What's cooking in git.git (Feb 2023, #04; Wed, 22)
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. A topic without enough support may be discarded after a long period of no activity.
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-vcs/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']
* ab/config-h-remove-unused (2023-02-07) 1 commit (merged to 'next' on 2023-02-09 at fa83258b81)
+ config.h: remove unused git_configset_add_parameters()
Code clean-up.
source: <[email protected]>
* ab/hook-api-with-stdin (2023-02-08) 5 commits
(merged to 'next' on 2023-02-13 at 4cb74ecb6b)
+ hook: support a --to-stdin=<path> option
+ sequencer: use the new hook API for the simpler "post-rewrite" call
+ hook API: support passing stdin to hooks, convert am's 'post-rewrite'
+ run-command: allow stdin for run_processes_parallel
+ run-command.c: remove dead assignment in while-loop
Extend the run-hooks API to allow feeding data from the standard input when running the hook script(s).
source: <[email protected]>
* ab/retire-scripted-add-p (2023-02-06) 3 commits
(merged to 'next' on 2023-02-09 at 1b8c7bce7d)
+ docs & comments: replace mentions of "git-add--interactive.perl"
+ add API: remove run_add_interactive() wrapper function
+ add: remove "add.interactive.useBuiltin" & Perl "git add--interactive"
Finally retire the scripted "git add -p/-i" implementation and have everybody use the one reimplemented in C.
source: <[email protected]>
* ab/sequencer-unleak (2023-02-06) 8 commits
(merged to 'next' on 2023-02-09 at 98bff1953f)
+ commit.c: free() revs.commit in get_fork_point()
+ builtin/rebase.c: free() "options.strategy_opts"
+ sequencer.c: always free() the "msgbuf" in do_pick_commit()
+ builtin/rebase.c: fix "options.onto_name" leak
+ builtin/revert.c: move free-ing of "revs" to replay_opts_release()
+ sequencer API users: fix get_replay_opts() leaks
+ sequencer.c: split up sequencer_remove_state()
+ rebase: use "cleanup" pattern in do_interactive_rebase()
Plug leaks in sequencer subsystem and its users.
source: <[email protected]>
* ab/the-index-compatibility (2023-02-10) 6 commits
(merged to 'next' on 2023-02-13 at 4f9869d4cf)
+ cocci & cache.h: remove "USE_THE_INDEX_COMPATIBILITY_MACROS"
+ cache-tree API: remove redundant update_main_cache_tree()
+ cocci & cache-tree.h: migrate "write_cache_as_tree" to "*_index_*"
+ cocci & cache.h: apply pending "index_cache_pos" rule
+ cocci & cache.h: fully apply "active_nr" part of index-compatibility
+ builtin/rm.c: use narrower "USE_THE_INDEX_VARIABLE"
Remove more remaining uses of macros that relies on the_index singleton instance without explicitly spelling it out.
source: <[email protected]>
* ab/various-leak-fixes (2023-02-06) 19 commits
(merged to 'next' on 2023-02-13 at 0cee9746f7)
+ push: free_refs() the "local_refs" in set_refspecs()
+ push: refactor refspec_append_mapped() for subsequent leak-fix
+ receive-pack: release the linked "struct command *" list
+ grep API: plug memory leaks by freeing "header_list"
+ grep.c: refactor free_grep_patterns()
+ builtin/merge.c: free "&buf" on "Your local changes..." error
+ builtin/merge.c: use fixed strings, not "strbuf", fix leak
+ show-branch: free() allocated "head" before return
+ commit-graph: fix a parse_options_concat() leak
+ http-backend.c: fix cmd_main() memory leak, refactor reg{exec,free}()
+ http-backend.c: fix "dir" and "cmd_arg" leaks in cmd_main()
+ worktree: fix a trivial leak in prune_worktrees()
+ repack: fix leaks on error with "goto cleanup"
+ name-rev: don't xstrdup() an already dup'd string
+ various: add missing clear_pathspec(), fix leaks
+ clone: use free() instead of UNLEAK()
+ commit-graph: use free_commit_graph() instead of UNLEAK()
+ bundle.c: don't leak the "args" in the "struct child_process"
+ tests: mark tests as passing with SANITIZE=leak
Leak fixes.
