diff --git a/Documentation/Makefile b/Documentation/Makefile
index 2aae4c9..f5605b7 100644
--- a/Documentation/Makefile
+++ b/Documentation/Makefile
@@ -139,6 +139,7 @@
 ASCIIDOC_COMMON = $(ASCIIDOC) $(ASCIIDOC_EXTRA) $(ASCIIDOC_CONF) \
 		-amanversion=$(GIT_VERSION) \
 		-amanmanual='Git Manual' -amansource='Git'
+ASCIIDOC_DEPS = asciidoc.conf GIT-ASCIIDOCFLAGS
 TXT_TO_HTML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_HTML)
 TXT_TO_XML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_DOCBOOK)
 MANPAGE_XSL = manpage-normal.xsl
@@ -193,6 +194,7 @@
 ASCIIDOC_EXTRA += -acompat-mode -atabsize=8
 ASCIIDOC_EXTRA += -I. -rasciidoctor-extensions
 ASCIIDOC_EXTRA += -alitdd='&\#x2d;&\#x2d;'
+ASCIIDOC_DEPS = asciidoctor-extensions.rb GIT-ASCIIDOCFLAGS
 DBLATEX_COMMON =
 XMLTO_EXTRA += --skip-validation
 XMLTO_EXTRA += -x manpage.xsl
@@ -294,9 +296,7 @@
 	cmd-list.made $(cmds_txt)
 
 doc.dep : $(docdep_prereqs) $(DOC_DEP_TXT) build-docdep.perl
-	$(QUIET_GEN)$(RM) $@+ $@ && \
-	$(PERL_PATH) ./build-docdep.perl >$@+ $(QUIET_STDERR) && \
-	mv $@+ $@
+	$(QUIET_GEN)$(PERL_PATH) ./build-docdep.perl >$@ $(QUIET_STDERR)
 
 ifneq ($(MAKECMDGOALS),clean)
 -include doc.dep
@@ -316,8 +316,7 @@
 $(cmds_txt): cmd-list.made
 
 cmd-list.made: cmd-list.perl ../command-list.txt $(MAN1_TXT)
-	$(QUIET_GEN)$(RM) $@ && \
-	$(PERL_PATH) ./cmd-list.perl ../command-list.txt $(cmds_txt) $(QUIET_STDERR) && \
+	$(QUIET_GEN)$(PERL_PATH) ./cmd-list.perl ../command-list.txt $(cmds_txt) $(QUIET_STDERR) && \
 	date >$@
 
 mergetools_txt = mergetools-diff.txt mergetools-merge.txt
@@ -325,7 +324,7 @@
 $(mergetools_txt): mergetools-list.made
 
 mergetools-list.made: ../git-mergetool--lib.sh $(wildcard ../mergetools/*)
-	$(QUIET_GEN)$(RM) $@ && \
+	$(QUIET_GEN) \
 	$(SHELL_PATH) -c 'MERGE_TOOLS_DIR=../mergetools && \
 		. ../git-mergetool--lib.sh && \
 		show_tool_names can_diff "* " || :' >mergetools-diff.txt && \
@@ -354,32 +353,23 @@
 	$(RM) manpage-base-url.xsl
 	$(RM) GIT-ASCIIDOCFLAGS
 
-$(MAN_HTML): %.html : %.txt asciidoc.conf asciidoctor-extensions.rb GIT-ASCIIDOCFLAGS
-	$(QUIET_ASCIIDOC)$(RM) $@+ $@ && \
-	$(TXT_TO_HTML) -d manpage -o $@+ $< && \
-	mv $@+ $@
+$(MAN_HTML): %.html : %.txt $(ASCIIDOC_DEPS)
+	$(QUIET_ASCIIDOC)$(TXT_TO_HTML) -d manpage -o $@ $<
 
-$(OBSOLETE_HTML): %.html : %.txto asciidoc.conf asciidoctor-extensions.rb GIT-ASCIIDOCFLAGS
-	$(QUIET_ASCIIDOC)$(RM) $@+ $@ && \
-	$(TXT_TO_HTML) -o $@+ $< && \
-	mv $@+ $@
+$(OBSOLETE_HTML): %.html : %.txto $(ASCIIDOC_DEPS)
+	$(QUIET_ASCIIDOC)$(TXT_TO_HTML) -o $@ $<
 
 manpage-base-url.xsl: manpage-base-url.xsl.in
 	$(QUIET_GEN)sed "s|@@MAN_BASE_URL@@|$(MAN_BASE_URL)|" $< > $@
 
 %.1 %.5 %.7 : %.xml manpage-base-url.xsl $(wildcard manpage*.xsl)
-	$(QUIET_XMLTO)$(RM) $@ && \
-	$(XMLTO) -m $(MANPAGE_XSL) $(XMLTO_EXTRA) man $<
+	$(QUIET_XMLTO)$(XMLTO) -m $(MANPAGE_XSL) $(XMLTO_EXTRA) man $<
 
-%.xml : %.txt asciidoc.conf asciidoctor-extensions.rb GIT-ASCIIDOCFLAGS
-	$(QUIET_ASCIIDOC)$(RM) $@+ $@ && \
-	$(TXT_TO_XML) -d manpage -o $@+ $< && \
-	mv $@+ $@
+%.xml : %.txt $(ASCIIDOC_DEPS)
+	$(QUIET_ASCIIDOC)$(TXT_TO_XML) -d manpage -o $@ $<
 
 user-manual.xml: user-manual.txt user-manual.conf asciidoctor-extensions.rb GIT-ASCIIDOCFLAGS
-	$(QUIET_ASCIIDOC)$(RM) $@+ $@ && \
-	$(TXT_TO_XML) -d book -o $@+ $< && \
-	mv $@+ $@
+	$(QUIET_ASCIIDOC)$(TXT_TO_XML) -d book -o $@ $<
 
 technical/api-index.txt: technical/api-index-skel.txt \
 	technical/api-index.sh $(patsubst %,%.txt,$(API_DOCS))
@@ -400,46 +390,35 @@
 XSLTOPTS += --param generate.consistent.ids 1
 
