What's cooking in git.git (May 2019, #01; Thu, 9)
Here are the topics that have been cooking. Commits prefixed with '-' are only in 'pu' (proposed updates) while commits prefixed with '+' are in 'next'. The ones marked with '.' do not appear in any of the integration branches, but I am still holding onto them.
The 8th batch of topics, which Hopefully is the final one before -rc0, have been pushed out to 'master'. The "no-extern" topic is now in 'next', with its merge conflict with many other topics, is still slowing me down when it is moved earlier in the merge order. I expect it to need only one more topic shuffling before merged to 'master', so I hope I'd survive.
You can find the changes described here in the integration branches of the repositories listed at
http://git-blame.blogspot.com/p/git-public-repositories.html
--------------------------------------------------
[Graduated to "master"]
* cc/aix-has-fileno-as-a-macro (2019-04-25) 1 commit (merged to 'next' on 2019-04-25 at f1d6464f98)
+ Makefile: use fileno macro work around on AIX (this branch is used by cc/access-on-aix-workaround.)
AIX shared the same build issues with other BSDs around fileno(fp), which has been corrected.
* cc/replace-graft-peel-tags (2019-04-15) 4 commits (merged to 'next' on 2019-04-25 at f8d0db27ba)
+ replace: peel tag when passing a tag first to --graft
+ replace: peel tag when passing a tag as parent to --graft
+ t6050: redirect expected error output to a file
+ t6050: use test_line_count instead of wc -l
When given a tag that points at a commit-ish, "git replace --graft" failed to peel the tag before writing a replace ref, which did not make sense because the old graft mechanism the feature wants to mimick only allowed to replace one commit object with another. This has been fixed.
* dl/merge-cleanup-scissors-fix (2019-04-19) 10 commits (merged to 'next' on 2019-04-25 at 2014eef6b1)
+ cherry-pick/revert: add scissors line on merge conflict
+ sequencer.c: save and restore cleanup mode
+ merge: add scissors line on merge conflict
+ merge: cleanup messages like commit
+ parse-options.h: extract common --cleanup option
+ commit: extract cleanup_mode functions to sequencer
+ t7502: clean up style
+ t7604: clean up style
+ t3507: clean up style
+ t7600: clean up style
(this branch uses pw/sequencer-cleanup-with-signoff-x-fix.)
The list of conflicted paths shown in the editor while concluding a conflicted merge was shown above the scissors line when the clean-up mode is set to "scissors", even though it was commented out just like the list of updated paths and other information to help the user explain the merge better.
* dl/warn-tagging-a-tag (2019-04-12) 2 commits
(merged to 'next' on 2019-04-25 at 8b966d7fe8)
+ tag: advise on nested tags
+ tag: fix formatting
"git tag" learned to give an advice suggesting it might be a mistake when creating an annotated or signed tag that points at another tag.
* dr/ref-filter-push-track-fix (2019-04-18) 1 commit (merged to 'next' on 2019-04-25 at 07db067adc)
+ ref-filter: use correct branch for %(push:track)
%(push:track) token used in the --format option to "git for-each-ref" and friends was not showing the right branch, which has been fixed.
* en/merge-directory-renames (2019-04-08) 15 commits
(merged to 'next' on 2019-04-25 at fd5b4f57b4)
+ merge-recursive: switch directory rename detection default
+ merge-recursive: give callers of handle_content_merge() access to contents
+ merge-recursive: track information associated with directory renames
+ t6043: fix copied test description to match its purpose
+ merge-recursive: switch from (oid,mode) pairs to a diff_filespec
+ merge-recursive: cleanup handle_rename_* function signatures
+ merge-recursive: track branch where rename occurred in rename struct
+ merge-recursive: remove ren[12]_other fields from rename_conflict_info
+ merge-recursive: shrink rename_conflict_info
+ merge-recursive: move some struct declarations together
+ merge-recursive: use 'ci' for rename_conflict_info variable name
+ merge-recursive: rename locals 'o' and 'a' to 'obuf' and 'abuf'
+ merge-recursive: rename diff_filespec 'one' to 'o'
+ merge-recursive: rename merge_options argument from 'o' to 'opt'
+ Use 'unsigned short' for mode, like diff_filespec does
"git merge-recursive" backend recently learned a new heuristics to infer file movement based on how other files in the same directory moved. As this is inherently less robust heuristics than the one based on the content similarity of the file itself (rather than based on what its neighbours are doing), it sometimes gives an outcome unexpected by the end users. This has been toned down to leave the renamed paths in higher/conflicted stages in the index so that the user can examine and confirm the result.
* jk/pack-objects-reports-num-objects-to-trace2 (2019-04-12) 1 commit (merged to 'next' on 2019-04-25 at e79464c054)
+ pack-objects: write objects packed to trace2
The "git pack-objects" command learned to report the number of objects it packed via the trace2 mechanism.
* jk/prune-optim (2019-04-19) 1 commit
(merged to 'next' on 2019-04-25 at c50353b27f)
+ t5304: add a test for pruning with bitmaps
A follow-up test for an earlier "git prune" improvements.
* jk/untracked-cache-more-fixes (2019-04-19) 3 commits (merged to 'next' on 2019-04-25 at a6037ddd54)
+ untracked-cache: simplify parsing by dropping "len"
+ untracked-cache: simplify parsing by dropping "next"
+ untracked-cache: be defensive about missing NULs in index
Code clean-up.
* js/misc-doc-fixes (2019-04-19) 8 commits
(merged to 'next' on 2019-04-25 at 6898f709d0)
+ Turn git serve into a test helper
+ test-tool: handle the -C <directory> option just like git
+ check-docs: do not bother checking for legacy scripts' documentation
+ docs: exclude documentation for commands that have been excluded
+ check-docs: allow command-list.txt to contain excluded commands
+ help -a: do not list commands that are excluded from the build
+ Makefile: drop the NO_INSTALL variable
+ remote-testgit: move it into the support directory for t5801
"make check-docs", "git help -a", etc. did not account for cases where a particular build may deliberately omit some subcommands, which has been corrected.
* js/trace2-to-directory (2019-03-22) 1 commit
(merged to 'next' on 2019-04-25 at 53adf71c41)
+ trace2: write to directory targets
The trace2 tracing facility learned to auto-generate a filename when told to log to a directory.
* jt/clone-server-option (2019-04-18) 2 commits (merged to 'next' on 2019-04-25 at 21f07cc85d)
+ clone: send server options when using protocol v2
+ transport: die if server options are unsupported
"git clone" learned a new --server-option option when talking over the protocol version 2.
* jt/submodule-repo-is-with-worktree (2019-04-21) 1 commit (merged to 'next' on 2019-04-25 at da2c6d684d)
+ worktree: update is_bare heuristics
The logic to tell if a Git repository has a working tree protects "git branch -D" from removing the branch that is currently checked out by mistake. The implementation of this logic was broken for repositories with unusual name, which unfortunately is the norm for submodules these days. This has been fixed.
* km/empty-repo-is-still-a-repo (2019-04-10) 3 commits (merged to 'next' on 2019-04-25 at bb3d4406a5)
+ add: error appropriately on repository with no commits
+ dir: do not traverse repositories with no commits
+ submodule: refuse to add repository with no commits
Running "git add" on a repository created inside the current repository is an explicit indication that the user wants to add it as a submodule, but when the HEAD of the inner repository is on an unborn branch, it cannot be added as a submodule. Worse, the files in its working tree can be added as if they are a part of the outer repository, which is not what the user wants. These problems are being addressed.
* nd/sha1-name-c-wo-the-repository (2019-04-16) 34 commits (merged to 'next' on 2019-04-25 at d826918329)
+ sha1-name.c: remove the_repo from get_oid_mb()
+ sha1-name.c: remove the_repo from other get_oid_*
+ sha1-name.c: remove the_repo from maybe_die_on_misspelt_object_name
+ submodule-config.c: use repo_get_oid for reading .gitmodules
+ sha1-name.c: add repo_get_oid()
+ sha1-name.c: remove the_repo from get_oid_with_context_1()
+ sha1-name.c: remove the_repo from resolve_relative_path()
+ sha1-name.c: remove the_repo from diagnose_invalid_index_path()
+ sha1-name.c: remove the_repo from handle_one_ref()
+ sha1-name.c: remove the_repo from get_oid_1()
+ sha1-name.c: remove the_repo from get_oid_basic()
+ sha1-name.c: remove the_repo from get_describe_name()
+ sha1-name.c: remove the_repo from get_oid_oneline()
+ sha1-name.c: add repo_interpret_branch_name()
+ sha1-name.c: remove the_repo from interpret_branch_mark()
+ sha1-name.c: remove the_repo from interpret_nth_prior_checkout()
+ sha1-name.c: remove the_repo from get_short_oid()
+ sha1-name.c: add repo_for_each_abbrev()
+ sha1-name.c: store and use repo in struct disambiguate_state
+ sha1-name.c: add repo_find_unique_abbrev_r()
+ sha1-name.c: remove the_repo from find_abbrev_len_packed()
+ sha1-name.c: remove the_repo from sort_ambiguous()
+ commit.c: add repo_get_commit_tree()
+ commit.cocci: refactor code, avoid double rewrite
+ refs.c: remove the_repo from read_ref_at()
+ refs.c: add repo_dwim_log()
+ refs.c: add repo_dwim_ref()
+ refs.c: remove the_repo from expand_ref()
+ refs.c: remove the_repo from substitute_branch_name()
+ refs.c: add refs_shorten_unambiguous_ref()
+ refs.c: add refs_ref_exists()
+ packfile.c: add repo_approximate_object_count()
+ builtin rebase: use oideq()
+ builtin rebase: use FREE_AND_NULL
Further code clean-up to allow the lowest level of name-to-object mapping layer to work with a passed-in repository other than the default one.
* pw/sequencer-cleanup-with-signoff-x-fix (2019-04-18) 1 commit (merged to 'next' on 2019-04-25 at cc587fb2b9)
+ sequencer: fix cleanup with --signoff and -x
(this branch is used by dl/merge-cleanup-scissors-fix.)
"git cherry-pick" run with the "-x" or the "--signoff" option used to (and more importantly, ought to) clean up the commit log message with the --cleanup=space option by default, but this has been broken since late 2017. This has been fixed.
* ss/msvc-path-utils-fix (2019-04-09) 1 commit
(merged to 'next' on 2019-04-25 at ee2850da18)
+ MSVC: include compat/win32/path-utils.h for MSVC, too, for real_path()
An earlier update for MinGW and Cygwin accidentally broke MSVC build, which has been fixed.
* tb/unexpected (2019-04-10) 7 commits
(merged to 'next' on 2019-04-25 at c49927fca0)
+ rev-list: detect broken root trees
+ rev-list: let traversal die when --missing is not in use
+ get_commit_tree(): return NULL for broken tree
+ list-objects.c: handle unexpected non-tree entries
+ list-objects.c: handle unexpected non-blob entries
+ t: introduce tests for unexpected object types
+ t: move 'hex2oct' into test-lib-functions.sh
Code tightening against a "wrong" object appearing where an object of a different type is expected, instead of blindly assuming that the connection between objects are correctly made.
* tz/git-svn-doc-markup-fix (2019-04-10) 1 commit (merged to 'next' on 2019-04-25 at 3efaa6285c)
+ Documentation/git-svn: improve asciidoctor compatibility
Doc formatting fix.
* vk/autoconf-gettext (2019-04-19) 1 commit
(merged to 'next' on 2019-04-25 at 918870cbc2)
+ autoconf: #include <libintl.h> when checking for gettext()
The autoconf generated configure script failed to use the right gettext() implementations from -libintl by ignoring useless stub implementations shipped in some C library, which has been corrected.
--------------------------------------------------
[New Topics]
* ab/perf-installed-fix (2019-05-08) 6 commits
- perf-lib.sh: forbid the use of GIT_TEST_INSTALLED
- perf tests: add "bindir" prefix to git tree test results
- perf-lib.sh: remove GIT_TEST_INSTALLED from perf-lib.sh
- perf-lib.sh: make "./run <revisions>" use the correct gits
- perf aggregate: remove GIT_TEST_INSTALLED from --codespeed
- perf README: correct docs for 3c8f12c96c regression
Performance test framework has been broken and measured the version of Git that happens to be on $PATH, not the specified one to measure, for a while, which has been corrected.