source: <[email protected]>
* ar/userdiff-java-update (2023-02-08) 3 commits
(merged to 'next' on 2023-02-09 at f62497382a)
+ userdiff: support Java sealed classes
+ userdiff: support Java record types
+ userdiff: support Java type parameters
Userdiff regexp update for Java language.
source: <[email protected]>
* cb/grep-fallback-failing-jit (2023-01-31) 1 commit (merged to 'next' on 2023-02-08 at 3a70d6ef6c)
+ grep: fall back to interpreter if JIT memory allocation fails
In an environment where dynamically generated code is prohibited to run (e.g. SELinux), failure to JIT pcre patterns is expected. Fall back to interpreted execution in such a case.
source: <[email protected]>
* cw/doc-pushurl-vs-url (2023-02-07) 1 commit (merged to 'next' on 2023-02-09 at 21b0678d19)
+ Documentation: clarify multiple pushurls vs urls
Doc update.
source: <[email protected]>
* ds/bundle-uri-5 (2023-01-31) 11 commits (merged to 'next' on 2023-02-08 at 1ba9ba565d)
+ bundle-uri: test missing bundles with heuristic
+ bundle-uri: store fetch.bundleCreationToken
+ fetch: fetch from an external bundle URI
+ bundle-uri: drop bundle.flag from design doc
+ clone: set fetch.bundleURI if appropriate
+ bundle-uri: download in creationToken order
+ bundle-uri: parse bundle.<id>.creationToken values
+ bundle-uri: parse bundle.heuristic=creationToken
+ t5558: add tests for creationToken heuristic
+ bundle: verify using check_connected()
+ bundle: test unbundling with incomplete history
The bundle-URI subsystem adds support for creation-token heuristics to help incremental fetches.
source: <[email protected]>
* en/name-rev-make-taggerdate-much-less-important (2023-02-09) 1 commit (merged to 'next' on 2023-02-10 at 059bc956aa)
+ name-rev: fix names by dropping taggerdate workaround
"git name-rev" heuristics update.
source: <[email protected]>
* jk/doc-ls-remote-matching (2023-02-10) 2 commits (merged to 'next' on 2023-02-13 at 79a209bcb3)
+ doc/ls-remote: clarify pattern format
+ doc/ls-remote: cosmetic cleanups for examples
Doc update.
source: <[email protected]>
* kf/t5000-modernise (2023-02-06) 1 commit
(merged to 'next' on 2023-02-09 at 0341c54639)
+ t5000: modernise archive and :(glob) test
Test clean-up.
source: <[email protected]>
* po/attributes-text (2023-02-06) 1 commit
(merged to 'next' on 2023-02-09 at 8be7b134da)
+ .gitattributes: include text attribute for eol attributes
In-tree .gitattributes update to match the way we recommend our users to mark a file as text.
source: <[email protected]>
* rs/cache-tree-strbuf-growth-fix (2023-02-10) 1 commit (merged to 'next' on 2023-02-13 at fd9ae8790b)
+ cache-tree: fix strbuf growth in prime_cache_tree_rec()
Remove unnecessary explicit sizing of strbuf.
source: <[email protected]>
* rs/size-t-fixes (2023-02-06) 2 commits (merged to 'next' on 2023-02-09 at 4bf6998479)
+ pack-objects: use strcspn(3) in name_cmp_len()
+ read-cache: use size_t for {base,df}_name_compare()
Type fixes.
source: <[email protected]>
* wl/new-command-doc (2023-02-06) 1 commit (merged to 'next' on 2023-02-09 at 23bcd4d4d9)
+ new-command.txt: update reference to builtin docs
Comment fix.
source: <[email protected]>
--------------------------------------------------
[New Topics]
* jc/diff-algo-attribute (2023-02-21) 2 commits (merged to 'next' on 2023-02-22 at 2375640330)
+ diff: teach diff to read algorithm from diff driver
+ diff: consolidate diff algorithm option parsing
The "diff" drivers specified by the "diff" attribute attached to paths can now specify which algorithm (e.g. histogram) to use.
