| 1 | Git 2.28 Release Notes |
| 2 | ====================== |
| 3 | |
| 4 | Updates since v2.27 |
| 5 | ------------------- |
| 6 | |
| 7 | Backward compatibility notes |
| 8 | |
| 9 | * "fetch.writeCommitGraph" is deemed to be still a bit too risky and |
| 10 | is no longer part of the "feature.experimental" set. |
| 11 | |
| 12 | |
| 13 | UI, Workflows & Features |
| 14 | |
| 15 | * The commands in the "diff" family learned to honor "diff.relative" |
| 16 | configuration variable. |
| 17 | |
| 18 | * The check in "git fsck" to ensure that the tree objects are sorted |
| 19 | still had corner cases it missed unsorted entries. |
| 20 | |
| 21 | * The interface to redact sensitive information in the trace output |
| 22 | has been simplified. |
| 23 | |
| 24 | * The command line completion (in contrib/) learned to complete |
| 25 | options that the "git switch" command takes. |
| 26 | |
| 27 | * "git diff" used to take arguments in random and nonsense range |
| 28 | notation, e.g. "git diff A..B C", "git diff A..B C...D", etc., |
| 29 | which has been cleaned up. |
| 30 | |
| 31 | * "git diff-files" has been taught to say paths that are marked as |
| 32 | intent-to-add are new files, not modified from an empty blob. |
| 33 | |
| 34 | * "git status" learned to report the status of sparse checkout. |
| 35 | |
| 36 | * "git difftool" has trouble dealing with paths added to the index |
| 37 | with the intent-to-add bit. |
| 38 | |
| 39 | * "git fast-export --anonymize" learned to take customized mapping to |
| 40 | allow its users to tweak its output more usable for debugging. |
| 41 | |
| 42 | * The command line completion support (in contrib/) used to be |
| 43 | prepared to work with "set -u" but recent changes got a bit more |
| 44 | sloppy. This has been corrected. |
| 45 | |
| 46 | * "git gui" now allows opening work trees from the start-up dialog. |
| 47 | |
| 48 | |
| 49 | Performance, Internal Implementation, Development Support etc. |
| 50 | |
| 51 | * Code optimization for a common case. |
| 52 | (merge 8777616e4d an/merge-single-strategy-optim later to maint). |
| 53 | |
| 54 | * We've adopted a convention that any on-stack structure can be |
| 55 | initialized to have zero values in all fields with "= { 0 }", |
| 56 | even when the first field happens to be a pointer, but sparse |
| 57 | complained that a null pointer should be spelled NULL for a long |
| 58 | time. Start using -Wno-universal-initializer option to squelch |
| 59 | it (the latest sparse has it on by default). |
| 60 | |
| 61 | * "git log -L..." now takes advantage of the "which paths are touched |
| 62 | by this commit?" info stored in the commit-graph system. |
| 63 | |
| 64 | * As FreeBSD is not the only platform whose regexp library reports |
| 65 | a REG_ILLSEQ error when fed invalid UTF-8, add logic to detect that |
| 66 | automatically and skip the affected tests. |
| 67 | |
| 68 | * "git bugreport" learns to report what shell is in use. |
| 69 | |
| 70 | * Support for GIT_CURL_VERBOSE has been rewritten in terms of |
| 71 | GIT_TRACE_CURL. |
| 72 | |
| 73 | * Preliminary clean-ups around refs API, plus file format |
| 74 | specification documentation for the reftable backend. |
| 75 | |
| 76 | * Workaround breakage in MSVC build, where "curl-config --cflags" |
| 77 | gives settings appropriate for GCC build. |
| 78 | |
| 79 | * Code clean-up of "git clean" resulted in a fix of recent |
| 80 | performance regression. |
| 81 | |
| 82 | * Code clean-up in the codepath that serves "git fetch" continues. |
| 83 | |
| 84 | * "git merge-base --is-ancestor" is taught to take advantage of the |
| 85 | commit graph. |
| 86 | |
| 87 | * Rewrite of parts of the scripted "git submodule" Porcelain command |
| 88 | continues; this time it is "git submodule set-branch" subcommand's |
| 89 | turn. |
| 90 | |