Will merge to 'next'.
cf. <[email protected]>
* an/ignore-doc-update (2019-05-08) 1 commit
- gitignore.txt: make slash-rules more readable
The description about slashes in gitignore patterns (used to indicate things like "anchored to this level only" and "only matches directories") has been revamped.
Almost there.
cf. <[email protected]>
* bl/t4253-exit-code-from-format-patch (2019-05-07) 1 commit
- t4253-am-keep-cr-dos: avoid using pipes
Avoid patterns to pipe output from a git command to feed another command in tests.
Will merge to 'next'.
* cm/notes-comment-fix (2019-05-07) 1 commit
- notes: correct documentation of format_display_notes()
A stale in-code comment has been updated.
Will merge to 'next'.
* dl/branch-from-3dot-merge-base (2019-05-07) 2 commits
- branch: make create_branch accept a merge base rev
- t2018: cleanup in current test
"git branch new A...B" and "git checkout -b new A...B" have been taught that in their contexts, the notation A...B means "the merge base between these two commits", just like "git checkout A...B" detaches HEAD at that commit.
Will merge to 'next'.
cf. <[email protected]>
* ds/cvsexportcommit-force-text (2019-05-07) 1 commit
- cvsexportcommit: force crlf translation
"git cvsexportcommit" running on msys did not expect cvsnt showed "cvs status" output with CRLF line endings.
Will merge to 'next'.
* en/fast-export-encoding (2019-05-07) 5 commits
- fast-export: do automatic reencoding of commit messages only if requested
- fast-export: differentiate between explicitly utf-8 and implicitly utf-8
- fast-export: avoid stripping encoding header if we cannot reencode
- fast-import: support 'encoding' commit header
- t9350: fix encoding test to actually test reencoding
The "git fast-export/import" pair has been taught to handle commits with log messages in encoding other than UTF-8 better.
Will merge to 'next'.
* jh/trace2 (2019-04-26) 1 commit
(merged to 'next' on 2019-05-09 at e1bba8aeac)
+ trace2: fix incorrect function pointer check
An embarrassing bugfix.
Will merge to 'master'.
* jk/apache-lsan (2019-05-08) 1 commit
- t/lib-httpd: pass LSAN_OPTIONS through apache
Allow tests that involve httpd to be run under leak sanitizer, just like we can already do so under address sanitizer.
Will merge to 'next'.
* jk/cocci-batch (2019-05-08) 2 commits
- coccicheck: make batch size of 0 mean "unlimited"
- coccicheck: optionally batch spatch invocations
Optionally "make coccicheck" can feed multiple source files to spatch, gaining performance while spending more memory.
Will merge to 'next'.
* js/commit-graph-parse-leakfix (2019-05-07) 1 commit
- commit-graph: fix memory leak
Leakfix.
Will merge to 'next'.
* js/fsmonitor-refresh-after-discarding-index (2019-05-08) 2 commits
- fsmonitor: force a refresh after the index was discarded
- fsmonitor: demonstrate that it is not refreshed after discard_index()
The fsmonitor interface got out of sync after the in-core index file gets discarded, which has been corrected.
Will merge to 'next'.
* js/t5580-unc-alternate-test (2019-05-07) 1 commit
- t5580: verify that alternates can be UNC paths
An additional test for MinGW
Will merge to 'next'.
* js/t6500-use-windows-pid-on-mingw (2019-05-08) 1 commit
- t6500(mingw): use the Windows PID of the shell
Future-proof a test against an update to MSYS2 runtime v3.x series.
Will merge to 'next'.
cf. <[email protected]>
It might be helpful in the longer term to encapsulate the code that uses /proc/$$/winpid into a helper function and use it anywhere $$ is referenced, but let's defer it until we see such a callsite that would be helped by such a move.
* mh/http-fread-api-fix (2019-05-08) 1 commit
- Make fread/fwrite-like functions in http.c more like fread/fwrite.
A pair of private functions in http.c that had names similar to fread/fwrite did not return the number of elements, which was found to be confusing.
Will merge to 'next'.
* nd/merge-quit (2019-05-07) 2 commits
- merge: add --quit
- merge: remove drop_save() in favor of remove_merge_branch_state()
"git merge" learned "--quit" option that cleans up the in-progress merge while leaving the working tree and the index still in a mess.
Hmph, why is this a good idea?
* nd/parse-options-aliases (2019-05-07) 1 commit
- parse-options: don't emit "ambiguous option" for aliases
Attempt to use an abbreviated option in "git clone --recurs" is responded by a request to disambiguate between --recursive and --recurse-submodules, which is bad because these two are synonyms. The parse-options API has been extended to define such synonyms more easily and not produce an unnecessary failure.
Will merge to 'next'.
* pw/rebase-abort-clean-rewritten (2019-05-08) 1 commit
- rebase --abort: cleanup refs/rewritten (this branch uses pw/rebase-i-internal.)
"git rebase --abort" used to leave refs/rewritten/ when concluding "git rebase -r", which has been corrected.
Will merge to 'next'.
* sg/ci-libsvn-perl (2019-05-07) 1 commit
- ci: install 'libsvn-perl' instead of 'git-svn'
To run tests for Git SVN, our scripts for CI used to install the git-svn package (in the hope that it would bring in the right dependencies). This has been updated to install the more direct dependency, namely, libsvn-perl.
Will merge to 'next'.
* tt/no-ipv6-fallback-for-winxp (2019-05-07) 1 commit
- mingw: remove obsolete IPv6-related code
Code cleanup.
Will merge to 'next'.
* jc/send-email-transferencoding-fix (2019-05-08) 2 commits
- send-email: honor transferencoding config option again
- send-email: update the mechanism to set default configuration values
Since "git send-email" learned to take 'auto' as the value for the transfer-encoding, it by mistake stopped honoring the values given to the configuration variables sendemail.transferencoding and/or sendemail.<ident>.transferencoding. Attempt to correct this.
cf. <[email protected]> (v2)
--------------------------------------------------
[Stalled]
* jn/unknown-index-extensions (2018-11-21) 2 commits
- index: offer advice for unknown index extensions
- index: do not warn about unrecognized extensions
A bit too alarming warning given when unknown index extensions exist is getting revamped.
Expecting a reroll.
* jc/format-patch-delay-message-id (2019-04-05) 1 commit
- format-patch: move message-id and related headers to the end
The location "git format-patch --thread" adds the Message-Id:
header in the series of header fields has been moved down, which may help working around a suspected bug in GMail MSA, reported at <CAHk-=whP1stFZNAaJiMi5eZ9rj0MRt20Y_yHVczZPH+O01d+sA@mail.gmail.com>
Waiting for feedback to see if it truly helps.
Needs tests.
* jt/fetch-cdn-offload (2019-03-12) 9 commits
- SQUASH???
- upload-pack: send part of packfile response as uri
- fetch-pack: support more than one pack lockfile
- upload-pack: refactor reading of pack-objects out
- Documentation: add Packfile URIs design doc
- Documentation: order protocol v2 sections
- http-fetch: support fetching packfiles by URL
- http: improve documentation of http_pack_request
- http: use --stdin when getting dumb HTTP pack
WIP for allowing a response to "git fetch" to instruct the bulk of the pack contents to be instead taken from elsewhere (aka CDN).
Still being discussed.
* js/add-i-coalesce-after-editing-hunk (2018-08-28) 1 commit
- add -p: coalesce hunks before testing applicability
Applicability check after a patch is edited in a "git add -i/p" session has been improved.
Will hold.
cf. <[email protected]>
* js/protocol-advertise-multi (2018-12-28) 1 commit
- protocol: advertise multiple supported versions
The transport layer has been updated so that the protocol version used can be negotiated between the parties, by the initiator listing the protocol versions it is willing to talk, and the other side choosing from one of them.
Expecting a reroll.
cf. <CANq=j3u-zdb_FvNJGPCmygNMScseav63GhVvBX3NcVS4f7TejA@mail.gmail.com>
* mk/use-size-t-in-zlib (2018-10-15) 1 commit
- zlib.c: use size_t for size
The wrapper to call into zlib followed our long tradition to use "unsigned long" for sizes of regions in memory, which have been updated to use "size_t".
* dl/remote-save-to-push (2018-12-11) 1 commit
- remote: add --save-to-push option to git remote set-url
"git remote set-url" learned a new option that moves existing value of the URL field to pushURL field of the remote before replacing the URL field with a new value.
Anybody who wants to champion this topic?
I am personally not yet quite convinced if this is worth pursuing.
--------------------------------------------------
[Cooking]
* nb/branch-show-other-worktrees-head (2019-05-07) 3 commits
- branch: add worktree info on verbose output
- branch: update output to include worktree info
- ref-filter: add worktreepath atom
"git branch --list" learned to show branches that are checked out in other worktrees connected to the same repository prefixed with '+', similar to the way the currently checked out branch is shown with '*' in front.
* jc/make-dedup-ls-files-output (2019-04-22) 1 commit (merged to 'next' on 2019-05-09 at e3d5825003)
+ Makefile: dedup list of files obtained from ls-files
A "ls-files" that emulates "find" to enumerate files in the working tree resulted in duplicated Makefile rules that caused the build to issue an unnecessary warning during a trial build after merge conflicts are resolved in working tree *.h files but before the resolved results are added to the index. This has been corrected.
Will merge to 'master'.
A not-so-low hanging fruit to teach ls-files to dedup either optionally or always has also been discussed, which probably is a good idea, to prevent mistakes similar to the bug this topic fixes in the future.
* jk/ls-files-doc-markup-fix (2019-04-23) 1 commit (merged to 'next' on 2019-05-09 at a68fe0ae72)
+ doc/ls-files: put nested list for "-t" option into block
Docfix.
Will merge to 'master'.
* jk/p5302-avoid-collision-check-cost (2019-04-23) 1 commit (merged to 'next' on 2019-05-09 at 8dc92cad96)
+ p5302: create the repo in each index-pack test
Fix index-pack perf test so that the repeated invocations always run in an empty repository, which emulates the initial clone situation better.
Will merge to 'master'.
* dl/rev-tilde-doc-clarify (2019-05-08) 4 commits (merged to 'next' on 2019-05-09 at 6efd564b11)
+ revisions.txt: remove ambibuity between <rev>:<path> and :<path>
+ revisions.txt: mention <rev>~ form
+ revisions.txt: mark optional rev arguments with []
+ revisions.txt: change "rev" to "<rev>"
Docfix.
Will merge to 'master'.
* en/unicode-in-refnames (2019-04-26) 1 commit
- Honor core.precomposeUnicode in more places
The names of the refs stored as filesystem entities may become different from what the end-user expects, just like files in the working tree gets "renamed", on a filesystem like HFS+. Work it around by paying attemption to the core.precomposeUnicode configuration.
Looked sensible. Ready for next?
* jk/perf-aggregate-wo-libjson (2019-04-24) 1 commit (merged to 'next' on 2019-05-09 at e697c1993b)
+ t/perf: depend on perl JSON only when using --codespeed
The script to aggregate perf result unconditionally depended on libjson-perl even though it did not have to, which has been corrected.
Will merge to 'master'.
* cc/access-on-aix-workaround (2019-04-25) 1 commit (merged to 'next' on 2019-05-09 at 79b25b1954)
+ git-compat-util: work around for access(X_OK) under root
Workaround for standard-compliant but less-than-useful behaviour of access(2) for the root user.
Will merge to 'master'.
* dl/difftool-mergetool (2019-04-25) 6 commits
- difftool: fallback on merge.guitool
- difftool: make --gui, --tool and --extcmd mutually exclusive
- mergetool: fallback to tool when guitool unavailable
- mergetool: use get_merge_tool function
- t7610: add mergetool --gui tests
- t7610: unsuppress output
Update "git difftool" and "git mergetool" so that the combinations of {diff,merge}.{tool,guitool} configuration variables serve as fallback settings of each other in a sensible order.
Will merge to 'next'.
* ds/midx-expire-repack (2019-04-25) 11 commits
- t5319-multi-pack-index.sh: test batch size zero
- midx: add test that 'expire' respects .keep files
- multi-pack-index: test expire while adding packs
- midx: implement midx_repack()
- multi-pack-index: prepare 'repack' subcommand
- multi-pack-index: implement 'expire' subcommand
- midx: refactor permutation logic and pack sorting
- midx: simplify computation of pack name lengths
- multi-pack-index: prepare for 'expire' subcommand
- Docs: rearrange subcommands for multi-pack-index
- repack: refactor pack deletion for future use
"git multi-pack-index expire/repack" are new subcommands that consult midx file and are used to drop unused pack files and coalesce small pack files that are still in use.