Will merge to 'master'.
source: <[email protected]>
* jk/shorten-unambiguous-ref-wo-sscanf (2023-02-15) 3 commits
- shorten_unambiguous_ref(): avoid sscanf()
- shorten_unambiguous_ref(): use NUM_REV_PARSE_RULES constant
- shorten_unambiguous_ref(): avoid integer truncation
sscanf(3) used in "git symbolic-ref --short" implementation found to be not working reliably on macOS in UTF-8 locales. Rewrite the code to avoid sscanf() altogether to work it around.
Will merge to 'next'?
source: <[email protected]>
* js/gpg-errors (2023-02-15) 2 commits
(merged to 'next' on 2023-02-16 at 510898ff10)
+ gpg: do show gpg's error message upon failure
+ t7510: add a test case that does not need gpg
Error messages given upon a signature verification failure used to discard the errors from underlying gpg program, which has been corrected.
Will merge to 'master'.
source: <[email protected]>
* jc/genzeros-avoid-raw-write (2023-02-16) 1 commit (merged to 'next' on 2023-02-22 at a60fe69bce)
+ test-genzeros: avoid raw write(2)
A test helper had a single write(2) of 256kB, which was too big for some platforms (e.g. NonStop), which has been corrected by using xwrite() wrapper appropriately.
Will merge to 'master'.
source: <[email protected]>
* jk/http-proxy-tests (2023-02-16) 1 commit
(merged to 'next' on 2023-02-22 at f1f69d188f)
+ add basic http proxy tests
Test updates.
Will merge to 'master'.
source: <[email protected]>
* ps/free-island-marks (2023-02-21) 1 commit (merged to 'next' on 2023-02-22 at f46938bce3)
+ delta-islands: fix segfault when freeing island marks
Fix on a previous fix already in 'master'.
Will merge to 'master'.
source: <61e490595b80b34c55fd640e093e021ff6fa9591.1676542973.git.ps@pks.im>
* tb/drop-dir-iterator-follow-symlink-bit (2023-02-16) 2 commits (merged to 'next' on 2023-02-22 at 785e1f323e)
+ t0066: drop setup of "dir5"
+ dir-iterator: drop unused DIR_ITERATOR_FOLLOW_SYMLINKS
Remove leftover and unused code.
Will merge to 'master'.
source: <9bb10b607e46f867a3f8f5c71abf13c990d1ecfe.1676572031.git.me@ttaylorr.com>
source: <CAGdrTFhHBU2BNYdYr7LbOS7i1LOHGjWLw_d5ZJAXxvTKyLiFCA@mail.gmail.com>
* jc/countermand-format-attach (2023-02-17) 1 commit
- format.attach: allow empty value to disable multi-part messages
The format.attach configuration variable lacked a way to override a value defined in a lower-priority configuration file (e.g. the system one) by redefining it in a higher-priority configuration file. Now, setting format.attach to an empty string means show the patch inline in the e-mail message, without using MIME attachment.
This is a backward incompatible change.
Will merge to 'next'?
source: <[email protected]>
* rs/archive-mtime (2023-02-18) 1 commit
- archive: add --mtime
"git archive HEAD^{tree}" records the paths with the current timestamp in the archive, making it harder to obtain a stable output. The command learned the --mtime option to specify an arbitrary timestamp (e.g. --mtime="@0 +0000" for the epoch).
Will merge to 'next'?
source: <[email protected]>
* ap/t2015-style-update (2023-02-21) 1 commit (merged to 'next' on 2023-02-22 at 40883b85c7)
+ t2015-checkout-unborn.sh: changes the style for cd
Test clean-up.
Will merge to 'master'.
source: <[email protected]>
* if/simplify-trace-setup (2023-02-21) 1 commit
- trace.c, git.c: remove unnecessary parameter to trace_repo_setup()
Code clean-up.
Will merge to 'next'?
source: <[email protected]>
* ma/fetch-parallel-use-online-cpus (2023-02-21) 1 commit (merged to 'next' on 2023-02-22 at 756d379bf0)
+ fetch: choose a sensible default with --jobs=0 again
"git fetch --jobs=0" used to hit a BUG(), which has been corrected to use the available CPUs.