 user-manual.html: user-manual.xml $(XSLT)
-	$(QUIET_XSLTPROC)$(RM) $@+ $@ && \
-	xsltproc $(XSLTOPTS) -o $@+ $(XSLT) $< && \
-	mv $@+ $@
+	$(QUIET_XSLTPROC)xsltproc $(XSLTOPTS) -o $@ $(XSLT) $<
 
 git.info: user-manual.texi
 	$(QUIET_MAKEINFO)$(MAKEINFO) --no-split -o $@ user-manual.texi
 
 user-manual.texi: user-manual.xml
-	$(QUIET_DB2TEXI)$(RM) $@+ $@ && \
-	$(DOCBOOK2X_TEXI) user-manual.xml --encoding=UTF-8 --to-stdout >$@++ && \
-	$(PERL_PATH) fix-texi.perl <$@++ >$@+ && \
-	rm $@++ && \
-	mv $@+ $@
+	$(QUIET_DB2TEXI)$(DOCBOOK2X_TEXI) user-manual.xml --encoding=UTF-8 --to-stdout >$@+ && \
+	$(PERL_PATH) fix-texi.perl <$@+ >$@ && \
+	$(RM) $@+
 
 user-manual.pdf: user-manual.xml
-	$(QUIET_DBLATEX)$(RM) $@+ $@ && \
-	$(DBLATEX) -o $@+ $(DBLATEX_COMMON) $< && \
-	mv $@+ $@
+	$(QUIET_DBLATEX)$(DBLATEX) -o $@ $(DBLATEX_COMMON) $<
 
 gitman.texi: $(MAN_XML) cat-texi.perl texi.xsl
-	$(QUIET_DB2TEXI)$(RM) $@+ $@ && \
+	$(QUIET_DB2TEXI) \
 	($(foreach xml,$(sort $(MAN_XML)),xsltproc -o $(xml)+ texi.xsl $(xml) && \
 		$(DOCBOOK2X_TEXI) --encoding=UTF-8 --to-stdout $(xml)+ && \
-		rm $(xml)+ &&) true) > $@++ && \
-	$(PERL_PATH) cat-texi.perl $@ <$@++ >$@+ && \
-	rm $@++ && \
-	mv $@+ $@
+		$(RM) $(xml)+ &&) true) > $@+ && \
+	$(PERL_PATH) cat-texi.perl $@ <$@+ >$@ && \
+	$(RM) $@+
 
 gitman.info: gitman.texi
 	$(QUIET_MAKEINFO)$(MAKEINFO) --no-split --no-validate $*.texi
 
 $(patsubst %.txt,%.texi,$(MAN_TXT)): %.texi : %.xml
-	$(QUIET_DB2TEXI)$(RM) $@+ $@ && \
-	$(DOCBOOK2X_TEXI) --to-stdout $*.xml >$@+ && \
-	mv $@+ $@
+	$(QUIET_DB2TEXI)$(DOCBOOK2X_TEXI) --to-stdout $*.xml >$@
 
 howto-index.txt: howto-index.sh $(HOWTO_TXT)
-	$(QUIET_GEN)$(RM) $@+ $@ && \
-	'$(SHELL_PATH_SQ)' ./howto-index.sh $(sort $(HOWTO_TXT)) >$@+ && \
-	mv $@+ $@
+	$(QUIET_GEN)'$(SHELL_PATH_SQ)' ./howto-index.sh $(sort $(HOWTO_TXT)) >$@
 
 $(patsubst %,%.html,$(ARTICLES)) : %.html : %.txt
 	$(QUIET_ASCIIDOC)$(TXT_TO_HTML) $*.txt
@@ -448,10 +427,9 @@
 
 howto/%.html: ASCIIDOC_EXTRA += -a git-relative-html-prefix=../
 $(patsubst %.txt,%.html,$(HOWTO_TXT)): %.html : %.txt GIT-ASCIIDOCFLAGS
-	$(QUIET_ASCIIDOC)$(RM) $@+ $@ && \
+	$(QUIET_ASCIIDOC) \
 	sed -e '1,/^$$/d' $< | \
-	$(TXT_TO_HTML) - >$@+ && \
-	mv $@+ $@
+	$(TXT_TO_HTML) - >$@
 
 install-webdoc : html
 	'$(SHELL_PATH_SQ)' ./install-webdoc.sh $(WEBDOC_DEST)
@@ -492,4 +470,7 @@
 	$(MAKE) -C po $@
 endif
 