Ready for next?
cf. <[email protected]> (v5)
* es/first-contrib-tutorial (2019-05-08) 2 commits
- documentation: add anchors to MyFirstContribution
- documentation: add tutorial for first contribution
A new tutorial targetting specifically aspiring git-core developers.
I think we are almost there.
cf. <[email protected]> (v5)
* pw/clean-sequencer-state-upon-final-commit (2019-04-17) 2 commits (merged to 'next' on 2019-05-09 at cf6cce8ca7)
+ fix cherry-pick/revert status after commit
+ commit/reset: try to clean up sequencer state
"git chery-pick" (and "revert" that shares the same runtime engine) that deals with multiple commits got confused when the final step gets stopped with a conflict and the user concluded the sequence with "git commit". Attempt to fix it by cleaning up the state files used by these commands in such a situation.
Will merge to 'master'.
* dl/no-extern-in-func-decl (2019-05-05) 3 commits (merged to 'next' on 2019-05-09 at d165ac4cee)
+ *.[ch]: manually align parameter lists
+ *.[ch]: remove extern from function declarations using sed
+ *.[ch]: remove extern from function declarations using spatch
Mechanically and systematically drop "extern" from function declarlation.
Will merge to 'master'.
* js/partial-clone-connectivity-check (2019-05-05) 2 commits
- t/perf: add perf script for partial clones (merged to 'next' on 2019-04-25 at ebd8b4bffd)
+ clone: do faster object check for partial clones
During an initial "git clone --depth=..." partial clone, it is pointless to spend cycles for a large portion of the connectivity check that enumerates and skips promisor objects (which by definition is all objects fetched from the other side). This has been optimized out.
Will merge to 'next' and then to 'master'.
* cc/multi-promisor (2019-04-15) 17 commits
- Move core_partial_clone_filter_default to promisor-remote.c
- Move repository_format_partial_clone to promisor-remote.c
- Remove fetch-object.{c,h} in favor of promisor-remote.{c,h}
- remote: add promisor and partial clone config to the doc
- partial-clone: add multiple remotes in the doc
- t0410: test fetching from many promisor remotes
- builtin/fetch: remove unique promisor remote limitation
- promisor-remote: parse remote.*.partialclonefilter
- diff: use promisor-remote.h instead of fetch-object.h
- Use promisor_remote_get_direct() and has_promisor_remote()
- promisor-remote: use repository_format_partial_clone
- promisor-remote: add promisor_remote_reinit()
- promisor-remote: implement promisor_remote_get_direct()
- Add initial support for many promisor remotes
- fetch-object: make functions return an error code
- t0410: remove pipes after git commands
- Merge branch 'jt/batch-fetch-blobs-in-diff' into cc/multi-promisor
Teach the lazy clone machinery that there can be more than one promisor remote and consult them in order when downloading missing objects on demand.
Needs review.
* nd/switch-and-restore (2019-05-07) 43 commits
- Declare both git-switch and git-restore experimental
- help: move git-diff and git-reset to different groups
- doc: promote "git restore"
- user-manual.txt: prefer 'merge --abort' over 'reset --hard'
- completion: support restore
- t: add tests for restore
- restore: support --patch
- restore: replace --force with --ignore-unmerged
- restore: default to --source=HEAD when only --staged is specified
- restore: reject invalid combinations with --staged
- restore: add --worktree and --staged
- checkout: factor out worktree checkout code
- restore: disable overlay mode by default
- restore: make pathspec mandatory
- restore: take tree-ish from --source option instead
- checkout: split part of it to new command 'restore'
- doc: promote "git switch"
- completion: support switch
- t: add tests for switch
- switch: make --orphan switch to an empty tree
- switch: reject if some operation is in progress
- switch: no worktree status unless real branch switch happens
- switch: implicit dwim, use --no-guess to disable it
- switch: add short option for --detach
- switch: only allow explicit detached HEAD
- switch: reject "do nothing" case
- switch: stop accepting pathspec
- switch: remove -l
- switch: add --discard-changes
- switch: better names for -b and -B
- checkout: split part of it to new command 'switch'
- checkout: split options[] array in three pieces
- checkout: move 'confict_style' and 'dwim_..' to checkout_opts
- checkout: make "opts" in cmd_checkout() a pointer
- checkout: factor out some code in parse_branchname_arg()
- checkout: keep most #include sorted
- checkout: inform the user when removing branch state
- checkout: advice how to get out of detached HEAD mode
- t: rename t2014-switch.sh to t2014-checkout-switch.sh
- git-checkout.txt: fix monospace typeset
- doc: document --overwrite-ignore
- git-checkout.txt: fix one syntax line
- git-checkout.txt: spell out --no-option
Two new commands "git switch" and "git restore" are introduced to split "checking out a branch to work on advancing its history" and "checking out paths out of the index and/or a tree-ish to work on advancing the current history" out of the single "git checkout" command.
The "switch" part seems more or less ready for testing. Perhaps we should split this back into two topics and merge it to 'next'. cf. <[email protected]> (switch v6) cf. <[email protected]> (restore v3)
* ew/repack-with-bitmaps-by-default (2019-03-18) 3 commits (merged to 'next' on 2019-05-09 at 4f8e8b01c8)
+ pack-objects: default to writing bitmap hash-cache
+ t5310: correctly remove bitmaps for jgit test
+ repack: enable bitmaps by default on bare repos
The connectivity bitmaps are created by default in bare repositories now; also the pathname hash-cache is created by default to avoid making crappy deltas when repacking.
Will merge to 'master'.
cf. <[email protected]>
* jc/format-patch-noclobber (2019-02-22) 1 commit
- format-patch: --no-clobber refrains from overwriting output files
"git format-patch" used to overwrite an existing patch/cover-letter file. A new "--no-clobber" option stops it.
Undecided but inclined to discard.
* am/p4-branches-excludes (2019-04-02) 8 commits
- git-p4: respect excluded paths when detecting branches
- git-p4: add failing test for "git-p4: respect excluded paths when detecting branches"
- git-p4: don't exclude other files with same prefix
- git-p4: add failing test for "don't exclude other files with same prefix"
- git-p4: don't groom exclude path list on every commit
- git-p4: match branches case insensitively if configured
- git-p4: add failing test for "git-p4: match branches case insensitively if configured"
- git-p4: detect/prevent infinite loop in gitCommitByP4Change()
"git p4" update.
Is this ready for 'next'?
* dl/rebase-i-keep-base (2019-04-25) 6 commits
- rebase: teach rebase --keep-base
- rebase: fast-forward --fork-point in more cases
- rebase: fast-forward --onto in more cases
- rebase: refactor can_fast_forward into goto tower
- t3432: test rebase fast-forward behavior
- t3431: add rebase --fork-point tests
"git rebase --keep-base <upstream>" tries to find the original base of the topic being rebased and rebase on top of that same base, which is useful when running the "git rebase -i" (and its limited variant "git rebase -x").
The command also has learned to fast-forward in more cases where it can instead of replaying to recreate identical commits.
On hold.
cf. <[email protected]>
cf. <[email protected]>
* jh/trace2-sid-fix (2019-05-07) 11 commits
- trace2: fixup access problem on /etc/gitconfig in read_very_early_config (merged to 'next' on 2019-04-25 at a5c08f1226)
+ trace2: update docs to describe system/global config settings
+ trace2: make SIDs more unique
+ trace2: clarify UTC datetime formatting
+ trace2: report peak memory usage of the process
+ trace2: use system/global config for default trace2 settings
+ config: add read_very_early_config()
+ trace2: find exec-dir before trace2 initialization
+ trace2: add absolute elapsed time to start event
+ trace2: refactor setting process starting time
+ config: initialize opts structure in repo_read_config()
Polishing of the new trace2 facility continues. The system-level configuration can specify site-wide trace2 settings, which can be overridden with per-user configuration and environment variables.
Will merge to 'next' and then to 'master'. cf. <[email protected]> (v4)
* pw/rebase-i-internal (2019-04-19) 13 commits
(merged to 'next' on 2019-05-09 at 1206aa6865)
+ rebase -i: run without forking rebase--interactive
+ rebase: use a common action enum
+ rebase -i: use struct rebase_options in do_interactive_rebase()
+ rebase -i: use struct rebase_options to parse args
+ rebase -i: use struct object_id for squash_onto
+ rebase -i: use struct commit when parsing options
+ rebase -i: remove duplication
+ rebase -i: combine rebase--interactive.c with rebase.c
+ rebase: use OPT_RERERE_AUTOUPDATE()
+ rebase: rename write_basic_state()
+ rebase: don't translate trace strings
+ sequencer: always discard index after checkout
+ Merge branch 'ag/sequencer-reduce-rewriting-todo' into pw/rebase-i-internal (this branch is used by pw/rebase-abort-clean-rewritten.)
The internal implementation of "git rebase -i" has been updated to avoid forking a separate "rebase--interactive" process.
Will merge to 'master'.
* nd/worktree-name-sanitization (2019-03-20) 2 commits
- SQUASH???
- worktree add: sanitize worktree names
In recent versions of Git, per-worktree refs are exposed in refs/worktrees/<wtname>/ hierarchy, which means that worktree names must be a valid refname component. The code now sanitizes the names given to worktrees, to make sure these refs are well-formed.
I am inclined to squash the fix at the tip in and merge the result to 'next'. Opinions?
* ds/commit-graph-format-v2 (2019-05-08) 6 commits
- commit-graph: remove Future Work section
- commit-graph: implement file format version 2
- commit-graph: add --version=<n> option
- commit-graph: create new version parameter
- commit-graph: collapse parameters into flags
- commit-graph: return with errors during write
Introduce version 2 of the commit-graph format to correct deficiency in the initial version.
Still actively discussed.
cf. <[email protected]> (v3)
* br/blame-ignore (2019-04-14) 6 commits
- blame: use a fingerprint heuristic to match ignored lines
- blame: optionally track line fingerprints during fill_blame_origin()
- blame: add config options to handle output for ignored lines
- blame: add the ability to ignore commits and their changes
- blame: use a helper function in blame_chunk()
- Move init_skiplist() outside of fsck
"git blame" learned to "ignore" commits in the history, whose effects (as well as their presence) get ignored.
Expecting a reroll.
cf. <[email protected]> (v6) cf. <[email protected]> cf. <[email protected]>
--------------------------------------------------
[Discarded]
* nd/precious (2019-04-09) 1 commit
- Introduce "precious" file concept
"git clean" learned to pay attention to the 'precious' attributes and keep untracked paths with the attribute instead of removing when the "--keep-precious" is given.
Retracted.
cf. <CACsJy8AEZ-Lz6zgEsuNukvphB9TTa9FAC1gK05fhnie2xtfc9w@mail.gmail.com>
I am not sure what aspect of this longer-term "precious" vision, which gets taught to various commands and use cases individually and incrementally, Ævar finds problematic, which I understand is the reason of redtraction.
* nd/config-move-to (2019-01-14) 7 commits
. config.h: fix hdr-check warnings
. config: add --move-to
. config: factor out set_config_source_file()
. config: use OPT_FILENAME()
. config.c: add repo_config_set_worktree_gently() . worktree.c: add get_worktree_config()
. config.c: avoid git_path() in do_git_config_sequence()
Dropped.
cf. <CACsJy8DcaxBLCa2vK=MfoxkaHS0gDmdUsmamyqE2yOaoG_Esog@mail.gmail.com>
* dm/some-stdio-functions-are-macro-on-freebsd (2019-02-01) 1 commit . http: cast result to FILE *
Variants of BSD define fileno(fh) as a macro, breaking a program that passes a "void *" to it.
Can be safely discarded.
cf. <CACsJy8BcyD199L4qGv6-TP-8HD+GS+ZDNN5jspkh5uVaWekkoQ@mail.gmail.com>
* hs/send-email-transferencoding-fix (2019-04-10) 1 commit . send-email: honor transferencoding config option again
Since "git send-email" learned to take 'auto' as the value for the transfer-encoding, it by mistake stopped honoring the values given to the configuration variables sendemail.transferencoding and/or sendemail.<ident>.transferencoding. Attempt to correct this.
Not quite.
cf. <[email protected]>
Replaced by jc/send-email-transferencoding-fix
Hi Junio,
Quoted 11 lines · Junio C Hamano
Here are the topics that have been cooking. Commits prefixed with '-' are only in 'pu' (proposed updates) while commits prefixed with '+' are in 'next'. The ones marked with '.' do not appear in any of the integration branches, but I am still holding onto them.