Will merge to 'master'.
source: <[email protected]>
* pw/rebase-i-validate-labels-early (2023-02-21) 1 commit (merged to 'next' on 2023-02-22 at 9b37d42d24)
+ rebase -i: check labels and refs when parsing todo list
An invalid label or ref in the "rebase -i" todo file used to trigger an runtime error. SUch an error is now diagnosed while the todo file is parsed.
Will merge to 'master'.
source: <[email protected]>
* tl/range-diff-custom-abbrev (2023-02-21) 1 commit (merged to 'next' on 2023-02-22 at 49db7cce13)
+ range-diff: let '--abbrev' option takes effect
"git range-diff" learned --abbrev=<num> option.
Will merge to 'master'.
source: <[email protected]>
--------------------------------------------------
[Stalled]
* ja/worktree-orphan (2023-01-13) 4 commits
- worktree add: add hint to direct users towards --orphan
- worktree add: add --orphan flag
- worktree add: refactor opt exclusion tests
- worktree add: include -B in usage docs
'git worktree add' learned how to create a worktree based on an orphaned branch with --orphan.
Expecting a reroll.
cf. <[email protected]>
source: <[email protected]>
* tl/notes--blankline (2023-02-16) 3 commits
- notes.c: introduce "--separator" option
- notes.c: cleanup for "designated init"
- notes.c: cleanup 'strbuf_grow' call in 'append_edit'
'git notes append' was taught '--separator' to specify string to insert between paragraphs.
source: <[email protected]>
* ab/tag-object-type-errors (2022-11-22) 5 commits
- tag: don't emit potentially incorrect "object is a X, not a Y"
- tag: don't misreport type of tagged objects in errors
- object tests: add test for unexpected objects in tags
- object-file.c: free the "t.tag" in check_tag()
- Merge branch 'jk/parse-object-type-mismatch' into ab/tag-object-type-errors
Hardening checks around mismatched object types when one of those objects is a tag.
Expecting a reroll.
cf. <[email protected]>
cf. <[email protected]>
source: <[email protected]>
* ed/fsmonitor-inotify (2022-12-13) 6 commits
- fsmonitor: update doc for Linux
- fsmonitor: test updates
- fsmonitor: enable fsmonitor for Linux
- fsmonitor: implement filesystem change listener for Linux
- fsmonitor: determine if filesystem is local or remote
- fsmonitor: prepare to share code between Mac OS and Linux
Bundled fsmonitor for Linux using inotify API.
Needs review on the updated round.
source: <[email protected]>
* jc/spell-id-in-both-caps-in-message-id (2022-12-17) 1 commit
- e-mail workflow: Message-ID is spelled with ID in both capital letters
Consistently spell "Message-ID" as such, not "Message-Id".
Needs review.
source: <[email protected]>
* ad/test-record-count-when-harness-is-in-use (2022-12-25) 1 commit
- test-lib: allow storing counts with test harnesses
Allow summary results from tests to be written to t/test-results directory even when a test harness like 'prove' is in use.
Needs review.
source: <[email protected]>
* so/diff-merges-more (2022-12-18) 5 commits
- diff-merges: improve --diff-merges documentation
- diff-merges: issue warning on lone '-m' option
- diff-merges: support list of values for --diff-merges
- diff-merges: implement log.diffMerges-m-imply-p config
- diff-merges: implement [no-]hide option and log.diffMergesHide config
Assorted updates to "--diff-merges=X" option.
May want to discard. Breaking compatibility does not seem worth it.
source: <[email protected]>
--------------------------------------------------
[Cooking]
* ew/commit-reach-clean-up-flags-fix (2023-02-11) 1 commit
- commit-reach: avoid NULL dereference
Under discussion.
cf. <[email protected]>
source: <[email protected]>
* rs/ctype-test (2023-02-13) 3 commits
(merged to 'next' on 2023-02-15 at 279f198dec)
+ test-ctype: test iscntrl, ispunct, isxdigit and isprint
+ test-ctype: test islower and isupper
+ test-ctype: test isascii
Test safe_ctype
Will merge to 'master'.
source: <[email protected]>
* mh/credential-password-expiry (2023-02-18) 1 commit
- credential: new attribute password_expiry_utc
The credential subsystem learned that a password may have an explicit expiration.