+# Delete the target file on error
+.DELETE_ON_ERROR:
+
 .PHONY: FORCE
diff --git a/Documentation/RelNotes/2.33.0.txt b/Documentation/RelNotes/2.33.0.txt
new file mode 100644
index 0000000..57443c7
--- /dev/null
+++ b/Documentation/RelNotes/2.33.0.txt
@@ -0,0 +1,78 @@
+Git 2.33 Release Notes
+======================
+
+Backward compatibility notes
+----------------------------
+
+ * The "-m" option in "git log -m" that does not specify which format,
+   if any, of diff is desired did not have any visible effect; it now
+   implies some form of diff (by default "--patch") is produced.
+
+   You can disable the diff output with "git log -m --no-patch", but
+   then there probably isn't much point in passing "-m" in the first
+   place ;-).
+
+
+Updates since Git 2.32
+----------------------
+
+UI, Workflows & Features
+
+ * "git send-email" learned the "--sendmail-cmd" command line option
+   and the "sendemail.sendmailCmd" configuration variable, which is a
+   more sensible approach than the current way of repurposing the
+   "smtp-server" that is meant to name the server to instead name the
+   command to talk to the server.
+
+ * The "-m" option in "git log -m" that does not specify which format,
+   if any, of diff is desired did not have any visible effect; it now
+   implies some form of diff (by default "--patch") is produced.
+
+
+Performance, Internal Implementation, Development Support etc.
+
+ * The code to handle the "--format" option in "for-each-ref" and
+   friends made too many string comparisons on %(atom)s used in the
+   format string, which has been corrected by converting them into
+   enum when the format string is parsed.
+
+ * Use the hashfile API in the codepath that writes the index file to
+   reduce code duplication.
+
+ * Repeated rename detections in a sequence of mergy operations have
+   been optimize out.
+
+
+Fixes since v2.32
+-----------------
+
+ * We historically rejected a very short string as an author name
+   while accepting a patch e-mail, which has been loosened.
+   (merge 72ee47ceeb ef/mailinfo-short-name later to maint).
+
+ * The parallel checkout codepath did not initialize object ID field
+   used to talk to the worker processes in a futureproof way.
+
+ * Rewrite code that triggers undefined behaviour warning.
+   (merge aafa5df0df jn/size-t-casted-to-off-t-fix later to maint).
+
+ * The description of "fast-forward" in the glossary has been updated.
+   (merge e22f2daed0 ry/clarify-fast-forward-in-glossary later to maint).
+
+ * Recent "git clone" left a temporary directory behind when the
+   transport layer returned an failure.
+   (merge 6aacb7d861 jk/clone-clean-upon-transport-error later to maint).
+
+ * "git fetch" over protocol v2 left its side of the socket open after
+   it finished speaking, which unnecessarily wasted the resource on
+   the other side.
+   (merge ae1a7eefff jk/fetch-pack-v2-half-close-early later to maint).
+
+ * Other code cleanup, docfix, build fix, etc.
+   (merge bfe35a6165 ah/doc-describe later to maint).
+   (merge f302c1e4aa jc/clarify-revision-range later to maint).
+   (merge 3127ff90ea tl/fix-packfile-uri-doc later to maint).
+   (merge a84216c684 jk/doc-color-pager later to maint).
+   (merge 4e0a64a713 ab/trace2-squelch-gcc-warning later to maint).
+   (merge 225f7fa847 ps/rev-list-object-type-filter later to maint).
+   (merge 5317dfeaed dd/honor-users-tar-in-tests later to maint).
diff --git a/Documentation/config/color.txt b/Documentation/config/color.txt
index d5daacb..e05d520 100644
--- a/Documentation/config/color.txt
+++ b/Documentation/config/color.txt
@@ -127,8 +127,9 @@
 	interactive commands.
 
 color.pager::
-	A boolean to enable/disable colored output when the pager is in
-	use (default is true).
+	A boolean to specify whether `auto` color modes should colorize
+	output going to the pager. Defaults to true; set this to false
+	if your pager does not understand ANSI color codes.
 
 color.push::
 	A boolean to enable/disable color in push errors. May be set to
diff --git a/Documentation/config/merge.txt b/Documentation/config/merge.txt
index cb2ed58..6b66c83 100644
--- a/Documentation/config/merge.txt
+++ b/Documentation/config/merge.txt
@@ -14,7 +14,7 @@
 	branches at the remote named by `branch.<current branch>.remote`
 	are consulted, and then they are mapped via `remote.<remote>.fetch`
 	to their corresponding remote-tracking branches, and the tips of
-	these tracking branches are merged.
+	these tracking branches are merged. Defaults to true.
 
 merge.ff::
 	By default, Git does not create an extra merge commit when merging
diff --git a/Documentation/diff-options.txt b/Documentation/diff-options.txt
index 530d115..32e6dee 100644
--- a/Documentation/diff-options.txt
+++ b/Documentation/diff-options.txt
@@ -49,10 +49,9 @@
 --diff-merges=m:::
 -m:::
 	This option makes diff output for merge commits to be shown in
-	the default format. `-m` will produce the output only if `-p`
-	is given as well. The default format could be changed using
+	the default format. The default format could be changed using
 	`log.diffMerges` configuration parameter, which default value
-	is `separate`.
+	is `separate`. `-m` implies `-p`.
 +
 --diff-merges=first-parent:::
 --diff-merges=1:::
@@ -62,7 +61,8 @@
 --diff-merges=separate:::
 	This makes merge commits show the full diff with respect to
 	each of the parents. Separate log entry and diff is generated
-	for each parent.
+	for each parent. This is the format that `-m` produced
+	historically.
 +
 --diff-merges=combined:::
 --diff-merges=c:::
diff --git a/Documentation/git-describe.txt b/Documentation/git-describe.txt
index a88f6ae..c6a79c2 100644
--- a/Documentation/git-describe.txt
+++ b/Documentation/git-describe.txt
@@ -63,9 +63,10 @@
 	Automatically implies --tags.
 
 --abbrev=<n>::
-	Instead of using the default 7 hexadecimal digits as the
-	abbreviated object name, use <n> digits, or as many digits
-	as needed to form a unique object name.  An <n> of 0
+	Instead of using the default number of hexadecimal digits (which
+	will vary according to the number of objects in the repository with
+	a default of 7) of the abbreviated object name, use <n> digits, or
+	as many digits as needed to form a unique object name. An <n> of 0
 	will suppress long format, only showing the closest tag.
 
 --candidates=<n>::
@@ -139,8 +140,11 @@
 
 The number of additional commits is the number
 of commits which would be displayed by "git log v1.0.4..parent".
-The hash suffix is "-g" + unambiguous abbreviation for the tip commit
-of parent (which was `2414721b194453f058079d897d13c4e377f92dc6`).
+The hash suffix is "-g" + an unambigous abbreviation for the tip commit
+of parent (which was `2414721b194453f058079d897d13c4e377f92dc6`). The
+length of the abbreviation scales as the repository grows, using the
+approximate number of objects in the repository and a bit of math
+around the birthday paradox, and defaults to a minimum of 7.
 The "g" prefix stands for "git" and is used to allow describing the version of
 a software depending on the SCM the software is managed with. This is useful
 in an environment where people may use different SCMs.
diff --git a/Documentation/git-send-email.txt b/Documentation/git-send-email.txt
index 93708ae..3db4eab 100644
--- a/Documentation/git-send-email.txt
+++ b/Documentation/git-send-email.txt
@@ -167,6 +167,14 @@
 	`sendemail.envelopeSender` configuration variable; if that is
 	unspecified, choosing the envelope sender is left to your MTA.
 