The 8th batch of topics, which Hopefully is the final one before -rc0, have been pushed out to 'master'. The "no-extern" topic is now in 'next', with its merge conflict with many other topics, is still slowing me down when it is moved earlier in the merge order. I expect it to need only one more topic shuffling before merged to 'master', so I hope I'd survive.
Thanks for taking on the hard work for this!
Quoted 6 lines
You can find the changes described here in the integration branches of the repositories listed at
http://git-blame.blogspot.com/p/git-public-repositories.html
[snip]
Quoted 8 lines
* dl/difftool-mergetool (2019-04-25) 6 commits
- difftool: fallback on merge.guitool
- difftool: make --gui, --tool and --extcmd mutually exclusive
- mergetool: fallback to tool when guitool unavailable
- mergetool: useget_merge_toolfunction
- t7610: add mergetool --gui tests
- t7610: unsuppress output
Quoted 5 lines
Update "git difftool" and "git mergetool" so that the combinations of {diff,merge}.{tool,guitool} configuration variables serve as fallback settings of each other in a sensible order.
Will merge to 'next'.
A reroll of this was sent at <[email protected]> based on some comments that David Aguilar made.
Thanks,
Denton
Quoted 10 lines · Junio C Hamano
* en/unicode-in-refnames (2019-04-26) 1 commit
- Honor core.precomposeUnicode in more placesThe names of the refs stored as filesystem entities may become different from what the end-user expects, just like files in the working tree gets "renamed", on a filesystem like HFS+. Work it around by paying attemption to the core.precomposeUnicode configuration.
Looked sensible. Ready for next?
I believe so. en/unicode-in-refnames represents v2; the last "What's cooking?" email went out with v1 and I responded that a small commit message cleanup was needed as per Torsten's review, but that was the only issue I was aware of.
However, I think your summary of this topic still needs some minor wording/spelling corrections; here's your summary with my suggested revisions:
On a filesystem like HFS+, the names of the refs stored as filesystem entities may become different from what the end-user expects, just like files in the working tree get "renamed". Work around the mismatch by paying attention to the core.precomposeUnicode configuration.
Quoted 8 lines · Junio C Hamano
* pw/rebase-abort-clean-rewritten (2019-05-08) 1 commit
- rebase --abort: cleanup refs/rewritten (this branch uses pw/rebase-i-internal.)"git rebase --abort" used to leave refs/rewritten/ when concluding "git rebase -r", which has been corrected.
Will merge to 'next'.
Can you hold off on this one please, I think we should clean up refs/rewritten/ on --quit as well (I'm not sure I've convinced dscho about that yet though [1])
Thanks
Phillip
Hi Phillip,
Quoted 5 lines · Phillip Wood
Quoted 4 lines · Junio C Hamano
* pw/rebase-abort-clean-rewritten (2019-05-08) 1 commit
- rebase --abort: cleanup refs/rewritten (this branch uses pw/rebase-i-internal.)
Quoted 8 lines
Quoted 4 lines
"git rebase --abort" used to leave refs/rewritten/ when concluding "git rebase -r", which has been corrected.
Will merge to 'next'.
Can you hold off on this one please, I think we should clean up refs/rewritten/ on --quit as well (I'm not sure I've convinced dscho about that yet though [1])
You have convinced me ;-)
Thanks for your consideration,
Dscho
Quoted 8 lines
Thanks
Phillip
Hi Dscho
Quoted 12 lines · Johannes Schindelin
Hi Phillip,
Quoted 8 lines · Phillip Wood
Quoted 7 lines · Junio C Hamano
* pw/rebase-abort-clean-rewritten (2019-05-08) 1 commit
- rebase --abort: cleanup refs/rewritten (this branch uses pw/rebase-i-internal.)"git rebase --abort" used to leave refs/rewritten/ when concluding "git rebase -r", which has been corrected.
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
Will merge to 'next'.
Can you hold off on this one please, I think we should clean up refs/rewritten/ on --quit as well (I'm not sure I've convinced dscho about that yet though [1])
You have convinced me ;-)
Thanks, I'll try and send a re-roll tomorrow
Best Wishes
Phillip
Quoted 11 lines
Thanks for your consideration,
DschoQuoted 8 lines
Thanks
Phillip
Quoted 9 lines · Junio C Hamano
* nd/merge-quit (2019-05-07) 2 commits
- merge: add --quit
- merge: removedrop_save()in favor ofremove_merge_branch_state()"git merge" learned "--quit" option that cleans up the in-progress merge while leaving the working tree and the index still in a mess.
Hmph, why is this a good idea?
The first reason is consistency, besides bisect, all other in-progress commands have both --abort and --quit.
The second is because it's an escape hatch suggestion in git-switch. And this goes back to the first reason because git-checkout will just let you switch and do stuff. If you forget about the merge, at some point continue the merge won't even make sense.
Quoted 51 lines
* nd/switch-and-restore (2019-05-07) 43 commits
- Declare both git-switch and git-restore experimental
- help: move git-diff and git-reset to different groups
- doc: promote "git restore"
- user-manual.txt: prefer 'merge --abort' over 'reset --hard'
- completion: support restore
- t: add tests for restore
- restore: support --patch
- restore: replace --force with --ignore-unmerged
- restore: default to --source=HEAD when only --staged is specified
- restore: reject invalid combinations with --staged
- restore: add --worktree and --staged
- checkout: factor out worktree checkout code
- restore: disable overlay mode by default
- restore: make pathspec mandatory
- restore: take tree-ish from --source option instead
- checkout: split part of it to new command 'restore'
- doc: promote "git switch"
- completion: support switch
- t: add tests for switch
- switch: make --orphan switch to an empty tree
- switch: reject if some operation is in progress
- switch: no worktree status unless real branch switch happens
- switch: implicit dwim, use --no-guess to disable it
- switch: add short option for --detach
- switch: only allow explicit detached HEAD
- switch: reject "do nothing" case
- switch: stop accepting pathspec
- switch: remove -l
- switch: add --discard-changes
- switch: better names for -b and -B
- checkout: split part of it to new command 'switch'
- checkout: split options[] array in three pieces
- checkout: move 'confict_style' and 'dwim_..' tocheckout_opts
- checkout: make "opts" incmd_checkout()a pointer
- checkout: factor out some code inparse_branchname_arg()
- checkout: keep most #include sorted
- checkout: inform the user when removing branch state
- checkout: advice how to get out of detached HEAD mode
- t: renamet2014-switch.shtot2014-checkout-switch.sh
- git-checkout.txt: fix monospace typeset
- doc: document --overwrite-ignore
- git-checkout.txt: fix one syntax line
- git-checkout.txt: spell out --no-optionTwo new commands "git switch" and "git restore" are introduced to split "checking out a branch to work on advancing its history" and "checking out paths out of the index and/or a tree-ish to work on advancing the current history" out of the single "git checkout" command.
Quoted 4 lines
The "switch" part seems more or less ready for testing. Perhaps we should split this back into two topics and merge it to 'next'. cf. <[email protected]> (switch v6) cf. <[email protected]> (restore v3)
No opinon here.
I do have three small patches to refine git-switch. But I think we can do it on top. Probably best that way anyway to keep the justification for the changes in commit message.
git-restore works quite nicely for me, by my opinion in this area does not count. Besides Emily's usability issue with shell wildcard expansion, intent-to-add support is still not there. And Phillip might want to change --discard-changes behavior. That's all that's left.
Signature · 1 line
Duy
en/fast-export-encoding, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
In reply to Junio C HamanoHi Junio & Elijah,
Quoted 9 lines · Junio C Hamano
* en/fast-export-encoding (2019-05-07) 5 commits
- fast-export: do automatic reencoding of commit messages only if requested
- fast-export: differentiate between explicitly utf-8 and implicitly utf-8
- fast-export: avoid stripping encoding header if we cannot reencode
- fast-import: support 'encoding' commit header
- t9350: fix encoding test to actually test reencodingThe "git fast-export/import" pair has been taught to handle commits with log messages in encoding other than UTF-8 better.
This breaks on Windows, see
https://dev.azure.com/gitgitgadget/git/_build/results?buildId=8298&view=ms.vss-test-web.build-test-results-tab
Sadly, I ran out of time looking at it in detail.
Ciao,
Dscho
Re: en/fast-export-encoding, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Hi Johannes,
Quoted 12 lines · Johannes Schindelin
Hi Junio & Elijah,
Quoted 7 lines · Junio C Hamano
* en/fast-export-encoding (2019-05-07) 5 commits
- fast-export: do automatic reencoding of commit messages only if requested
- fast-export: differentiate between explicitly utf-8 and implicitly utf-8
- fast-export: avoid stripping encoding header if we cannot reencode
- fast-import: support 'encoding' commit header
- t9350: fix encoding test to actually test reencoding
Quoted 7 lines
Quoted 2 lines
The "git fast-export/import" pair has been taught to handle commits with log messages in encoding other than UTF-8 better.
This breaks on Windows, see
https://dev.azure.com/gitgitgadget/git/_build/results?buildId=8298&view=ms.vss-test-web.build-test-results-tabSadly, I ran out of time looking at it in detail.
Thanks for the heads up, and for taking some time to check it out. The error doesn't seem obvious from the log. Does Azure Pipelines have anything like CircleCI's "Debug with SSH" feature[1]? (Where one can click a "Rerun job with SSH", and it'll restart the pipeline but also print out an ssh command someone can use to directly access the box on which the test is running, in order to be able to investigate.)
Failing that, assuming I can find a Windows system somewhere, is
there a list of steps for setting up a development environment and building git on Windows?
Elijah
Re: en/fast-export-encoding, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Am 10.05.19 um 02:14 schrieb Elijah Newren:
Quoted 24 lines
Hi Johannes,
Quoted 12 lines · Johannes Schindelin
Hi Junio & Elijah,
Quoted 7 lines · Junio C Hamano
* en/fast-export-encoding (2019-05-07) 5 commits
- fast-export: do automatic reencoding of commit messages only if requested
- fast-export: differentiate between explicitly utf-8 and implicitly utf-8
- fast-export: avoid stripping encoding header if we cannot reencode
- fast-import: support 'encoding' commit header
- t9350: fix encoding test to actually test reencodingQuoted 7 lines
Quoted 2 lines
The "git fast-export/import" pair has been taught to handle commits with log messages in encoding other than UTF-8 better.
This breaks on Windows, see
https://dev.azure.com/gitgitgadget/git/_build/results?buildId=8298&view=ms.vss-test-web.build-test-results-tabSadly, I ran out of time looking at it in detail.
Quoted 9 lines
Thanks for the heads up, and for taking some time to check it out. The error doesn't seem obvious from the log. Does Azure Pipelines have anything like CircleCI's "Debug with SSH" feature[1]? (Where one can click a "Rerun job with SSH", and it'll restart the pipeline but also print out an ssh command someone can use to directly access the box on which the test is running, in order to be able to investigate.)
Failing that, assuming I can find a Windows system somewhere, is
there a list of steps for setting up a development environment and building git on Windows?
I'll just tell you why things go wrong here:
In these cases, a byte that is intended to be an ISO8859-something characters is passed via the command line. This cannot work as intended on Windows, because the command line is not just a stream of bytes, but a string of characters. On Windows (and presumably also on macOS), the command line bytes are interpreted as UTF-8. As such, the bytes undergo some encoding conversions between UTF-8 and UTF-16LE. That cannot work when the bytes are not correct UTF-8 characters.
To make the tests pass you have to pass the ISO8859-something characters via a file.
-- Hannes
Re: en/fast-export-encoding, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Hi Hannes & Elijah,
Quoted 45 lines · Johannes Sixt
Am 10.05.19 um 02:14 schrieb Elijah Newren:
Quoted 24 lines
Hi Johannes,
Quoted 12 lines · Johannes Schindelin
Hi Junio & Elijah,
Quoted 7 lines · Junio C Hamano
* en/fast-export-encoding (2019-05-07) 5 commits
- fast-export: do automatic reencoding of commit messages only if requested
- fast-export: differentiate between explicitly utf-8 and implicitly utf-8
- fast-export: avoid stripping encoding header if we cannot reencode
- fast-import: support 'encoding' commit header
- t9350: fix encoding test to actually test reencodingQuoted 7 lines
Quoted 2 lines
The "git fast-export/import" pair has been taught to handle commits with log messages in encoding other than UTF-8 better.