Will merge to 'next'?
source: <[email protected]>
* tk/pull-conflict-suggest-rebase-merge-not-rebase-true (2023-02-13) 1 commit
- pull: conflict hint pull.rebase suggestion should offer "merges" vs "true"
In an advice message after failed non-ff pull, we used to suggest setting pull.rebase=true, but these days pull.rebase=merges may be more inline with the original spirit of "rebuild your side on top of theirs".
cf. <CAMMLpeTPEoKVTbfc17w+Y9qn7jOGmQi_Ux0Y3sFW5QTgGWJ=SA@mail.gmail.com>
source: <[email protected]>
* jc/gpg-lazy-init (2023-02-09) 1 commit
- gpg-interface: lazily initialize and read the configuration
Instead of forcing each command to choose to honor GPG related configuration variables, make the subsystem lazily initialize itself.
Will merge to 'next'?
source: <[email protected]>
* mc/credential-helper-www-authenticate (2023-02-16) 3 commits
- credential: add WWW-Authenticate header to cred requests
- http: read HTTP WWW-Authenticate response headers
- t5563: add tests for basic and anoymous HTTP access
Allow information carried on the WWW-AUthenticate header to be passed to the credential helpers.
Will merge to 'next'?
source: <[email protected]>
* ab/avoid-losing-exit-codes-in-tests (2023-02-06) 6 commits
- tests: don't lose misc "git" exit codes
- tests: don't lose exit status with "test <op> $(git ...)"
- tests: don't lose "git" exit codes in "! ( git ... | grep )"
- tests: don't lose exit status with "(cd ...; test <op> $(git ...))"
- t/lib-patch-mode.sh: fix ignored exit codes
- auto-crlf tests: don't lose exit code in loops and outside tests
Test clean-up.
Will merge to 'next'?
source: <[email protected]>
* rj/branch-unborn-in-other-worktrees (2023-02-06) 3 commits
- branch: rename orphan branches in any worktree
- branch: description for orphan branch errors
- branch: avoid unnecessary worktrees traversals
Error messages given when working on an unborn branch that is checked out in another worktree have been improvved.
Expecting a reroll.
cf. <[email protected]>
source: <[email protected]>
* ab/imap-send-requires-curl (2023-02-02) 6 commits
- imap-send: correctly report "host" when using "tunnel"
- imap-send: remove old --no-curl codepath
- imap-send: make --curl no-optional
- imap-send: replace auto-probe libcurl with hard dependency
- imap-send doc: the imap.sslVerify is used with imap.tunnel
- imap-send: note "auth_method", not "host" on auth method failure
Give a hard dependency on cURL library to build "git imap-send", and remove the code to interact with IMAP server without using cURL.
Expecting a reroll.
The 'tunnel' part is still iffy.
cf. <[email protected]>
source: <[email protected]>
* rd/doc-default-date-format (2023-02-15) 1 commit (merged to 'next' on 2023-02-22 at 3818705ca9)
+ rev-list: clarify git-log default date format
Update --date=default documentation.
Will merge to 'master'.
source: <[email protected]>
* ab/config-multi-and-nonbool (2023-02-07) 10 commits
- for-each-repo: with bad config, don't conflate <path> and <cmd>
- config API: add "string" version of *_value_multi(), fix segfaults
- config API users: test for *_get_value_multi() segfaults
- for-each-repo: error on bad --config
- config API: don't lose the git_*get*() return values
- config API: have *_multi() return an "int" and take a "dest"
- versioncmp.c: refactor config reading next commit
- config API: add and use a "git_config_get()" family of functions
- config tests: add "NULL" tests for *_get_value_multi()
- config tests: cover blind spots in git_die_config() tests
Assorted config API updates.
Will merge to 'next'?
source: <[email protected]>
* cb/checkout-same-branch-twice (2023-01-20) 1 commit
- checkout/switch: disallow checking out same branch in multiple worktrees
"git checkout -B $branch" failed to protect against checking out a branch that is checked out elsewhere, unlike "git branch -f" did.
Expecting a hopefully minor and final reroll. cf. <[email protected]>
source: <[email protected]>
* rj/avoid-switching-to-already-used-branch (2023-01-22) 3 commits
- switch: reject if the branch is already checked out elsewhere (test)
- rebase: refuse to switch to a branch already checked out elsewhere (test)
- branch: fix die_if_checked_out() when ignore_current_worktree
A few subcommands have been taught to stop users from working on a branch that is being used in another worktree linked to the same repository.