| 91 | * The "fetch/clone" protocol has been updated to allow the server to |
| 92 | instruct the clients to grab pre-packaged packfile(s) in addition |
| 93 | to the packed object data coming over the wire. |
| 94 | |
| 95 | * A misdesigned strbuf_write_fd() function has been retired. |
| 96 | |
| 97 | * SHA-256 migration work continues, including CVS/SVN interface. |
| 98 | |
| 99 | * A few fields in "struct commit" that do not have to always be |
| 100 | present have been moved to commit slabs. |
| 101 | |
| 102 | * API cleanup for get_worktrees() |
| 103 | |
| 104 | * By renumbering object flag bits, "struct object" managed to lose |
| 105 | bloated inter-field padding. |
| 106 | |
| 107 | * The name of the primary branch in existing repositories, and the |
| 108 | default name used for the first branch in newly created |
| 109 | repositories, is made configurable, so that we can eventually wean |
| 110 | ourselves off of the hardcoded 'master'. |
| 111 | |
| 112 | * The effort to avoid using test_must_fail on non-git command continues. |
| 113 | |
| 114 | * In 2.28-rc0, we corrected a bug that some repository extensions are |
| 115 | honored by mistake even in a version 0 repositories (these |
| 116 | configuration variables in extensions.* namespace were supposed to |
| 117 | have special meaning in repositories whose version numbers are 1 or |
| 118 | higher), but this was a bit too big a change. The behaviour in |
| 119 | recent versions of Git where certain extensions.* were honored by |
| 120 | mistake even in version 0 repositories has been restored. |
| 121 | |
| 122 | |
| 123 | Fixes since v2.27 |
| 124 | ----------------- |
| 125 | |
| 126 | * The "--prepare-p4-only" option of "git p4" is supposed to stop |
| 127 | after replaying one changeset, but kept going (by mistake?) |
| 128 | |
| 129 | * The error message from "git checkout -b foo -t bar baz" was |
| 130 | confusing. |
| 131 | |
| 132 | * Some repositories in the wild have commits that record nonsense |
| 133 | committer timezone (e.g. rails.git); "git fast-import" learned an |
| 134 | option to pass these nonsense timestamps intact to allow recreating |
| 135 | existing repositories as-is. |
| 136 | (merge d42a2fb72f en/fast-import-looser-date later to maint). |
| 137 | |
| 138 | * The command line completion script (in contrib/) tried to complete |
| 139 | "git stash -p" as if it were "git stash push -p", but it was too |
| 140 | aggressive and also affected "git stash show -p", which has been |
| 141 | corrected. |
| 142 | (merge fffd0cf520 vs/complete-stash-show-p-fix later to maint). |
| 143 | |
| 144 | * On-the-wire protocol v2 easily falls into a deadlock between the |
| 145 | remote-curl helper and the fetch-pack process when the server side |
| 146 | prematurely throws an error and disconnects. The communication has |
| 147 | been updated to make it more robust. |
| 148 | |
| 149 | * "git checkout -p" did not handle a newly added path at all. |
| 150 | (merge 2c8bd8471a js/checkout-p-new-file later to maint). |
| 151 | |
| 152 | * The code to parse "git bisect start" command line was lax in |
| 153 | validating the arguments. |
| 154 | (merge 4d9005ff5d cb/bisect-helper-parser-fix later to maint). |
| 155 | |
| 156 | * Reduce memory usage during "diff --quiet" in a worktree with too |
| 157 | many stat-unmatched paths. |
| 158 | (merge d2d7fbe129 jk/diff-memuse-optim-with-stat-unmatch later to maint). |
| 159 | |
| 160 | * The reflog entries for "git clone" and "git fetch" did not |
| 161 | anonymize the URL they operated on. |
| 162 | (merge 46da295a77 js/reflog-anonymize-for-clone-and-fetch later to maint). |
| 163 | |
| 164 | * The behaviour of "sparse-checkout" in the state "git clone |
| 165 | --no-checkout" left was changed accidentally in 2.27, which has |
| 166 | been corrected. |
