yet, but will mostly call it a day and get back to this tomorrow.

Karthik

#17 lore raw
Quoted 33 lines · Karthik Nayak
Quoted 22 lines · Karthik Nayak
Quoted 10 lines · Junio C Hamano

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)

"git refs migrate" for migrating reflog data was broken.

Will merge to 'master'.
cf. <[email protected]>
source: <CAOLa=ZTL9n_DPhNr49XAd6bT838kc09oVx_AH7Pb4o8VK_xQ9w@mail.gmail.com>

This seems to be breaking on 'next'. I tested it locally with

  GIT_TEST_DEFAULT_REF_FORMAT=reftable meson test -v --test-args='-i' t1400-update-ref

my local tests were made on files backend, and it didn't trigger on the CI either for some reason (I shall investigate that soon). But dscho (CC'd) reported that macos builds for reftable were failing [1] for his branch and I could bisect it to this.

Quoted 5 lines

I'm yet to understand why this fails and also why the CI didn't notify of the issue. But that is something I shall do next. For now we need to remove it from next.

[1]: https://github.com/dscho/git/actions/runs/12906424058/job/35987723223

This is reproducible when the leak sanitizier is enabled and tested against reftable:

Quoted 8 lines

So setting up meson with:
CC=clang meson setup --reconfigure -Db_sanitize=address,undefined build
and running the test in the build folder with:
GIT_TEST_DEFAULT_REF_FORMAT=reftable meson test -v
--test-args='-ixd' t1400-update-ref

reproduces the issue. I haven't found the root cause yet, but will mostly call it a day and get back to this tomorrow.

Thanks. I'll mark the topic as on-hold.

#18 lore raw
In reply to Karthik Nayak
Quoted 9 lines · Karthik Nayak
Quoted 7 lines · Karthik Nayak
Quoted 5 lines · Junio C Hamano

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)
...

Quoted 4 lines
Quoted 1 line

This seems to be breaking on 'next'.

...
reproduces the issue. I haven't found the root cause yet, but will mostly call it a day and get back to this tomorrow.

We have a handful of topics related to refs subsystem in flight, and I am a bit lost here.

(1) kn/reflog-migration-fix (the above) was done as a "fix" for the issue reported by brian in
https://lore.kernel.org/all/[email protected]/

(2) You mention that (1) is broken in the message I am responding to. There is no known fix yet, so (1) needs to wait in 'next' until it gets fixed.

(3) kn/reflog-migration-fix-followup is a code clean-up for (1); it has to wait for (2) as well.

(4) kn/reflog-symref-fix is a fix for a different bug the commit that introduced the bug (1) addresses. It can proceed independently from the other topics.

(5) ps/reflog-migration-with-logall-fix is another fix for a different bug introduced by the same series whose bugs are addressed by (1) and (4). It can proceed independently from the other topics.

The above is my current understanding; did I miss any other relevant topics that are related to these efforts, and/or did I misunderstand the dependencies among them?

If I am not misunderstanding the current status of these topics, I'll be marking (4) and (5) for 'next'; I am undecided for (3).

Thanks.

#19 lore raw
Quoted 39 lines · Junio C Hamano
Quoted 9 lines · Karthik Nayak
Quoted 7 lines · Karthik Nayak
Quoted 5 lines · Junio C Hamano

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)
...

Quoted 4 lines
Quoted 1 line

This seems to be breaking on 'next'.

...
reproduces the issue. I haven't found the root cause yet, but will mostly call it a day and get back to this tomorrow.

We have a handful of topics related to refs subsystem in flight, and I am a bit lost here.

(1) kn/reflog-migration-fix (the above) was done as a "fix" for the issue reported by brian in
https://lore.kernel.org/all/[email protected]/

(2) You mention that (1) is broken in the message I am responding to. There is no known fix yet, so (1) needs to wait in 'next' until it gets fixed.

(3) kn/reflog-migration-fix-followup is a code clean-up for (1); it has to wait for (2) as well.

(4) kn/reflog-symref-fix is a fix for a different bug the commit that introduced the bug (1) addresses. It can proceed independently from the other topics.

(5) ps/reflog-migration-with-logall-fix is another fix for a different bug introduced by the same series whose bugs are addressed by (1) and (4). It can proceed independently from the other topics.

Quoted 6 lines

The above is my current understanding; did I miss any other relevant topics that are related to these efforts, and/or did I misunderstand the dependencies among them?

If I am not misunderstanding the current status of these topics, I'll be marking (4) and (5) for 'next'; I am undecided for (3).

Karthik has meanwhile sent a v2 [1] of the broken patch in (1) that fixes the issue discovered in (2). Given that (1) has already been in next, (2) probably needs to be rerolled to be a patch on top of what we already have in next.

Other than that yes, I think (4) and (5) can be merged independently of
(1) to (3).

Patrick

[1]: <[email protected]>

#20 lore raw
Quoted 7 lines · Patrick Steinhardt

Karthik has meanwhile sent a v2 [1] of the broken patch in (1) that fixes the issue discovered in (2). Given that (1) has already been in next, (2) probably needs to be rerolled to be a patch on top of what we already have in next.

Other than that yes, I think (4) and (5) can be merged independently of
(1) to (3).

Thanks for sanity checking me.

#21 lore raw
In reply to Patrick Steinhardt
Quoted 52 lines · Patrick Steinhardt
Quoted 39 lines · Junio C Hamano
Quoted 9 lines · Karthik Nayak
Quoted 7 lines · Karthik Nayak
Quoted 5 lines · Junio C Hamano

* kn/reflog-migration-fix (2025-01-15) 1 commit (merged to 'next' on 2025-01-16 at ae8f9ce9a0)
+ reftable: write correct max_update_index to header (this branch is used by kn/reflog-migration-fix-followup.)
...

Quoted 4 lines
Quoted 1 line

This seems to be breaking on 'next'.

...
reproduces the issue. I haven't found the root cause yet, but will mostly call it a day and get back to this tomorrow.

We have a handful of topics related to refs subsystem in flight, and I am a bit lost here.

(1) kn/reflog-migration-fix (the above) was done as a "fix" for the issue reported by brian in
https://lore.kernel.org/all/[email protected]/

(2) You mention that (1) is broken in the message I am responding to. There is no known fix yet, so (1) needs to wait in 'next' until it gets fixed.

(3) kn/reflog-migration-fix-followup is a code clean-up for (1); it has to wait for (2) as well.

(4) kn/reflog-symref-fix is a fix for a different bug the commit that introduced the bug (1) addresses. It can proceed independently from the other topics.

(5) ps/reflog-migration-with-logall-fix is another fix for a different bug introduced by the same series whose bugs are addressed by (1) and (4). It can proceed independently from the other topics.

Quoted 6 lines

The above is my current understanding; did I miss any other relevant topics that are related to these efforts, and/or did I misunderstand the dependencies among them?

If I am not misunderstanding the current status of these topics, I'll be marking (4) and (5) for 'next'; I am undecided for (3).

Karthik has meanwhile sent a v2 [1] of the broken patch in (1) that fixes the issue discovered in (2). Given that (1) has already been in next, (2) probably needs to be rerolled to be a patch on top of what we already have in next.

Quoted 6 lines

Other than that yes, I think (4) and (5) can be merged independently of
(1) to (3).

Patrick

[1]: <[email protected]>

This seems right, just providing another set of eyes here.

Thanks!

#22 lore raw
Quoted 1 line · Karthik Nayak

This seems right, just providing another set of eyes here.

Thanks for helping me out.