Expecting a hopefully minor and final reroll. cf. <[email protected]>
source: <[email protected]>
* rj/bisect-already-used-branch (2023-01-22) 1 commit
- bisect: fix "reset" when branch is checked out elsewhere
Allow "git bisect reset [name]" to check out the named branch (or the original one) even when the branch is already checked out in a different worktree linked to the same repository.
Will merge to 'next'?
source: <[email protected]>
* ms/send-email-feed-header-to-validate-hook (2023-01-19) 2 commits
- send-email: expose header information to git-send-email's sendemail-validate hook
- send-email: refactor header generation functions
"git send-email" learned to give the e-mail headers to the validate hook by passing an extra argument from the command line.
Expecting a hopefully minor and final reroll.
cf. <[email protected]>
source: <[email protected]>
* tc/cat-file-z-use-cquote (2023-01-16) 1 commit
- cat-file: quote-format name in error when using -z
"cat-file" in the batch mode that is fed NUL-terminated pathnames learned to cquote them in its error output (otherwise, a funny pathname with LF in it would break the lines in the output stream).
Expecting a reroll.
cf. <[email protected]>
source: <[email protected]>
* cw/submodule-status-in-parallel (2023-02-09) 6 commits
- diff-lib: parallelize run_diff_files for submodules
- diff-lib: refactor out diff_change logic
- submodule: refactor is_submodule_modified()
- submodule: move status parsing into function
- submodule: strbuf variable rename
- run-command: add duplicate_output_fn to run_processes_parallel_opts
"git submodule status" learned to run the comparison in submodule repositories in parallel.
Not quite there yet...
source: <[email protected]>
--------------------------------------------------
[Discarded]
* mc/switch-advice (2022-11-09) 1 commit
. po: use switch over checkout in error message
Use 'switch' instead of 'checkout' in an error message.
Will discard.
Getting tired of waiting for review response.
* js/range-diff-mbox (2022-11-23) 1 commit
. range-diff: support reading mbox files
'git range-diff' gained support for reading either side from an .mbox file instead of a revision range.
Will discard.
Getting tired of waiting for review response.
cf. <[email protected]>
source: <[email protected]>
* jc/test-prereq-symlink-update (2023-02-09) 1 commit . test: make SYMLINKS prerequisite more robust
For some reason, "ln -s x y && test -h y" started passing on Windows, even though we do not have working readlink(3) there. Tighten the check to enable SYMLINKS prerequisite in our tests.
It turns out that this was caused by using a faulty build of the Windows dev environment, which was fixed.
source: <[email protected]>
* jc/t5559-workaround (2023-02-09) 1 commit
. t5559: skip a known-to-be-broken test
source: <[email protected]>
mc/credential-helper-www-authenticate (Re: What's cooking in git.git (Feb 2023, #04; Wed, 22))
Quoted 10 lines · Junio C Hamano
* mc/credential-helper-www-authenticate (2023-02-16) 3 commits
- credential: add WWW-Authenticate header to cred requests
- http: read HTTP WWW-Authenticate response headers
- t5563: add tests for basic and anoymous HTTP accessAllow information carried on the WWW-AUthenticate header to be passed to the credential helpers.
Will merge to 'next'?
source: <[email protected]>
Aside from some relatively minor touch-ups, this has been fairly stable since v8. I'm happy with it at this point, but I'm also curious to hear Peff's thoughts (since it was his review that prompted the switch to Apache for the test helper, among other changes).
Re: mc/credential-helper-www-authenticate (Re: What's cooking in git.git (Feb 2023, #04; Wed, 22))
Quoted 9 lines · Victoria Dye
Quoted 8 lines · Junio C Hamano
* mc/credential-helper-www-authenticate (2023-02-16) 3 commits
- credential: add WWW-Authenticate header to cred requests
- http: read HTTP WWW-Authenticate response headers
- t5563: add tests for basic and anoymous HTTP accessAllow information carried on the WWW-AUthenticate header to be passed to the credential helpers.