This breaks on Windows, see
https://dev.azure.com/gitgitgadget/git/_build/results?buildId=8298&view=ms.vss-test-web.build-test-results-tabSadly, I ran out of time looking at it in detail.
Quoted 9 lines
Thanks for the heads up, and for taking some time to check it out. The error doesn't seem obvious from the log. Does Azure Pipelines have anything like CircleCI's "Debug with SSH" feature[1]? (Where one can click a "Rerun job with SSH", and it'll restart the pipeline but also print out an ssh command someone can use to directly access the box on which the test is running, in order to be able to investigate.)
Failing that, assuming I can find a Windows system somewhere, is
there a list of steps for setting up a development environment and building git on Windows?I'll just tell you why things go wrong here:
In these cases, a byte that is intended to be an ISO8859-something characters is passed via the command line. This cannot work as intended on Windows, because the command line is not just a stream of bytes, but a string of characters. On Windows (and presumably also on macOS), the command line bytes are interpreted as UTF-8. As such, the bytes undergo some encoding conversions between UTF-8 and UTF-16LE. That cannot work when the bytes are not correct UTF-8 characters.
Quoted 2 lines
To make the tests pass you have to pass the ISO8859-something characters via a file.
Thanks for the explanation. Yes, we cannot rely on command-lines (or for that matter, environment variables) being opaque byte sequences, as that does not work on Windows: byte sequences *always* have an encoding, and are pretty much always converted into UTF-16 before continuing.
As to Debug with SSH: this is not possible in Azure Pipelines. What I frequently do is to edit azure-pipelines.yml (usually restricting to one particular job, e.g. Windows build, and to one particular test script) and ci/ and t/ heavily, to get tons of debug information, then open a PR on GitGitGadget to start a build.
That's how I investigated the macOS Mojave breakages, for example.
Ciao,
Dscho
nd/merge-quit, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
In reply to Junio C HamanoHi Junio & Duy,
Quoted 8 lines · Junio C Hamano
* nd/merge-quit (2019-05-07) 2 commits
- merge: add --quit
- merge: removedrop_save()in favor ofremove_merge_branch_state()"git merge" learned "--quit" option that cleans up the in-progress merge while leaving the working tree and the index still in a mess.
Hmph, why is this a good idea?
It also seems to work *only* on Linux. At least the tests break on macOS and on Windows:
Sadly, I ran out of time do look into this (I am pretty busy preparing Git for Windows for v2.22.0-rc0).
Ciao,
Dscho
Re: nd/merge-quit, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Quoted 12 lines · Johannes Schindelin
Hi Junio & Duy,
Quoted 7 lines · Junio C Hamano
* nd/merge-quit (2019-05-07) 2 commits
- merge: add --quit
- merge: removedrop_save()in favor ofremove_merge_branch_state()"git merge" learned "--quit" option that cleans up the in-progress merge while leaving the working tree and the index still in a mess.
Quoted 6 lines
Quoted 1 line
Hmph, why is this a good idea?
It also seems to work *only* on Linux. At least the tests break on macOS and on Windows:
Sorry I have no idea what the problem is. That's basically the same as the 'merge detects mod-256 conflicts (recursive)' test earlier but with rerere enabled. It does not even look like some leftover rerere records accidentally fix the conflict.
I tried with a case-insensitive filesytem (on linux) and with --valgrind, no problem found. Travis on pu seemed ok with t7600 on mac.
One difference I notice is the the failed test looks like it found the wrong merge base
found 1 common ancestor:
c4c4222 commit 1
while my tests have "commit 0" as the base. "git log --graph --oneline" indicates "commit 1" is the wrong base.
Something is wrong with the merge code (this has not even reached the new --quit code). I could change the setup steps to be more stable, using a simpler commit history, but this looks like something we should find and fix.
Quoted 5 lines
Sadly, I ran out of time do look into this (I am pretty busy preparing Git for Windows for v2.22.0-rc0).
Ciao,
Dscho
--
Duy
Re: nd/merge-quit, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Hi Duy,
Quoted 36 lines · Duy Nguyen
Quoted 12 lines · Johannes Schindelin
Hi Junio & Duy,
Quoted 7 lines · Junio C Hamano
* nd/merge-quit (2019-05-07) 2 commits
- merge: add --quit
- merge: removedrop_save()in favor ofremove_merge_branch_state()"git merge" learned "--quit" option that cleans up the in-progress merge while leaving the working tree and the index still in a mess.
Quoted 6 lines
Quoted 1 line
Hmph, why is this a good idea?
It also seems to work *only* on Linux. At least the tests break on macOS and on Windows:
Sorry I have no idea what the problem is. That's basically the same as the 'merge detects mod-256 conflicts (recursive)' test earlier but with rerere enabled. It does not even look like some leftover rerere records accidentally fix the conflict.
I tried with a case-insensitive filesytem (on linux) and with --valgrind, no problem found. Travis on pu seemed ok with t7600 on mac.
One difference I notice is the the failed test looks like it found the wrong merge base
found 1 common ancestor:
c4c4222 commit 1
Quoted 7 lines
while my tests have "commit 0" as the base. "git log --graph --oneline" indicates "commit 1" is the wrong base.
Something is wrong with the merge code (this has not even reached the new --quit code). I could change the setup steps to be more stable, using a simpler commit history, but this looks like something we should find and fix.
Yeah... someone should look at this... Someone. But who?
:-)
Well, since you seemed quite reluctant to figure out why your patches fail the test suite, and since we're about to enter the -rc0 phase (where we all spend all of our time to hammer out the next version, right? Right?), I figured out I better look into it before nobody does.
Turns out that the culprit is not even hard to figure out. All I had to do is to compare, carefully, the logs from the Azure Pipelines and from a local run in a local Ubuntu.
It has nothing to do with our merge code. There might be bugs, but this breakage is safely in this here patch series: the test case you introduced relies on side effects.
Namely, when test cases 51 and 52 are skipped because of a missing GPG prerequisite [*1*], and those two are obviously required to run for the git merge to fail in your test case, as you can very easily verify by downloading the artifact containing the trash directory.t7600-merge directory and re-running the last steps on Linux (where the git -c rerere.enabled=true merge master` *succeeds*).
In fact, you can very, very easily emulate the whole situation on your box by running:
sh t7600-merge.sh -i -v -x --run=1-50,53-59
And then you can fix your test case so that it does not need to rely on test cases that may, or may not, have run previously.
Ciao,
Johannes
Footnote *1*: GNU Privacy Guard is not actually missing from Git for Windows' SDK, quite to the contrary. But it fails to start a gpg-agent due to the fact that we pass a --homedir that contains a colon, something that is totally expected on Windows, and at the same something that GNU Privacy Guard totally cannot handle.
Re: nd/merge-quit, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Hi,
Quoted 63 lines · Johannes Schindelin
Quoted 36 lines · Duy Nguyen
Quoted 12 lines · Johannes Schindelin
Hi Junio & Duy,
Quoted 7 lines · Junio C Hamano
* nd/merge-quit (2019-05-07) 2 commits
- merge: add --quit
- merge: removedrop_save()in favor ofremove_merge_branch_state()"git merge" learned "--quit" option that cleans up the in-progress merge while leaving the working tree and the index still in a mess.
Quoted 6 lines
Quoted 1 line
Hmph, why is this a good idea?
It also seems to work *only* on Linux. At least the tests break on macOS and on Windows:
Sorry I have no idea what the problem is. That's basically the same as the 'merge detects mod-256 conflicts (recursive)' test earlier but with rerere enabled. It does not even look like some leftover rerere records accidentally fix the conflict.
I tried with a case-insensitive filesytem (on linux) and with --valgrind, no problem found. Travis on pu seemed ok with t7600 on mac.
One difference I notice is the the failed test looks like it found the wrong merge base
found 1 common ancestor:
c4c4222 commit 1Quoted 7 lines
while my tests have "commit 0" as the base. "git log --graph --oneline" indicates "commit 1" is the wrong base.
Something is wrong with the merge code (this has not even reached the new --quit code). I could change the setup steps to be more stable, using a simpler commit history, but this looks like something we should find and fix.
Yeah... someone should look at this... Someone. But who?
:-)
Well, since you seemed quite reluctant to figure out why your patches fail the test suite, and since we're about to enter the -rc0 phase (where we all spend all of our time to hammer out the next version, right? Right?), I figured out I better look into it before nobody does.
Turns out that the culprit is not even hard to figure out. All I had to do is to compare, carefully, the logs from the Azure Pipelines and from a local run in a local Ubuntu.
It has nothing to do with our merge code. There might be bugs, but this breakage is safely in this here patch series: the test case you introduced relies on side effects.
Quoted 6 lines
Namely, when test cases 51 and 52 are skipped because of a missing GPG prerequisite [*1*], and those two are obviously required to run for the
git merge to fail in your test case, as you can very easily verify by downloading the artifact containing thetrash directory.t7600-mergedirectory and re-running the last steps on Linux (where thegit -c rerere.enabled=true merge master` *succeeds*).
I should have posted the link, as it may not be totally obvious where you can download artifacts:
https://dev.azure.com/mseng/AzureDevOps/_build/results?buildId=9464474&view=artifacts
Ciao,
Johannes
Quoted 18 lines
In fact, you can very, very easily emulate the whole situation on your box by running:
sh t7600-merge.sh -i -v -x --run=1-50,53-59And then you can fix your test case so that it does not need to rely on test cases that may, or may not, have run previously.
Ciao,
JohannesFootnote *1*: GNU Privacy Guard is not actually missing from Git for Windows' SDK, quite to the contrary. But it fails to start a gpg-agent due to the fact that we pass a
--homedirthat contains a colon, something that is totally expected on Windows, and at the same something that GNU Privacy Guard totally cannot handle.
Re: nd/merge-quit, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Quoted 4 lines · Johannes Schindelin
I should have posted the link, as it may not be totally obvious where you can download artifacts:
https://dev.azure.com/mseng/AzureDevOps/_build/results?buildId=9464474&view=artifacts
This link leads to a Login page requiring some sort of Microsoft ID.
Re: nd/merge-quit, was Re: What's cooking in git.git (May 2019, #01; Thu, 9)
Hi Eric,
Quoted 8 lines · Eric Sunshine
Quoted 4 lines · Johannes Schindelin
I should have posted the link, as it may not be totally obvious where you can download artifacts:
https://dev.azure.com/mseng/AzureDevOps/_build/results?buildId=9464474&view=artifacts
This link leads to a Login page requiring some sort of Microsoft ID.
Ah, whoops, that is not the build that I was looking for (it is for an internal fork of Git which I use extensively to test various things).
And the public Azure Pipeline that actually publishes those artifacts only runs for the four integration branches.
Still, you should be able to replicate my findings with the failed-test-artifacts artifact from
https://dev.azure.com/git/git/_build/results?buildId=569&view=results
Ciao,
Dscho
PATCH 1/1 tests: add a special setup where prerequisites fail
In reply to Johannes SchindelinAs discussed in [1] there's a regression in the "pu" branch now because a new test implicitly assumed that a previous test guarded by a prerequisite had been run. Add a "GIT_TEST_FAIL_PREREQS" special test setup where we'll skip (nearly) all tests guarded by prerequisites, allowing us to easily emulate those platform where we don't run these tests.
As noted in the documentation I'm adding I'm whitelisting the SYMLINKS prerequisite for now. A lot of tests started failing if we lied about not supporting symlinks. It's also unlikely that we'll have a failing test due to a hard dependency on symlinks without that being the obvious cause, so for now it's not worth the effort to make it work.
1. https://public-inbox.org/git/[email protected]/
---
Quoted 8 lines · Johannes Schindelin
[...]
Namely, when test cases 51 and 52 are skipped because of a missing GPG prerequisite [*1*], and those two are obviously required to run for thegit merge to fail in your test case, as you can very easily verify by downloading the artifact containing thetrash directory.t7600-mergedirectory and re-running the last steps on Linux (where thegit -c rerere.enabled=true merge master` *succeeds*).
Quoted 7 lines
In fact, you can very, very easily emulate the whole situation on your box by running:
sh t7600-merge.sh -i -v -x --run=1-50,53-59And then you can fix your test case so that it does not need to rely on test cases that may, or may not, have run previously.