+--sendmail-cmd=<command>::
+	Specify a command to run to send the email. The command should
+	be sendmail-like; specifically, it must support the `-i` option.
+	The command will be executed in the shell if necessary.  Default
+	is the value of `sendemail.sendmailcmd`.  If unspecified, and if
+	--smtp-server is also unspecified, git-send-email will search
+	for `sendmail` in `/usr/sbin`, `/usr/lib` and $PATH.
+
 --smtp-encryption=<encryption>::
 	Specify the encryption to use, either 'ssl' or 'tls'.  Any other
 	value reverts to plain SMTP.  Default is the value of
@@ -211,13 +219,16 @@
 
 --smtp-server=<host>::
 	If set, specifies the outgoing SMTP server to use (e.g.
-	`smtp.example.com` or a raw IP address).  Alternatively it can
-	specify a full pathname of a sendmail-like program instead;
-	the program must support the `-i` option.  Default value can
-	be specified by the `sendemail.smtpServer` configuration
-	option; the built-in default is to search for `sendmail` in
-	`/usr/sbin`, `/usr/lib` and $PATH if such program is
-	available, falling back to `localhost` otherwise.
+	`smtp.example.com` or a raw IP address).  If unspecified, and if
+	`--sendmail-cmd` is also unspecified, the default is to search
+	for `sendmail` in `/usr/sbin`, `/usr/lib` and $PATH if such a
+	program is available, falling back to `localhost` otherwise.
++
+For backward compatibility, this option can also specify a full pathname
+of a sendmail-like program instead; the program must support the `-i`
+option.  This method does not support passing arguments or using plain
+command names.  For those use cases, consider using `--sendmail-cmd`
+instead.
 
 --smtp-server-port=<port>::
 	Specifies a port different from the default port (SMTP
diff --git a/Documentation/glossary-content.txt b/Documentation/glossary-content.txt
index 67c7a50..c077971 100644
--- a/Documentation/glossary-content.txt
+++ b/Documentation/glossary-content.txt
@@ -146,8 +146,8 @@
 	<<def_revision,revision>> and you are "merging" another
 	<<def_branch,branch>>'s changes that happen to be a descendant of what
 	you have. In such a case, you do not make a new <<def_merge,merge>>
-	<<def_commit,commit>> but instead just update to his
-	revision. This will happen frequently on a
+	<<def_commit,commit>> but instead just update your branch to point at the same
+	revision as the branch you are merging. This will happen frequently on a
 	<<def_remote_tracking_branch,remote-tracking branch>> of a remote
 	<<def_repository,repository>>.
 
diff --git a/Documentation/revisions.txt b/Documentation/revisions.txt
index d9169c0..f5f17b6 100644
--- a/Documentation/revisions.txt
+++ b/Documentation/revisions.txt
@@ -260,6 +260,9 @@
 A commit's reachable set is the commit itself and the commits in
 its ancestry chain.
 
+There are several notations to specify a set of connected commits
+(called a "revision range"), illustrated below.
+
 