Quoted 7 lines
Quoted 2 lines
Will merge to 'next'?
source: <[email protected]>Aside from some relatively minor touch-ups, this has been fairly stable since v8. I'm happy with it at this point, but I'm also curious to hear Peff's thoughts (since it was his review that prompted the switch to Apache for the test helper, among other changes).
I just gave v10 a careful read. It looks quite nice to me, though there were a few tiny nits that I think are worth a re-roll (the most important being the strcasecmp buffer over-read).
-Peff
Re: mc/credential-helper-www-authenticate (Re: What's cooking in git.git (Feb 2023, #04; Wed, 22))
Quoted 14 lines · Jeff King
Quoted 12 lines · Victoria Dye
Quoted 10 lines · Junio C Hamano
* mc/credential-helper-www-authenticate (2023-02-16) 3 commits
- credential: add WWW-Authenticate header to cred requests
- http: read HTTP WWW-Authenticate response headers
- t5563: add tests for basic and anoymous HTTP accessAllow information carried on the WWW-AUthenticate header to be passed to the credential helpers.
Will merge to 'next'?
source: <[email protected]>
Quoted 8 lines
Quoted 4 lines
Aside from some relatively minor touch-ups, this has been fairly stable since v8. I'm happy with it at this point, but I'm also curious to hear Peff's thoughts (since it was his review that prompted the switch to Apache for the test helper, among other changes).
I just gave v10 a careful read. It looks quite nice to me, though there were a few tiny nits that I think are worth a re-roll (the most important being the strcasecmp buffer over-read).
Thanks for the thorough review all :)
Quoted 1 line
-Peff
ab/config-multi-and-nonbool (was Re: What's cooking in git.git (Feb 2023, #04; Wed, 22))
In reply to Junio C HamanoQuoted 12 lines · Junio C Hamano
* ab/config-multi-and-nonbool (2023-02-07) 10 commits
- for-each-repo: with bad config, don't conflate <path> and <cmd>
- config API: add "string" version of*_value_multi(), fix segfaults
- config API users: test for*_get_value_multi()segfaults
- for-each-repo: error on bad --config
- config API: don't lose the git_*get*() return values
- config API: have*_multi()return an "int" and take a "dest"
-versioncmp.c: refactor config reading next commit
- config API: add and use a "git_config_get()" family of functions
- config tests: add "NULL" tests for*_get_value_multi()
- config tests: cover blind spots ingit_die_config()tests
Quoted 4 lines
Assorted config API updates.
Will merge to 'next'?
source: <[email protected]>
I would prefer to eject 06/10 [1] for now and leave it in for a future cleanup series. I haven't confirmed whether it's safe (the practical effect of that patch is that the *_get() functions can now return -1 instead of 1, so the patch is safe if all callers only check if the return value is zero, and not whether it has a particular sign), and I don't think 06/10 is absolutely necessary to the series.
But if we are already quite convinced that this is safe, I have no objections to merging it as-is.
Re: ab/config-multi-and-nonbool (was Re: What's cooking in git.git (Feb 2023, #04; Wed, 22))
Quoted 6 lines · Glen Choo
I would prefer to eject 06/10 [1] for now and leave it in for a future cleanup series. I haven't confirmed whether it's safe (the practical effect of that patch is that the
*_get()functions can now return -1 instead of 1, so the patch is safe if all callers only check if the return value is zero, and not whether it has a particular sign), and I don't think 06/10 is absolutely necessary to the series.
If somebody wants to help auditing potential breakage the patch in question causes, "are callers happy to declare an error upon non-zero return?" is not exactly sufficient. While all of these functions return 0 upon success, some of them return 1 on certain cases while error() on other cases, and we need to ensure that the callers are not behaving differently upon these values.
So let's see how many functions does the patch in question touch.
Quoted 8 lines
config.c +0
-0@@ -2449,86 +2449,93 @@ int git_configset_get(struct config_set *cs, const char *key)int git_configset_get_string (struct config_set *cs,const char *key,char **dest) {const char *value;
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2457
Quoted 10 lines
config.c +5
-4-if (!git_configset_get_value (cs, key, &value)) -return git_config_string ((const char **)dest, key, value); -else -return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; +return git_config_string ((const char **)dest, key, value); }
This used to return 1 when there is no such value, and get whatever error signal git_config_string() gave if git_config_string() failed.