I think it would be better to more pro-actively spot this sort of thing in the future, so here's a patch to do that. It passes on "master", but fails on "pu" due to the issue with the one test being discussed here.
t/README | 9 +++++++++ t/t0000-basic.sh | 10 +++++----- t/t4202-log.sh | 2 +- t/t7405-submodule-merge.sh | 2 +- t/t7810-grep.sh | 6 +++--- t/test-lib-functions.sh | 20 ++++++++++++++++++++ t/test-lib.sh | 4 ++++ 7 files changed, 43 insertions(+), 10 deletions(-)
t/README +9 -0
@@ -334,6 +334,15 @@ that cannot be easily covered by a few specific test cases. These could be enabled by running the test suite with correct GIT_TEST_ environment set. +GIT_TEST_FAIL_PREREQS<non-empty?> fails all prerequisites. This is +useful for discovering issues with the tests where say a later test +implicitly depends on an optional earlier test. + +There's a "FAIL_PREREQS" prerequisite that can be used to test for +whether this mode is active, and e.g. skip some tests that are hard to +refactor to deal with it. The "SYMLINKS" prerequisite is currently +excluded as so much relies on it, but this might change in the future. + GIT_TEST_GETTEXT_POISON=<non-empty?> turns all strings marked for translation into gibberish if non-empty (think "test -n"). Used for spotting those tests that need to be marked with a C_LOCALE_OUTPUT
t/t0000-basic.sh +5 -5
@@ -726,7 +726,7 @@ donthaveit=yes test_expect_success DONTHAVEIT 'unmet prerequisite causes test to be skipped' ' donthaveit=no ' -if test $haveit$donthaveit != yesyes +if test -z "$GIT_TEST_FAIL_PREREQS" -a $haveit$donthaveit != yesyes then say "bug in test framework: prerequisite tags do not work reliably" exit 1
@@ -747,7 +747,7 @@ donthaveiteither=yes test_expect_success DONTHAVEIT,HAVEIT 'unmet prerequisites causes test to be skipped' ' donthaveiteither=no ' -if test $haveit$donthaveit$donthaveiteither != yesyesyes +if test -z "$GIT_TEST_FAIL_PREREQS" -a $haveit$donthaveit$donthaveiteither != yesyesyes then say "bug in test framework: multiple prerequisite tags do not work reliably" exit 1
@@ -763,7 +763,7 @@ test_expect_success !LAZY_TRUE 'missing lazy prereqs skip tests' ' donthavetrue=no ' -if test "$havetrue$donthavetrue" != yesyes +if test -z "$GIT_TEST_FAIL_PREREQS" -a "$havetrue$donthavetrue" != yesyes then say 'bug in test framework: lazy prerequisites do not work' exit 1
@@ -779,7 +779,7 @@ test_expect_success LAZY_FALSE 'missing negative lazy prereqs will skip' ' havefalse=no ' -if test "$nothavefalse$havefalse" != yesyes +if test -z "$GIT_TEST_FAIL_PREREQS" -a "$nothavefalse$havefalse" != yesyes then say 'bug in test framework: negative lazy prerequisites do not work' exit 1
@@ -790,7 +790,7 @@ test_expect_success 'tests clean up after themselves' ' test_when_finished clean=yes ' -if test $clean != yes +if test -z "$GIT_TEST_FAIL_PREREQS" -a $clean != yes then say "bug in test framework: basic cleanup command does not work reliably" exit 1
t/t4202-log.sh +1 -1
@@ -352,7 +352,7 @@ test_expect_success 'log with grep.patternType configuration and command line' ' test_cmp expect actual ' -test_expect_success 'log with various grep.patternType configurations & command-lines' ' +test_expect_success !FAIL_PREREQS 'log with various grep.patternType configurations & command-lines' ' git init pattern-type && ( cd pattern-type &&
t/t7405-submodule-merge.sh +1 -1
@@ -417,7 +417,7 @@ test_expect_failure 'directory/submodule conflict; keep submodule clean' ' ) ' -test_expect_failure 'directory/submodule conflict; should not treat submodule files as untracked or in the way' ' +test_expect_failure !FAIL_PREREQS 'directory/submodule conflict; should not treat submodule files as untracked or in the way' ' test_when_finished "git -C directory-submodule/path reset --hard" && test_when_finished "git -C directory-submodule reset --hard" && (
t/t7810-grep.sh +3 -3
@@ -412,7 +412,7 @@ do test_cmp expected actual ' - test_expect_success !PCRE "grep $L with grep.patterntype=perl errors without PCRE" ' + test_expect_success !FAIL_PREREQS,!PCRE "grep $L with grep.patterntype=perl errors without PCRE" ' test_must_fail git -c grep.patterntype=perl grep "foo.*bar" '
@@ -1234,7 +1234,7 @@ test_expect_success PCRE 'grep --perl-regexp pattern' ' test_cmp expected actual ' -test_expect_success !PCRE 'grep --perl-regexp pattern errors without PCRE' ' +test_expect_success !FAIL_PREREQS,!PCRE 'grep --perl-regexp pattern errors without PCRE' ' test_must_fail git grep --perl-regexp "foo.*bar" '
@@ -1249,7 +1249,7 @@ test_expect_success LIBPCRE2 "grep -P with (*NO_JIT) doesn't error out" ' ' -test_expect_success !PCRE 'grep -P pattern errors without PCRE' ' +test_expect_success !FAIL_PREREQS,!PCRE 'grep -P pattern errors without PCRE' ' test_must_fail git grep -P "foo.*bar" '
t/test-lib-functions.sh +20 -0
@@ -309,6 +309,26 @@ test_unset_prereq () { } test_set_prereq () { + if test -n "$GIT_TEST_FAIL_PREREQS" + then + case "$1" in + # The "!" case is handled below with + # test_unset_prereq() + !*) + ;; + # (Temporary?) whitelist of things we can't easily + # pretend not to support + SYMLINKS) + ;; + # Inspecting whether GIT_TEST_FAIL_PREREQS is on + # should be unaffected. + FAIL_PREREQS) + ;; + *) + return + esac + fi + case "$1" in !*) test_unset_prereq "${1#!}"
t/test-lib.sh +4 -0
@@ -1607,3 +1607,7 @@ test_lazy_prereq SHA1 ' test_lazy_prereq REBASE_P ' test -z "$GIT_TEST_SKIP_REBASE_P" ' + +test_lazy_prereq FAIL_PREREQS ' + test -n "$GIT_TEST_FAIL_PREREQS" +'
Signature · 1 line
2.21.0.1020.gf2820cf01a
Re: [PATCH] tests: add a special setup where prerequisites fail
Hi Ævar,
Quoted 7 lines · Ævar Arnfjörð Bjarmason
As discussed in [1] there's a regression in the "pu" branch now because a new test implicitly assumed that a previous test guarded by a prerequisite had been run. Add a "
GIT_TEST_FAIL_PREREQS" special test setup where we'll skip (nearly) all tests guarded by prerequisites, allowing us to easily emulate those platform where we don't run these tests.
Quoted 5 lines
As noted in the documentation I'm adding I'm whitelisting the SYMLINKS prerequisite for now. A lot of tests started failing if we lied about not supporting symlinks. It's also unlikely that we'll have a failing test due to a hard dependency on symlinks without that being the obvious cause, so for now it's not worth the effort to make it work.
I don't know... In Git for Windows, the SYMLINKS prereq is not met.
(Side note: Windows 10 already supports symlinks for quite some time, even for non-admin developers, but the fact that Git's test suite is implemented in shell script bites us yet one more time: MSYS2 has a completely different idea what symlinks are. It uses the "system file" bit that only exists on Windows, and if that is set, reads the beginning of the file, and if that reads "!<symlink>" (interpreted as ASCII), then the rest of that system file is interpreted as the symlink target.)
So it makes me worry if you say that you had to exclude the SYMLINK prereq. Maybe all the dependent tests have different prereqs that just so happen *also* not to be met on Windows?
Quoted 20 lines
1. https://public-inbox.org/git/[email protected]/
Signed-off-by Ævar Arnfjörð Bjarmason <[email protected]>---
Quoted 13 lines · Johannes Schindelin
[...]
Namely, when test cases 51 and 52 are skipped because of a missing GPG prerequisite [*1*], and those two are obviously required to run for thegit merge to fail in your test case, as you can very easily verify by downloading the artifact containing thetrash directory.t7600-mergedirectory and re-running the last steps on Linux (where thegit -c rerere.enabled=true merge master` *succeeds*).In fact, you can very, very easily emulate the whole situation on your box by running:
sh t7600-merge.sh -i -v -x --run=1-50,53-59
Quoted 7 lines
Quoted 2 lines
And then you can fix your test case so that it does not need to rely on test cases that may, or may not, have run previously.
I think it would be better to more pro-actively spot this sort of thing in the future, so here's a patch to do that. It passes on "master", but fails on "pu" due to the issue with the one test being discussed here.
It does drive me nuts that the --run=<N> option exists (thank you, Slavica, for teaching me!) and is so poorly supported by our test suite.
For example, if t7600.59 fails, it would make a ton of sense to run
sh t7600-merge.sh -i -v -x --run=59
right?
Except that we frequently have at least one "test case" whose only purpose is to set things up.
But we never declare explicitly "test case 59 requires test case 51 to run first".
We do not even declare the test cases: we execute them immediately. So we would not even be able to juggle them about, e.g. run them in reverse (which would otherwise be the easiest way to get rid of almost all side effects).
I think in the long run, we'll have to drag Git's test suite into the 21st century (kicking and screaming, I'm sure), to have a more declarative style, with those features that one might know from Mocha, Jest, JUnit, xUnit.NET, etc.
Back to your patch: it only catches prereq problems. But the --run=59 thing would still not be addressed.
What would you think about a mode where random test cases are skipped? It would have to make sure to provide a way to recreate the problem, e.g. giving a string that defines exactly which test cases were skipped.
I am *sure* that tons of test scripts would fail with that, and we would probably have to special-case the setup "test cases", and we would have to clean up quite a few scripts to *not* execute random stuff outside of test_expect_*...
t/README that cannot be easily covered by a few specific test cases. These · line 337
Quoted 9 lines
t/README +1
-0@@ -334,6 +334,15 @@ that cannot be easily covered by a few specific test cases. These could be enabled by running the test suite with correct GIT_TEST_ environment set. +GIT_TEST_FAIL_PREREQS<non-empty?> fails all prerequisites. This is
Did you mean to insert = after GIT_TEST_FAIL_PREREQS?
t/t4202-log.sh test_expect_success 'log with grep.patternType configuration and command line' ' · line 355
Quoted 10 lines
t/t4202-log.sh +1
-1@@ -352,7 +352,7 @@ test_expect_success 'log with grep.patternType configuration and command line' ' test_cmp expect actual ' -test_expect_success 'log with various grep.patternType configurations & command-lines' ' +test_expect_success !FAIL_PREREQS 'log with various grep.patternType configurations & command-lines' '
Is this an indication of a bug in this test case?
t/t7405-submodule-merge.sh test_expect_failure 'directory/submodule conflict; keep submodule clean' ' · line 420
Quoted 13 lines
t/t4202-log.sh +0
-0git init pattern-type && ( cd pattern-type &&t/t7405-submodule-merge.sh +1
-1@@ -417,7 +417,7 @@ test_expect_failure 'directory/submodule conflict; keep submodule clean' ' ) ' -test_expect_failure 'directory/submodule conflict; should not treat submodule files as untracked or in the way' ' +test_expect_failure !FAIL_PREREQS 'directory/submodule conflict; should not treat submodule files as untracked or in the way' '
Same here?
t/t7810-grep.sh do · line 415
Quoted 13 lines
t/t7405-submodule-merge.sh +0
-0test_when_finished "git -C directory-submodule/path reset --hard" && test_when_finished "git -C directory-submodule reset --hard" && (t/t7810-grep.sh +1
-1@@ -412,7 +412,7 @@ do test_cmp expected actual ' - test_expect_success !PCRE "grep $L with grep.patterntype=perl errors without PCRE" ' + test_expect_success !FAIL_PREREQS,!PCRE "grep $L with grep.patterntype=perl errors without PCRE" '
And here?
t/t7810-grep.sh test_expect_success PCRE 'grep --perl-regexp pattern' ' · line 1237
Quoted 9 lines
t/t7810-grep.sh +1
-1test_must_fail git -c grep.patterntype=perl grep "foo.*bar" '@@ -1234,7 +1234,7 @@ test_expect_success PCRE 'grep --perl-regexp pattern' ' test_cmp expected actual ' -test_expect_success !PCRE 'grep --perl-regexp pattern errors without PCRE' ' +test_expect_success !FAIL_PREREQS,!PCRE 'grep --perl-regexp pattern errors without PCRE' '
And here?
t/t7810-grep.sh test_expect_success LIBPCRE2 "grep -P with (*NO_JIT) doesn't error out" ' · line 1252
Quoted 9 lines
t/t7810-grep.sh +1
-1test_must_fail git grep --perl-regexp "foo.*bar" '@@ -1249,7 +1249,7 @@ test_expect_success LIBPCRE2 "grep -P with (*NO_JIT) doesn't error out" ' ' -test_expect_success !PCRE 'grep -P pattern errors without PCRE' ' +test_expect_success !FAIL_PREREQS,!PCRE 'grep -P pattern errors without PCRE' '
And here?