 Commit Exclusions
 ~~~~~~~~~~~~~~~~~
@@ -294,6 +297,26 @@
 I forked from them?"  Note that '..' would mean 'HEAD..HEAD' which is an
 empty range that is both reachable and unreachable from HEAD.
 
+Commands that are specifically designed to take two distinct ranges
+(e.g. "git range-diff R1 R2" to compare two ranges) do exist, but
+they are exceptions.  Unless otherwise noted, all "git" commands
+that operate on a set of commits work on a single revision range.
+In other words, writing two "two-dot range notation" next to each
+other, e.g.
+
+    $ git log A..B C..D
+
+does *not* specify two revision ranges for most commands.  Instead
+it will name a single connected set of commits, i.e. those that are
+reachable from either B or D but are reachable from neither A or C.
+In a linear history like this:
+
+    ---A---B---o---o---C---D
+
+because A and B are reachable from C, the revision range specified
+by these two dotted ranges is a single commit D.
+
+
 Other <rev>{caret} Parent Shorthand Notations
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Three other shorthands exist, particularly useful for merge commits,
diff --git a/Documentation/technical/api-trace2.txt b/Documentation/technical/api-trace2.txt
index 3f52f98..037a91c 100644
--- a/Documentation/technical/api-trace2.txt
+++ b/Documentation/technical/api-trace2.txt
@@ -396,14 +396,14 @@
 }
 ------------
 
-`"discard"`::
+`"too_many_files"`::
 	This event is written to the git-trace2-discard sentinel file if there
 	are too many files in the target trace directory (see the
 	trace2.maxFiles config option).
 +
 ------------
 {
-	"event":"discard",
+	"event":"too_many_files",
 	...
 }
 ------------
diff --git a/Documentation/technical/packfile-uri.txt b/Documentation/technical/packfile-uri.txt
index f7eabc6..1eb525f 100644
--- a/Documentation/technical/packfile-uri.txt
+++ b/Documentation/technical/packfile-uri.txt
@@ -35,13 +35,14 @@
 at least so that we can test the client.
 
 This is the implementation: a feature, marked experimental, that allows the
-server to be configured by one or more `uploadpack.blobPackfileUri=<sha1>
-<uri>` entries. Whenever the list of objects to be sent is assembled, all such
-blobs are excluded, replaced with URIs. As noted in "Future work" below, the
-server can evolve in the future to support excluding other objects (or other
-implementations of servers could be made that support excluding other objects)
-without needing a protocol change, so clients should not expect that packfiles
-downloaded in this way only contain single blobs.
+server to be configured by one or more `uploadpack.blobPackfileUri=
+<object-hash> <pack-hash> <uri>` entries. Whenever the list of objects to be
+sent is assembled, all such blobs are excluded, replaced with URIs. As noted
+in "Future work" below, the server can evolve in the future to support
+excluding other objects (or other implementations of servers could be made
+that support excluding other objects) without needing a protocol change, so
+clients should not expect that packfiles downloaded in this way only contain
+single blobs.
 
 Client design
 -------------
diff --git a/Documentation/technical/partial-clone.txt b/Documentation/technical/partial-clone.txt
index 0780d30..a0dd7c6 100644
--- a/Documentation/technical/partial-clone.txt
+++ b/Documentation/technical/partial-clone.txt
@@ -242,8 +242,7 @@
   repository and can satisfy all such requests.
 
 - Repack essentially treats promisor and non-promisor packfiles as 2
-  distinct partitions and does not mix them.  Repack currently only works
-  on non-promisor packfiles and loose objects.
+  distinct partitions and does not mix them.
 
 - Dynamic object fetching invokes fetch-pack once *for each item*
   because most algorithms stumble upon a missing object and need to have
@@ -273,9 +272,6 @@
 The user might want to work in a triangular work flow with multiple
 promisor remotes that each have an incomplete view of the repository.
 
-- Allow repack to work on promisor packfiles (while keeping them distinct
-  from non-promisor packfiles).
-
 - Allow non-pathname-based filters to make use of packfile bitmaps (when
   present).  This was just an omission during the initial implementation.
 
diff --git a/Documentation/technical/remembering-renames.txt b/Documentation/technical/remembering-renames.txt
new file mode 100644
index 0000000..2fd5cc8
--- /dev/null
+++ b/Documentation/technical/remembering-renames.txt
@@ -0,0 +1,671 @@
+Rebases and cherry-picks involve a sequence of merges whose results are
+recorded as new single-parent commits.  The first parent side of those
+merges represent the "upstream" side, and often include a far larger set of
+changes than the second parent side.  Traditionally, the renames on the
+first-parent side of that sequence of merges were repeatedly re-detected
+for every merge.  This file explains why it is safe and effective during
+rebases and cherry-picks to remember renames on the upstream side of
+history as an optimization, assuming all merges are automatic and clean
+(i.e. no conflicts and not interrupted for user input or editing).
+
+Outline:
+
+  0. Assumptions
+
+  1. How rebasing and cherry-picking work
+
+  2. Why the renames on MERGE_SIDE1 in any given pick are *always* a
+     superset of the renames on MERGE_SIDE1 for the next pick.
+
+  3. Why any rename on MERGE_SIDE1 in any given pick is _almost_ always also
+     a rename on MERGE_SIDE1 for the next pick
+
+  4. A detailed description of the the counter-examples to #3.
+
+  5. Why the special cases in #4 are still fully reasonable to use to pair
+     up files for three-way content merging in the merge machinery, and why
+     they do not affect the correctness of the merge.
+
+  6. Interaction with skipping of "irrelevant" renames
+
+  7. Additional items that need to be cached
+
+  8. How directory rename detection interacts with the above and why this
+     optimization is still safe even if merge.directoryRenames is set to
+     "true".
+
+
+=== 0. Assumptions ===
+
+There are two assumptions that will hold throughout this document:
+
+  * The upstream side where commits are transplanted to is treated as the
+    first parent side when rebase/cherry-pick call the merge machinery
+
+  * All merges are fully automatic
+
+and a third that will hold in sections 2-5 for simplicity, that I'll later
+address in section 8:
+
+  * No directory renames occur
+
+
+Let me explain more about each assumption and why I include it:
+
+
+The first assumption is merely for the purposes of making this document
+clearer; the optimization implementation does not actually depend upon it.
+However, the assumption does hold in all cases because it reflects the way
+that both rebase and cherry-pick were implemented; and the implementation
+of cherry-pick and rebase are not readily changeable for backwards
+compatibility reasons (see for example the discussion of the --ours and
+--theirs flag in the documentation of `git checkout`, particularly the
+comments about how they behave with rebase).  The optimization avoids
+checking first-parent-ness, though.  It checks the conditions that make the
+optimization valid instead, so it would still continue working if someone
+changed the parent ordering that cherry-pick and rebase use.  But making
+this assumption does make this document much clearer and prevents me from
+having to repeat every example twice.
+
+If the second assumption is violated, then the optimization simply is