Luckily, git_configset_get_value() returns 1 when there is no such value, so I think this hunk is an expensive-to-audit no-op.
Quoted 11 lines
config.c +0
-7static int git_configset_get_string_tmp (struct config_set *cs,const char *key,const char **dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) { -if (!value) -return config_error_nonbool (key); - *dest = value; -return 0 ; - }else { -return 1 ;
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2470
Quoted 10 lines
config.c +8
-1- } +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; +if (!value) +return config_error_nonbool (key); + *dest = value; +return 0 ; }
Ditto.
Quoted 5 lines
config.c +0
-2int git_configset_get_int (struct config_set *cs,const char *key,int *dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) { - *dest =git_config_int (key, value);
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2480
Quoted 10 lines
config.c +6
-3-return 0 ; - }else -return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; + *dest =git_config_int (key, value); +return 0 ; }
Ditto.
Quoted 5 lines
config.c +0
-2int git_configset_get_ulong (struct config_set *cs,const char *key,unsigned long *dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) { - *dest =git_config_ulong (key, value);
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2490
Quoted 10 lines
config.c +6
-3-return 0 ; - }else -return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; + *dest =git_config_ulong (key, value); +return 0 ; }
Ditto.
Quoted 5 lines
config.c +0
-2int git_configset_get_bool (struct config_set *cs,const char *key,int *dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) { - *dest =git_config_bool (key, value);
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2500
Quoted 10 lines
config.c +6
-3-return 0 ; - }else -return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; + *dest =git_config_bool (key, value); +return 0 ; }
Ditto.
Quoted 6 lines
config.c +0
-2int git_configset_get_bool_or_int (struct config_set *cs,const char *key,int *is_bool,int *dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) { - *dest =git_config_bool_or_int (key, value, is_bool);
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2511
Quoted 10 lines
config.c +6
-3-return 0 ; - }else -return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; + *dest =git_config_bool_or_int (key, value, is_bool); +return 0 ; }
Ditto.
Quoted 9 lines
config.c +0
-6int git_configset_get_maybe_bool (struct config_set *cs,const char *key,int *dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) { - *dest =git_parse_maybe_bool (value); -if (*dest ==-1 ) -return -1 ; -return 0 ; - }else
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2523
Quoted 10 lines
config.c +8
-1-return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; + *dest =git_parse_maybe_bool (value); +if (*dest ==-1 ) +return -1 ; +return 0 ; }
Ditto.
config.c int git_configset_get(struct config_set *cs, const char *key) · line 2532
Quoted 13 lines
config.c +5
-4int git_configset_get_pathname (struct config_set *cs,const char *key,const char **dest) {const char *value; -if (!git_configset_get_value (cs, key, &value)) -return git_config_pathname (dest, key, value); -else -return 1 ; +int ret; + +if ((ret =git_configset_get_value (cs, key, &value))) +return ret; +return git_config_pathname (dest, key, value); }
Ditto.
config.c int git_config_get_expiry_in_days(const char *key, timestamp_t *expiry, timestam · line 2789
Quoted 11 lines
config.c +4
-2@@ -2775,9 +2782,11 @@ int git_config_get_expiry_in_days(const char *key, timestamp_t *expiry, timestamconst char *expiry_string;intmax_t days;timestamp_t when; +int ret; -if (git_config_get_string_tmp (key, &expiry_string)) -return 1 ;/* no such thing */ +if ((ret =git_config_get_string_tmp (key, &expiry_string))) +/* no such thing, or git_config_parse_key() failure etc. */ +return ret;
This was returning -1 for an error (at the end of the function), as well as 1 for "no such thing". Now, this is a breaking change if any callers wanted to tell between -1 and 1.
Luckily, I think its two callers discard the error return; they initialize expiry to some reasonable default and rely on the fact that the expiry is not updated when these configuration API calls did not find anything interesting.
config.c int git_config_get_index_threads(int *dest)
Quoted 1 line
config.c +0
-0@@ -2827,15 +2837,14 @@ int git_config_get_index_threads(int *dest)
You can easily find output in "git grep git_config_get_index_threads" that all four callers only care if they get 0 or not, so this one should be safe.