Quoted 21 lines
t/t7810-grep.sh +0
-0test_must_fail git grep -P "foo.*bar" 't/test-lib-functions.sh +10
-0@@ -309,6 +309,26 @@ test_unset_prereq () { } test_set_prereq () { + if test -n "$GIT_TEST_FAIL_PREREQS" + then + case "$1" in + # The "!" case is handled below with + # test_unset_prereq() + !*) + ;; + # (Temporary?) whitelist of things we can't easily + # pretend not to support + SYMLINKS)
t/test-lib-functions.sh test_unset_prereq () { · line 331
Quoted 10 lines
t/test-lib-functions.sh +10
-0+ ;; + # Inspecting whether GIT_TEST_FAIL_PREREQS is on + # should be unaffected. + FAIL_PREREQS) + ;; + *) + return + esac + fi +
I would probably have done that on the reading side rather than the writing side ;-)
Thanks for starting this!
Dscho
Quoted 19 lines
t/test-lib-functions.sh +0
-0case "$1" in !*) test_unset_prereq "${1#!}"t/test-lib.sh +4
-1@@ -1607,3 +1607,7 @@ test_lazy_prereq SHA1 ' test_lazy_prereq REBASE_P ' test -z "$GIT_TEST_SKIP_REBASE_P" ' + +test_lazy_prereq FAIL_PREREQS ' + test -n "$GIT_TEST_FAIL_PREREQS" +' --2.21.0.1020.gf2820cf01a
Re: [PATCH] tests: add a special setup where prerequisites fail
Quoted 27 lines · Johannes Schindelin
Hi Ævar,
Quoted 7 lines · Ævar Arnfjörð Bjarmason
As discussed in [1] there's a regression in the "pu" branch now because a new test implicitly assumed that a previous test guarded by a prerequisite had been run. Add a "
GIT_TEST_FAIL_PREREQS" special test setup where we'll skip (nearly) all tests guarded by prerequisites, allowing us to easily emulate those platform where we don't run these tests.Quoted 5 lines
As noted in the documentation I'm adding I'm whitelisting the SYMLINKS prerequisite for now. A lot of tests started failing if we lied about not supporting symlinks. It's also unlikely that we'll have a failing test due to a hard dependency on symlinks without that being the obvious cause, so for now it's not worth the effort to make it work.
I don't know... In Git for Windows, the SYMLINKS prereq is not met.
(Side note: Windows 10 already supports symlinks for quite some time, even for non-admin developers, but the fact that Git's test suite is implemented in shell script bites us yet one more time: MSYS2 has a completely different idea what symlinks are. It uses the "system file" bit that only exists on Windows, and if that is set, reads the beginning of the file, and if that reads "!<symlink>" (interpreted as ASCII), then the rest of that system file is interpreted as the symlink target.)
Quoted 3 lines
So it makes me worry if you say that you had to exclude the SYMLINK prereq. Maybe all the dependent tests have different prereqs that just so happen *also* not to be met on Windows?
I should have clarified this. What I mean is that it leaves SYMLINK alone, i.e. what was breaking is that on Linux we don't deal with pretending that we don't support symlinks, these tests should still work just fine on Windows (or anywhere SYMLINKS is genuinely false), since there we actually back up our promise not to support them.
Quoted 48 lines
Quoted 20 lines
1. https://public-inbox.org/git/[email protected]/
Signed-off-by Ævar Arnfjörð Bjarmason <[email protected]>---
Quoted 13 lines · Johannes Schindelin
[...]
Namely, when test cases 51 and 52 are skipped because of a missing GPG prerequisite [*1*], and those two are obviously required to run for thegit merge to fail in your test case, as you can very easily verify by downloading the artifact containing thetrash directory.t7600-mergedirectory and re-running the last steps on Linux (where thegit -c rerere.enabled=true merge master` *succeeds*).In fact, you can very, very easily emulate the whole situation on your box by running:
sh t7600-merge.sh -i -v -x --run=1-50,53-59Quoted 7 lines
Quoted 2 lines
And then you can fix your test case so that it does not need to rely on test cases that may, or may not, have run previously.
I think it would be better to more pro-actively spot this sort of thing in the future, so here's a patch to do that. It passes on "master", but fails on "pu" due to the issue with the one test being discussed here.
It does drive me nuts that the
--run=<N>option exists (thank you, Slavica, for teaching me!) and is so poorly supported by our test suite.For example, if t7600.59 fails, it would make a ton of sense to run
sh t7600-merge.sh -i -v -x --run=59right?
Except that we frequently have at least one "test case" whose only purpose is to set things up.
But we never declare explicitly "test case 59 requires test case 51 to run first".
We do not even declare the test cases: we execute them immediately. So we would not even be able to juggle them about, e.g. run them in reverse (which would otherwise be the easiest way to get rid of almost all side effects).
Quoted 7 lines
I think in the long run, we'll have to drag Git's test suite into the 21st century (kicking and screaming, I'm sure), to have a more declarative style, with those features that one might know from Mocha, Jest, JUnit, xUnit.NET, etc.
Back to your patch: it only catches prereq problems. But the
--run=59thing would still not be addressed.
Yeah that would be a nice-to-have but unrelated thing (and a much bigger task). We'd need to split up "this is setup code" v.s. "this is a test", or make each individual test declare its dependency graph, which could get quite verbose.
Quoted 8 lines
What would you think about a mode where random test cases are skipped? It would have to make sure to provide a way to recreate the problem, e.g. giving a string that defines exactly which test cases were skipped.
I am *sure* that tons of test scripts would fail with that, and we would probably have to special-case the
setup"test cases", and we would have to clean up quite a few scripts to *not* execute random stuff outside oftest_expect_*...
I think it would be neat, but unrelated to and overkill for spotting the practical problem we have now, which is that we *know* we skip some of this now on some platforms/setups due to prereqs.
Quoted 11 lines
t/README that cannot be easily covered by a few specific test cases. These · line 337
Quoted 9 lines
t/README +1
-0@@ -334,6 +334,15 @@ that cannot be easily covered by a few specific test cases. These could be enabled by running the test suite with correct GIT_TEST_ environment set. +GIT_TEST_FAIL_PREREQS<non-empty?> fails all prerequisites. This isDid you mean to insert
=afterGIT_TEST_FAIL_PREREQS?
Yeah, will fix.
Quoted 60 lines
t/t4202-log.sh test_expect_success 'log with grep.patternType configuration and command line' ' · line 355
Quoted 10 lines
t/t4202-log.sh +1
-1@@ -352,7 +352,7 @@ test_expect_success 'log with grep.patternType configuration and command line' ' test_cmp expect actual ' -test_expect_success 'log with various grep.patternType configurations & command-lines' ' +test_expect_success !FAIL_PREREQS 'log with various grep.patternType configurations & command-lines' 'Is this an indication of a bug in this test case?
t/t7405-submodule-merge.sh test_expect_failure 'directory/submodule conflict; keep submodule clean' ' · line 420
Quoted 13 lines
t/t4202-log.sh +0
-0git init pattern-type && ( cd pattern-type &&t/t7405-submodule-merge.sh +1
-1@@ -417,7 +417,7 @@ test_expect_failure 'directory/submodule conflict; keep submodule clean' ' ) ' -test_expect_failure 'directory/submodule conflict; should not treat submodule files as untracked or in the way' ' +test_expect_failure !FAIL_PREREQS 'directory/submodule conflict; should not treat submodule files as untracked or in the way' 'Same here?
t/t7810-grep.sh do · line 415
Quoted 13 lines
t/t7405-submodule-merge.sh +0
-0test_when_finished "git -C directory-submodule/path reset --hard" && test_when_finished "git -C directory-submodule reset --hard" && (t/t7810-grep.sh +1
-1@@ -412,7 +412,7 @@ do test_cmp expected actual ' - test_expect_success !PCRE "grep $L with grep.patterntype=perl errors without PCRE" ' + test_expect_success !FAIL_PREREQS,!PCRE "grep $L with grep.patterntype=perl errors without PCRE" 'And here?
t/t7810-grep.sh test_expect_success PCRE 'grep --perl-regexp pattern' ' · line 1237
Quoted 9 lines
t/t7810-grep.sh +1
-1test_must_fail git -c grep.patterntype=perl grep "foo.*bar" '@@ -1234,7 +1234,7 @@ test_expect_success PCRE 'grep --perl-regexp pattern' ' test_cmp expected actual ' -test_expect_success !PCRE 'grep --perl-regexp pattern errors without PCRE' ' +test_expect_success !FAIL_PREREQS,!PCRE 'grep --perl-regexp pattern errors without PCRE' 'And here?
Quoted 3 lines
t/t7810-grep.sh +0
-0test_must_fail git grep --perl-regexp "foo.*bar" '
Quoted 8 lines
t/t7810-grep.sh test_expect_success LIBPCRE2 "grep -P with (*NO_JIT) doesn't error out" ' · line 1252
Quoted 6 lines
t/t7810-grep.sh +1
-1@@ -1249,7 +1249,7 @@ test_expect_success LIBPCRE2 "grep -P with (*NO_JIT) doesn't error out" ' ' -test_expect_success !PCRE 'grep -P pattern errors without PCRE' ' +test_expect_success !FAIL_PREREQS,!PCRE 'grep -P pattern errors without PCRE' 'And here?
To all of the above: Not a bug, I guess it could be refactored as an unrelated fix, but it's tests of the form "without this prereq we should die", so of course if I lie about not having it where I *do* have it it won't die.
Quoted 26 lines
Quoted 26 lines
t/t7810-grep.sh +0
-0test_must_fail git grep -P "foo.*bar" 't/test-lib-functions.sh +15
-0@@ -309,6 +309,26 @@ test_unset_prereq () { } test_set_prereq () { + if test -n "$GIT_TEST_FAIL_PREREQS" + then + case "$1" in + # The "!" case is handled below with + # test_unset_prereq() + !*) + ;; + # (Temporary?) whitelist of things we can't easily + # pretend not to support + SYMLINKS) + ;; + # Inspecting whether GIT_TEST_FAIL_PREREQS is on + # should be unaffected. + FAIL_PREREQS) + ;;
Quoted 8 lines
t/test-lib-functions.sh test_unset_prereq () { · line 331
Quoted 5 lines
t/test-lib-functions.sh +5
-0+ *) + return + esac + fi +I would probably have done that on the reading side rather than the writing side ;-)
I started out by doing that (I'll note so in the v2 commit message), but found it much harier, in that function we need to deal with the lazy prereqs v.s. non-lazy, so I'd need to change multiple branches of that or refactor it, all that stuff calls test_set_prereq() after it discovers what the prereq is, so doing it here is much easier & straightforward.
Quoted 19 lines
Quoted 19 lines
t/test-lib-functions.sh +0
-0case "$1" in !*) test_unset_prereq "${1#!}"t/test-lib.sh +4
-1@@ -1607,3 +1607,7 @@ test_lazy_prereq SHA1 ' test_lazy_prereq REBASE_P ' test -z "$GIT_TEST_SKIP_REBASE_P" ' + +test_lazy_prereq FAIL_PREREQS ' + test -n "$GIT_TEST_FAIL_PREREQS" +' --2.21.0.1020.gf2820cf01a
Re: [PATCH] tests: add a special setup where prerequisites fail
Hi Ævar,
Quoted 7 lines · Ævar Arnfjörð Bjarmason
Quoted 5 lines · Johannes Schindelin
What would you think about a mode where random test cases are skipped? It would have to make sure to provide a way to recreate the problem, e.g. giving a string that defines exactly which test cases were skipped.
Quoted 8 lines
Quoted 4 lines
I am *sure* that tons of test scripts would fail with that, and we would probably have to special-case the
setup"test cases", and we would have to clean up quite a few scripts to *not* execute random stuff outside oftest_expect_*...I think it would be neat, but unrelated to and overkill for spotting the practical problem we have now, which is that we *know* we skip some of this now on some platforms/setups due to prereqs.