+turned off and thus isn't relevant to consider.  The second assumption can
+also be stated as "there is no interruption for a user to resolve conflicts
+or to just further edit or tweak files".  While real rebases and
+cherry-picks are often interrupted (either because it's an interactive
+rebase where the user requested to stop and edit, or because there were
+conflicts that the user needs to resolve), the cache of renames is not
+stored on disk, and thus is thrown away as soon as the rebase or cherry
+pick stops for the user to resolve the operation.
+
+The third assumption makes sections 2-5 simpler, and allows people to
+understand the basics of why this optimization is safe and effective, and
+then I can go back and address the specifics in section 8.  It is probably
+also worth noting that if directory renames do occur, then the default of
+merge.directoryRenames being set to "conflict" means that the operation
+will stop for users to resolve the conflicts and the cache will be thrown
+away, and thus that there won't be an optimization to apply.  So, the only
+reason we need to address directory renames specifically, is that some
+users will have set merge.directoryRenames to "true" to allow the merges to
+continue to proceed automatically.  The optimization is still safe with
+this config setting, but we have to discuss a few more cases to show why;
+this discussion is deferred until section 8.
+
+
+=== 1. How rebasing and cherry-picking work ===
+
+Consider the following setup (from the git-rebase manpage):
+
+		     A---B---C topic
+		    /
+	       D---E---F---G main
+
+After rebasing or cherry-picking topic onto main, this will appear as:
+
+			     A'--B'--C' topic
+			    /
+	       D---E---F---G main
+
+The way the commits A', B', and C' are created is through a series of
+merges, where rebase or cherry-pick sequentially uses each of the three
+A-B-C commits in a special merge operation.  Let's label the three commits
+in the merge operation as MERGE_BASE, MERGE_SIDE1, and MERGE_SIDE2.  For
+this picture, the three commits for each of the three merges would be:
+
+To create A':
+   MERGE_BASE:   E
+   MERGE_SIDE1:  G
+   MERGE_SIDE2:  A
+
+To create B':
+   MERGE_BASE:   A
+   MERGE_SIDE1:  A'
+   MERGE_SIDE2:  B
+
+To create C':
+   MERGE_BASE:   B
+   MERGE_SIDE1:  B'
+   MERGE_SIDE2:  C
+
+Sometimes, folks are surprised that these three-way merges are done.  It
+can be useful in understanding these three-way merges to view them in a
+slightly different light.  For example, in creating C', you can view it as
+either:
+
+  * Apply the changes between B & C to B'
+  * Apply the changes between B & B' to C
+
+Conceptually the two statements above are the same as a three-way merge of
+B, B', and C, at least the parts before you decide to record a commit.
+
+
+=== 2. Why the renames on MERGE_SIDE1 in any given pick are always a ===
+===    superset of the renames on MERGE_SIDE1 for the next pick.     ===
+
+The merge machinery uses the filenames it is fed from MERGE_BASE,
+MERGE_SIDE1, and MERGE_SIDE2.  It will only move content to a different
+filename under one of three conditions:
+
+  * To make both pieces of a conflict available to a user during conflict
+    resolution (examples: directory/file conflict, add/add type conflict
+    such as symlink vs. regular file)
+
+  * When MERGE_SIDE1 renames the file.
+
+  * When MERGE_SIDE2 renames the file.
+
+First, let's remember what commits are involved in the first and second
+picks of the cherry-pick or rebase sequence:
+
+To create A':
+   MERGE_BASE:   E
+   MERGE_SIDE1:  G
+   MERGE_SIDE2:  A
+
+To create B':
+   MERGE_BASE:   A
+   MERGE_SIDE1:  A'
+   MERGE_SIDE2:  B
+
+So, in particular, we need to show that the renames between E and G are a
+superset of those between A and A'.
+
+A' is created by the first merge.  A' will only have renames for one of the
+three reasons listed above.  The first case, a conflict, results in a
+situation where the cache is dropped and thus this optimization doesn't
+take effect, so we need not consider that case.  The third case, a rename
+on MERGE_SIDE2 (i.e. from G to A), will show up in A' but it also shows up
+in A -- therefore when diffing A and A' that path does not show up as a
+rename.  The only remaining way for renames to show up in A' is for the
+rename to come from MERGE_SIDE1.  Therefore, all renames between A and A'
+are a subset of those between E and G.  Equivalently, all renames between E
+and G are a superset of those between A and A'.
+
+
+=== 3. Why any rename on MERGE_SIDE1 in any given pick is _almost_   ===
+===    always also a rename on MERGE_SIDE1 for the next pick.        ===
+
+Let's again look at the first two picks:
+
+To create A':
+   MERGE_BASE:   E
+   MERGE_SIDE1:  G
+   MERGE_SIDE2:  A
+
+To create B':
+   MERGE_BASE:   A
+   MERGE_SIDE1:  A'
+   MERGE_SIDE2:  B
+
+Now let's look at any given rename from MERGE_SIDE1 of the first pick, i.e.
+any given rename from E to G.  Let's use the filenames 'oldfile' and
+'newfile' for demonstration purposes.  That first pick will function as
+follows; when the rename is detected, the merge machinery will do a
+three-way content merge of the following:
+    E:oldfile
+    G:newfile
+    A:oldfile
+and produce a new result:
+    A':newfile
+
+Note above that I've assumed that E->A did not rename oldfile.  If that
+side did rename, then we most likely have a rename/rename(1to2) conflict
+that will cause the rebase or cherry-pick operation to halt and drop the
+in-memory cache of renames and thus doesn't need to be considered further.
+In the special case that E->A does rename the file but also renames it to
+newfile, then there is no conflict from the renaming and the merge can
+succeed.  In this special case, the rename is not valid to cache because
+the second merge will find A:newfile in the MERGE_BASE (see also the new
+testcases in t6429 with "rename same file identically" in their
+description).  So a rename/rename(1to1) needs to be specially handled by
+pruning renames from the cache and decrementing the dir_rename_counts in
+the current and leading directories associated with those renames.  Or,
+since these are really rare, one could just take the easy way out and
+disable the remembering renames optimization when a rename/rename(1to1)
+happens.
+
+The previous paragraph handled the cases for E->A renaming oldfile, let's
+continue assuming that oldfile is not renamed in A.
+
+As per the diagram for creating B', MERGE_SIDE1 involves the changes from A
+to A'.  So, we are curious whether A:oldfile and A':newfile will be viewed
+as renames.  Note that:
+
+  * There will be no A':oldfile (because there could not have been a
+    G:oldfile as we do not do break detection in the merge machinery and
+    G:newfile was detected as a rename, and by the construction of the
+    rename above that merged cleanly, the merge machinery will ensure there
+    is no 'oldfile' in the result).
+
+  * There will be no A:newfile (if there had been, we would have had a
+    rename/add conflict).
+
+  * Clearly A:oldfile and A':newfile are "related" (A':newfile came from a
+    clean three-way content merge involving A:oldfile).
+
+We can also expound on the third point above, by noting that three-way
+content merges can also be viewed as applying the differences between the
+base and one side to the other side.  Thus we can view A':newfile as
+having been created by taking the changes between E:oldfile and G:newfile
+(which were detected as being related, i.e. <50% changed) to A:oldfile.