| 167 | |
| 168 | * Use of negative pathspec, while collecting paths including |
| 169 | untracked ones in the working tree, was broken. |
| 170 | |
| 171 | * The same worktree directory must be registered only once, but |
| 172 | "git worktree move" allowed this invariant to be violated, which |
| 173 | has been corrected. |
| 174 | (merge 810382ed37 es/worktree-duplicate-paths later to maint). |
| 175 | |
| 176 | * The effect of sparse checkout settings on submodules is documented. |
| 177 | (merge e7d7c73249 en/sparse-with-submodule-doc later to maint). |
| 178 | |
| 179 | * Code clean-up around "git branch" with a minor bugfix. |
| 180 | (merge dc44639904 dl/branch-cleanup later to maint). |
| 181 | |
| 182 | * A branch name used in a test has been clarified to match what is |
| 183 | going on. |
| 184 | (merge 08dc26061f pb/t4014-unslave later to maint). |
| 185 | |
| 186 | * An in-code comment in "git diff" has been updated. |
| 187 | (merge c592fd4c83 dl/diff-usage-comment-update later to maint). |
| 188 | |
| 189 | * The documentation and some tests have been adjusted for the recent |
| 190 | renaming of "pu" branch to "seen". |
| 191 | (merge 6dca5dbf93 js/pu-to-seen later to maint). |
| 192 | |
| 193 | * The code to push changes over "dumb" HTTP had a bad interaction |
| 194 | with the commit reachability code due to incorrect allocation of |
| 195 | object flag bits, which has been corrected. |
| 196 | (merge 64472d15e9 bc/http-push-flagsfix later to maint). |
| 197 | |
| 198 | * "git send-email --in-reply-to=<msg>" did not use the In-Reply-To: |
| 199 | header with the value given from the command line, and let it be |
| 200 | overridden by the value on In-Reply-To: header in the messages |
| 201 | being sent out (if exists). |
| 202 | (merge f9f60d7066 ra/send-email-in-reply-to-from-command-line-wins later to maint). |
| 203 | |
| 204 | * "git log -Lx,y:path --before=date" lost track of where the range |
| 205 | should be because it didn't take the changes made by the youngest |
| 206 | commits that are omitted from the output into account. |
| 207 | |
| 208 | * When "fetch.writeCommitGraph" configuration is set in a shallow |
| 209 | repository and a fetch moves the shallow boundary, we wrote out |
| 210 | broken commit-graph files that do not match the reality, which has |
| 211 | been corrected. |
| 212 | |
| 213 | * "git checkout" failed to catch an error from fstat() after updating |
| 214 | a path in the working tree. |
| 215 | (merge 35e6e212fd mt/entry-fstat-fallback-fix later to maint). |
| 216 | |
| 217 | * When an aliased command, whose output is piped to a pager by git, |
| 218 | gets killed by a signal, the pager got into a funny state, which |
| 219 | has been corrected (again). |
| 220 | (merge c0d73a59c9 ta/wait-on-aliased-commands-upon-signal later to maint). |
| 221 | |
| 222 | * The code to produce progress output from "git commit-graph --write" |
| 223 | had a few breakages, which have been fixed. |
| 224 | |
| 225 | * Other code cleanup, docfix, build fix, etc. |
| 226 | (merge 2c31a7aa44 jx/pkt-line-doc-count-fix later to maint). |
| 227 | (merge d63ae31962 cb/t5608-cleanup later to maint). |
| 228 | (merge 788db145c7 dl/t-readme-spell-git-correctly later to maint). |
| 229 | (merge 45a87a83bb dl/python-2.7-is-the-floor-version later to maint). |
| 230 | (merge b75a219904 es/advertise-contribution-doc later to maint). |
| 231 | (merge 0c9a4f638a rs/pull-leakfix later to maint). |
| 232 | (merge d546fe2874 rs/commit-reach-leakfix later to maint). |
| 233 | (merge 087bf5409c mk/pb-pretty-email-without-domain-part-fix later to maint). |
| 234 | (merge 5f4ee57ad9 es/worktree-code-cleanup later to maint). |
| 235 | (merge 0172f7834a cc/cat-file-usage-update later to maint). |
| 236 | (merge 81de0c01cf ma/rebase-doc-typofix later to maint). |