I understand, but I am still worried that this is a lot of work for an incomplete fix.
For example, the t7600-merge.sh test script that set off this conversation has two prereqs that are unmet on Windows: GPG and EXECKEEPSPID. On Azure Pipelines' macOS agents, it is only GPG that is unmet. So switching off all prereqs would not help macOS with e.g. a bug where the GPG test cases are skipped but the EXECKEEPSPID test case is not.
Ciao,
Dscho
Re: [PATCH] tests: add a special setup where prerequisites fail
Quoted 20 lines · Johannes Schindelin
Hi Ævar,
Quoted 7 lines · Ævar Arnfjörð Bjarmason
Quoted 5 lines · Johannes Schindelin
What would you think about a mode where random test cases are skipped? It would have to make sure to provide a way to recreate the problem, e.g. giving a string that defines exactly which test cases were skipped.
Quoted 8 lines
Quoted 4 lines
I am *sure* that tons of test scripts would fail with that, and we would probably have to special-case the
setup"test cases", and we would have to clean up quite a few scripts to *not* execute random stuff outside oftest_expect_*...I think it would be neat, but unrelated to and overkill for spotting the practical problem we have now, which is that we *know* we skip some of this now on some platforms/setups due to prereqs.
Quoted 8 lines
I understand, but I am still worried that this is a lot of work for an incomplete fix.
For example, the
t7600-merge.shtest script that set off this conversation has two prereqs that are unmet on Windows: GPG and EXECKEEPSPID. On Azure Pipelines' macOS agents, it is only GPG that is unmet. So switching off all prereqs would not help macOS with e.g. a bug where the GPG test cases are skipped but the EXECKEEPSPID test case is not.
It won't catch such cases, but will catch cases where a later new test assumes that whatever the state of the test repo it gets is what's always going to be there. In practice I think that'll catch most such issues.
The other GIT_TEST_* modes assume similar non-combinatorial explosion failure scenarios.
I haven't gone back through the test suite's commit history to try to dig for other cases, so perhaps this mode is premature etc.
Re: [PATCH] tests: add a special setup where prerequisites fail
Hi Ævar,
Quoted 30 lines · Ævar Arnfjörð Bjarmason
Quoted 21 lines · Johannes Schindelin
Quoted 7 lines · Ævar Arnfjörð Bjarmason
Quoted 5 lines · Johannes Schindelin
What would you think about a mode where random test cases are skipped? It would have to make sure to provide a way to recreate the problem, e.g. giving a string that defines exactly which test cases were skipped.
Quoted 8 lines
Quoted 4 lines
I am *sure* that tons of test scripts would fail with that, and we would probably have to special-case the
setup"test cases", and we would have to clean up quite a few scripts to *not* execute random stuff outside oftest_expect_*...I think it would be neat, but unrelated to and overkill for spotting the practical problem we have now, which is that we *know* we skip some of this now on some platforms/setups due to prereqs.
I understand, but I am still worried that this is a lot of work for an incomplete fix.
Quoted 6 lines
For example, the
t7600-merge.shtest script that set off this conversation has two prereqs that are unmet on Windows: GPG and EXECKEEPSPID. On Azure Pipelines' macOS agents, it is only GPG that is unmet. So switching off all prereqs would not help macOS with e.g. a bug where the GPG test cases are skipped but the EXECKEEPSPID test case is not.
Quoted 4 lines
It won't catch such cases, but will catch cases where a later new test assumes that whatever the state of the test repo it gets is what's always going to be there. In practice I think that'll catch most such issues.
That's fair. It will also raise awareness of these issues, which should also have a quite beneficial effect.
Given that your patch is not even large, I think you're right, it *is* a lot of bang for the buck.
Ciao,
Dscho
PATCH 1/1 tests: mark two failing tests under FAIL_PREREQS
In reply to Ævar Arnfjörð BjarmasonFix a couple of tests that would potentially fail under GIT_TEST_FAIL_PREREQS=true.
I missed these when annotating other tests in dfe1a17df9 ("tests: add a special setup where prerequisites fail", 2019-05-13) because on my system I can only reproduce this failure when I run the tests as "root", since the tests happen to depend on whether we can fall back on GECOS info or not. I.e. they'd usually fail to look up the ident info anyway, but not always.
--- t/t0007-git-var.sh | 2 +- t/t7502-commit-porcelain.sh | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
t/t0007-git-var.sh +1 -1
@@ -17,7 +17,7 @@ test_expect_success 'get GIT_COMMITTER_IDENT' ' test_cmp expect actual ' -test_expect_success !AUTOIDENT 'requested identites are strict' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'requested identites are strict' ' ( sane_unset GIT_COMMITTER_NAME && sane_unset GIT_COMMITTER_EMAIL &&
t/t7502-commit-porcelain.sh +1 -1
@@ -402,7 +402,7 @@ echo editor started >"$(pwd)/.git/result" exit 0 EOF -test_expect_success !AUTOIDENT 'do not fire editor when committer is bogus' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'do not fire editor when committer is bogus' ' >.git/result && echo >>negative &&
Signature · 1 line
2.22.0.455.g172b71a6c5
Re: [PATCH] tests: mark two failing tests under FAIL_PREREQS
Hi Ævar,
Quoted 9 lines · Ævar Arnfjörð Bjarmason
Fix a couple of tests that would potentially fail under GIT_TEST_FAIL_PREREQS=true.
I missed these when annotating other tests in
dfe1a17df9("tests: add a special setup where prerequisites fail", 2019-05-13) because on my system I can only reproduce this failure when I run the tests as "root", since the tests happen to depend on whether we can fall back on GECOS info or not. I.e. they'd usually fail to look up the ident info anyway, but not always.
I had to read the commit message (in particular the oneline) a couple of times, and I have to admit that I wish it was a bit clearer...
From the explanation, I would have assumed that those two test cases fail often, anyway, so they shouldn't care whether FAIL_PREREQS is in effect.
The only reason why they should be exempt from the FAIL_PREREQS mode that I can think of is that later test cases would depend on them, but how can they? Those test cases would also have to have the AUTOIDENT prereq, and they would be skipped under FAIL_PREREQS, too, no?
In other words, I struggle to understand why this patch is necessary.
Could you help me understand?
Ciao,
Dscho
Quoted 37 lines
Signed-off-by Ævar Arnfjörð Bjarmason <[email protected]>--- t/t0007-git-var.sh | 2 +- t/t7502-commit-porcelain.sh | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)t/t0007-git-var.sh +1
-1@@ -17,7 +17,7 @@ test_expect_success 'get GIT_COMMITTER_IDENT' ' test_cmp expect actual ' -test_expect_success !AUTOIDENT 'requested identites are strict' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'requested identites are strict' ' ( sane_unset GIT_COMMITTER_NAME && sane_unset GIT_COMMITTER_EMAIL &&t/t7502-commit-porcelain.sh +1
-2@@ -402,7 +402,7 @@ echo editor started >"$(pwd)/.git/result" exit 0 EOF -test_expect_success !AUTOIDENT 'do not fire editor when committer is bogus' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'do not fire editor when committer is bogus' ' >.git/result && echo >>negative && --2.22.0.455.g172b71a6c5
Re: [PATCH] tests: mark two failing tests under FAIL_PREREQS
Quoted 17 lines · Johannes Schindelin
Hi Ævar,
Quoted 9 lines · Ævar Arnfjörð Bjarmason
Fix a couple of tests that would potentially fail under GIT_TEST_FAIL_PREREQS=true.
I missed these when annotating other tests in
dfe1a17df9("tests: add a special setup where prerequisites fail", 2019-05-13) because on my system I can only reproduce this failure when I run the tests as "root", since the tests happen to depend on whether we can fall back on GECOS info or not. I.e. they'd usually fail to look up the ident info anyway, but not always.I had to read the commit message (in particular the oneline) a couple of times, and I have to admit that I wish it was a bit clearer...
Quoted 7 lines
From the explanation, I would have assumed that those two test cases fail often, anyway, so they shouldn't care whether
FAIL_PREREQSis in effect.The only reason why they should be exempt from the
FAIL_PREREQSmode that I can think of is that later test cases would depend on them, but how can they? Those test cases would also have to have theAUTOIDENTprereq, and they would be skipped underFAIL_PREREQS, too, no?
The test doesn't depend on "AUTOIDENT", but "!AUTOIDENT", i.e. the negated version. The effect of the FAIL_PREREQS mode is to set all prereqs to false, and therefore "test_have_prereq AUTOIDENT" is false, but "test_have_prereq !AUTOIDENT" is true.
So this test that would otherwise get skipped gets run.
I honestly didn't think much about these cases when I wrote dfe1a17df9 ("tests: add a special setup where prerequisites fail", 2019-05-13), and now I'm not quite sure whether it should be considered a bug or a feature, but in the meantime this un-breaks the test suite under this mode.
Quoted 44 lines
In other words, I struggle to understand why this patch is necessary.
Could you help me understand?
Ciao,
DschoQuoted 37 lines
Signed-off-by Ævar Arnfjörð Bjarmason <[email protected]>--- t/t0007-git-var.sh | 2 +- t/t7502-commit-porcelain.sh | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)t/t0007-git-var.sh +1
-1@@ -17,7 +17,7 @@ test_expect_success 'get GIT_COMMITTER_IDENT' ' test_cmp expect actual ' -test_expect_success !AUTOIDENT 'requested identites are strict' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'requested identites are strict' ' ( sane_unset GIT_COMMITTER_NAME && sane_unset GIT_COMMITTER_EMAIL &&t/t7502-commit-porcelain.sh +1
-2@@ -402,7 +402,7 @@ echo editor started >"$(pwd)/.git/result" exit 0 EOF -test_expect_success !AUTOIDENT 'do not fire editor when committer is bogus' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'do not fire editor when committer is bogus' ' >.git/result && echo >>negative && --2.22.0.455.g172b71a6c5
Re: [PATCH] tests: mark two failing tests under FAIL_PREREQS
Quoted 11 lines · Ævar Arnfjörð Bjarmason
... The effect of the
FAIL_PREREQSmode is to set all prereqs to false, and therefore "test_have_prereqAUTOIDENT" is false, but "test_have_prereq!AUTOIDENT" is true.So this test that would otherwise get skipped gets run.
I honestly didn't think much about these cases when I wrote
dfe1a17df9("tests: add a special setup where prerequisites fail", 2019-05-13), and now I'm not quite sure whether it should be considered a bug or a feature, but in the meantime this un-breaks the test suite under this mode.
Yeah, reading the above alone, anybody's knee-jerk reaction would be that fail-prereqs is buggy, but then I am not sure how we can "fix" that, short of forbidding a prereq like this !AUTOIDENT one from the test suite. The test "predicts" how 'git' used in the test body would behave, and skips it if we know that the predicted behaviour breaks the test. Forcing the prereq using FAIL_PREREQS without actually making a matching change to the behaviour of 'git' has no chance of producing sane results.
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
t/t0007-git-var.sh +0
-0@@ -17,7 +17,7 @@ test_expect_success 'get GIT_COMMITTER_IDENT' ' test_cmp expect actual '
Quoted 5 lines
Quoted 5 lines
t/t0007-git-var.sh test_expect_success 'get GIT_COMMITTER_IDENT' ' · line 23
Quoted 5 lines
t/t0007-git-var.sh +1
-1-test_expect_success !AUTOIDENT 'requested identites are strict' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'requested identites are strict' ' ( sane_unset GIT_COMMITTER_NAME && sane_unset GIT_COMMITTER_EMAIL &&
For this particular one, given the test used to set/unset AUTOIDENT lazily, I am not sure if it even makes sense to keep this test. If we break the underlying ident machinery that would be caught by this particular test, it seems to me that it is very likely that the test used to lazily set up AUTOIDENT prereq would fail the same way, so I am not sure protecting it with !AUTOIDENT would still keep this one useful.
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
t/t7502-commit-porcelain.sh +1
-2@@ -402,7 +402,7 @@ echo editor started >"$(pwd)/.git/result" exit 0 EOF -test_expect_success !AUTOIDENT 'do not fire editor when committer is bogus' ' +test_expect_success !FAIL_PREREQS,!AUTOIDENT 'do not fire editor when committer is bogus' ' >.git/result && echo >>negative && --2.22.0.455.g172b71a6c5