This repository has been archived on 2026-11-17. You can view files and clone it, but cannot push or open issues or pull requests.
caelum0x 49b183818c Unify gitgram + aitimemachine into the monorepo (de-submodule) and merge vendored engines
Convert gitgram/ and aitimemachine/ from embedded git repos (which showed as
opaque submodule gitlinks on GitHub) into fully-tracked first-party content.

AITM now merges its 7 vendored repos as first-party engines:
- memory: beads (Go task DAG) + context7 (docs SDK), unified into saved_contexts
- git: vendored C git (bundled binary) + gitnexus (in-process code intelligence)
- agentic: a2a (@a2a-js/sdk) + pal (AITM-managed Python subprocess)

Includes the full session's product build: Gitgram web/mobile/backend features,
AITM website/CLI/MCP, the unified memory layer, engine build tooling
(scripts/build-engines.sh, Dockerfile, CI), docs (README, ENGINES.md, ARCHITECTURE.md),
and extensive test suites.

Build artifacts, node_modules, external/, and gemini-cli are gitignored.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-23 14:14:09 +03:00

62 lines
2.2 KiB
Plaintext

GIT v1.6.3.2 Release Notes
==========================
Fixes since v1.6.3.1
--------------------
* A few codepaths picked up the first few bytes from an sha1[] by
casting the (char *) pointer to (int *); GCC 4.4 did not like this,
and aborted compilation.
* Some unlink(2) failures went undiagnosed.
* The "recursive" merge strategy misbehaved when faced rename/delete
conflicts while coming up with an intermediate merge base.
* The low-level merge algorithm did not handle a degenerate case of
merging a file with itself using itself as the common ancestor
gracefully. It should produce the file itself, but instead
produced an empty result.
* GIT_TRACE mechanism segfaulted when tracing a shell-quoted aliases.
* OpenBSD also uses st_ctimspec in "struct stat", instead of "st_ctim".
* With NO_CROSS_DIRECTORY_HARDLINKS, "make install" can be told not to
create hardlinks between $(gitexecdir)/git-$builtin_commands and
$(bindir)/git.
* command completion code in bash did not reliably detect that we are
in a bare repository.
* "git add ." in an empty directory complained that pathspec "." did not
match anything, which may be technically correct, but not useful. We
silently make it a no-op now.
* "git add -p" (and "patch" action in "git add -i") was broken when
the first hunk that adds a line at the top was split into two and
both halves are marked to be used.
* "git blame path" misbehaved at the commit where path became file
from a directory with some files in it.
* "git for-each-ref" had a segfaulting bug when dealing with a tag object
created by an ancient git.
* "git format-patch -k" still added patch numbers if format.numbered
configuration was set.
* "git grep --color ''" did not terminate. The command also had
subtle bugs with its -w option.
* http-push had a small use-after-free bug.
* "git push" was converting OFS_DELTA pack representation into less
efficient REF_DELTA representation unconditionally upon transfer,
making the transferred data unnecessarily larger.
* "git remote show origin" segfaulted when origin was still empty.
Many other general usability updates around help text, diagnostic messages
and documentation are included as well.