+
+Thus A:oldfile and A':newfile are just as related as E:oldfile and
+G:newfile are -- they have exactly identical differences.  Since the latter
+were detected as renames, A:oldfile and A':newfile should also be
+detectable as renames almost always.
+
+
+=== 4. A detailed description of the counter-examples to #3.         ===
+
+We already noted in section 3 that rename/rename(1to1) (i.e. both sides
+renaming a file the same way) was one counter-example.  The more
+interesting bit, though, is why did we need to use the "almost" qualifier
+when stating that A:oldfile and A':newfile are "almost" always detectable
+as renames?
+
+Let's repeat an earlier point that section 3 made:
+
+  A':newfile was created by applying the changes between E:oldfile and
+  G:newfile to A:oldfile.  The changes between E:oldfile and G:newfile were
+  <50% of the size of E:oldfile.
+
+If those changes that were <50% of the size of E:oldfile are also <50% of
+the size of A:oldfile, then A:oldfile and A':newfile will be detectable as
+renames.  However, if there is a dramatic size reduction between E:oldfile
+and A:oldfile (but the changes between E:oldfile, G:newfile, and A:oldfile
+still somehow merge cleanly), then traditional rename detection would not
+detect A:oldfile and A':newfile as renames.
+
+Here's an example where that can happen:
+  * E:oldfile had 20 lines
+  * G:newfile added 10 new lines at the beginning of the file
+  * A:oldfile kept the first 3 lines of the file, and deleted all the rest
+then
+  => A':newfile would have 13 lines, 3 of which matches those in A:oldfile.
+E:oldfile -> G:newfile would be detected as a rename, but A:oldfile and
+A':newfile would not be.
+
+
+=== 5. Why the special cases in #4 are still fully reasonable to use to    ===
+===    pair up files for three-way content merging in the merge machinery, ===
+===    and why they do not affect the correctness of the merge.            ===
+
+In the rename/rename(1to1) case, A:newfile and A':newfile are not renames
+since they use the *same* filename.  However, files with the same filename
+are obviously fine to pair up for three-way content merging (the merge
+machinery has never employed break detection).  The interesting
+counter-example case is thus not the rename/rename(1to1) case, but the case
+where A did not rename oldfile.  That was the case that we spent most of
+the time discussing in sections 3 and 4.  The remainder of this section
+will be devoted to that case as well.
+
+So, even if A:oldfile and A':newfile aren't detectable as renames, why is
+it still reasonable to pair them up for three-way content merging in the
+merge machinery?  There are multiple reasons:
+
+  * As noted in sections 3 and 4, the diff between A:oldfile and A':newfile
+    is *exactly* the same as the diff between E:oldfile and G:newfile.  The
+    latter pair were detected as renames, so it seems unlikely to surprise
+    users for us to treat A:oldfile and A':newfile as renames.
+
+  * In fact, "oldfile" and "newfile" were at one point detected as renames
+    due to how they were constructed in the E..G chain.  And we used that
+    information once already in this rebase/cherry-pick.  I think users
+    would be unlikely to be surprised at us continuing to treat the files
+    as renames and would quickly understand why we had done so.
+
+  * Marking or declaring files as renames is *not* the end goal for merges.
+    Merges use renames to determine which files make sense to be paired up
+    for three-way content merges.
+
+  * A:oldfile and A':newfile were _already_ paired up in a three-way
+    content merge; that is how A':newfile was created.  In fact, that
+    three-way content merge was clean.  So using them again in a later
+    three-way content merge seems very reasonable.
+
+However, the above is focusing on the common scenarios.  Let's try to look
+at all possible unusual scenarios and compare without the optimization to
+with the optimization.  Consider the following theoretical cases; we will
+then dive into each to determine which of them are possible,
+and if so, what they mean:
+
+  1. Without the optimization, the second merge results in a conflict.
+     With the optimization, the second merge also results in a conflict.
+     Questions: Are the conflicts confusingly different?  Better in one case?
+
+  2. Without the optimization, the second merge results in NO conflict.
+     With the optimization, the second merge also results in NO conflict.
+     Questions: Are the merges the same?
+
+  3. Without the optimization, the second merge results in a conflict.
+     With the optimization, the second merge results in NO conflict.
+     Questions: Possible?  Bug, bugfix, or something else?
+
+  4. Without the optimization, the second merge results in NO conflict.
+     With the optimization, the second merge results in a conflict.
+     Questions: Possible?  Bug, bugfix, or something else?
+
+I'll consider all four cases, but out of order.
+
+The fourth case is impossible.  For the code without the remembering
+renames optimization to not get a conflict, B:oldfile would need to exactly
+match A:oldfile -- if it doesn't, there would be a modify/delete conflict.
+If A:oldfile matches B:oldfile exactly, then a three-way content merge
+between A:oldfile, A':newfile, and B:oldfile would have no conflict and
+just give us the version of newfile from A' as the result.
+
+From the same logic as the above paragraph, the second case would indeed
+result in identical merges.  When A:oldfile exactly matches B:oldfile, an
+undetected rename would say, "Oh, I see one side didn't modify 'oldfile'
+and the other side deleted it.  I'll delete it.  And I see you have this
+brand new file named 'newfile' in A', so I'll keep it."  That gives the
+same results as three-way content merging A:oldfile, A':newfile, and
+B:oldfile -- a removal of oldfile with the version of newfile from A'
+showing up in the result.
+
+The third case is interesting.  It means that A:oldfile and A':newfile were
+not just similar enough, but that the changes between them did not conflict
+with the changes between A:oldfile and B:oldfile.  This would validate our
+hunch that the files were similar enough to be used in a three-way content
+merge, and thus seems entirely correct for us to have used them that way.
+(Sidenote: One particular example here may be enlightening.  Let's say that
+B was an immediate revert of A.  B clearly would have been a clean revert
+of A, since A was B's immediate parent.  One would assume that if you can
+pick a commit, you should also be able to cherry-pick its immediate revert.
+However, this is one of those funny corner cases; without this
+optimization, we just successfully picked a commit cleanly, but we are
+unable to cherry-pick its immediate revert due to the size differences
+between E:oldfile and A:oldfile.)
+
+That leaves only the first case to consider -- when we get conflicts both
+with or without the optimization.  Without the optimization, we'll have a
+modify/delete conflict, where both A':newfile and B:oldfile are left in the
+tree for the user to deal with and no hints about the potential similarity
+between the two.  With the optimization, we'll have a three-way content
+merged A:oldfile, A':newfile, and B:oldfile with conflict markers
+suggesting we thought the files were related but giving the user the chance
+to resolve.  As noted above, I don't think users will find us treating
+'oldfile' and 'newfile' as related as a surprise since they were between E
+and G.  In any event, though, this case shouldn't be concerning since we
+hit a conflict in both cases, told the user what we know, and asked them to
+resolve it.
+
+So, in summary, case 4 is impossible, case 2 yields the same behavior, and
+cases 1 and 3 seem to provide as good or better behavior with the
+optimization than without.
+
+
+=== 6. Interaction with skipping of "irrelevant" renames ===
+
+Previous optimizations involved skipping rename detection for paths
+considered to be "irrelevant".  See for example the following commits:
+
+  * 32a56dfb99 ("merge-ort: precompute subset of sources for which we
+		need rename detection", 2021-03-11)
+  * 2fd9eda462 ("merge-ort: precompute whether directory rename
+		detection is needed", 2021-03-11)
+  * 9bd342137e ("diffcore-rename: determine which relevant_sources are
+		no longer relevant", 2021-03-13)
+
+Relevance is always determined by what the _other_ side of history has
+done, in terms of modifing a file that our side renamed, or adding a
+file to a directory which our side renamed.  This means that a path
+that is "irrelevant" when picking the first commit of a series in a
+rebase or cherry-pick, may suddenly become "relevant" when picking the
+next commit.
+
+The upshot of this is that we can only cache rename detection results
+for relevant paths, and need to re-check relevance in subsequent
+commits.  If those subsequent commits have additional paths that are
+relevant for rename detection, then we will need to redo rename
+detection -- though we can limit it to the paths for which we have not
+already detected renames.
+
+
+=== 7. Additional items that need to be cached ===
+
+It turns out we have to cache more than just renames; we also cache:
+
+  A) non-renames (i.e. unpaired deletes)
+  B) counts of renames within directories
+  C) sources that were marked as RELEVANT_LOCATION, but which were
+     downgraded to RELEVANT_NO_MORE
+  D) the toplevel trees involved in the merge
+
+These are all stored in struct rename_info, and respectively appear in
+  * cached_pairs (along side actual renames, just with a value of NULL)
+  * dir_rename_counts
+  * cached_irrelevant
+  * merge_trees
+
+The reason for (A) comes from the irrelevant renames skipping
+optimization discussed in section 6.  The fact that irrelevant renames
+are skipped means we only get a subset of the potential renames
+detected and subsequent commits may need to run rename detection on
+the upstream side on a subset of the remaining renames (to get the
+renames that are relevant for that later commit).  Since unpaired
+deletes are involved in rename detection too, we don't want to
+repeatedly check that those paths remain unpaired on the upstream side
+with every commit we are transplanting.
+
+The reason for (B) is that diffcore_rename_extended() is what
+generates the counts of renames by directory which is needed in
+directory rename detection, and if we don't run
+diffcore_rename_extended() again then we need to have the output from
+it, including dir_rename_counts, from the previous run.
+
+The reason for (C) is that merge-ort's tree traversal will again think
+those paths are relevant (marking them as RELEVANT_LOCATION), but the
+fact that they were downgraded to RELEVANT_NO_MORE means that
+dir_rename_counts already has the information we need for directory
+rename detection.  (A path which becomes RELEVANT_CONTENT in a
+subsequent commit will be removed from cached_irrelevant.)
+
+The reason for (D) is that is how we determine whether the remember
+renames optimization can be used.  In particular, remembering that our
+sequence of merges looks like:
+
+   Merge 1:
+   MERGE_BASE:   E
+   MERGE_SIDE1:  G
+   MERGE_SIDE2:  A
+   => Creates    A'
+
+   Merge 2:
+   MERGE_BASE:   A
+   MERGE_SIDE1:  A'
+   MERGE_SIDE2:  B
+   => Creates    B'
+
+It is the fact that the trees A and A' appear both in Merge 1 and in
+Merge 2, with A as a parent of A' that allows this optimization.  So
+we store the trees to compare with what we are asked to merge next
+time.
+
+
+=== 8. How directory rename detection interacts with the above and   ===
+===    why this optimization is still safe even if                   ===
+===    merge.directoryRenames is set to "true".                      ===
+
+As noted in the assumptions section:
+
+    """
+    ...if directory renames do occur, then the default of
+    merge.directoryRenames being set to "conflict" means that the operation
+    will stop for users to resolve the conflicts and the cache will be
+    thrown away, and thus that there won't be an optimization to apply.
+    So, the only reason we need to address directory renames specifically,
+    is that some users will have set merge.directoryRenames to "true" to
+    allow the merges to continue to proceed automatically.
+    """
+
+Let's remember that we need to look at how any given pick affects the next
+one.  So let's again use the first two picks from the diagram in section
+one:
+
+  First pick does this three-way merge:
+    MERGE_BASE:   E
+    MERGE_SIDE1:  G
+    MERGE_SIDE2:  A
+    => creates A'
+
+  Second pick does this three-way merge:
+    MERGE_BASE:   A
+    MERGE_SIDE1:  A'
+    MERGE_SIDE2:  B
+    => creates B'
+
+Now, directory rename detection exists so that if one side of history
+renames a directory, and the other side adds a new file to the old
+directory, then the merge (with merge.directoryRenames=true) can move the
+file into the new directory.  There are two qualitatively different ways to
+add a new file to an old directory: create a new file, or rename a file
+into that directory.  Also, directory renames can be done on either side of
+history, so there are four cases to consider:
+
+  * MERGE_SIDE1 renames old dir, MERGE_SIDE2 adds new file to   old dir
+  * MERGE_SIDE1 renames old dir, MERGE_SIDE2 renames  file into old dir
+  * MERGE_SIDE1 adds new file to   old dir, MERGE_SIDE2 renames old dir
+  * MERGE_SIDE1 renames  file into old dir, MERGE_SIDE2 renames old dir
+
+One last note before we consider these four cases: There are some
+important properties about how we implement this optimization with
+respect to directory rename detection that we need to bear in mind
+while considering all of these cases:
+
+  * rename caching occurs *after* applying directory renames
+
+  * a rename created by directory rename detection is recorded for the side
+    of history that did the directory rename.
+
+  * dir_rename_counts, the nested map of
+	{oldname => {newname => count}},
+    is cached between runs as well.  This basically means that directory
+    rename detection is also cached, though only on the side of history
+    that we cache renames for (MERGE_SIDE1 as far as this document is
+    concerned; see the assumptions section).  Two interesting sub-notes
+    about these counts:
+
+    * If we need to perform rename-detection again on the given side (e.g.
+      some paths are relevant for rename detection that weren't before),
+      then we clear dir_rename_counts and recompute it, making use of
+      cached_pairs.  The reason it is important to do this is optimizations
+      around RELEVANT_LOCATION exist to prevent us from computing
+      unnecessary renames for directory rename detection and from computing
+      dir_rename_counts for irrelevant directories; but those same renames
+      or directories may become necessary for subsequent merges.  The
+      easiest way to "fix up" dir_rename_counts in such cases is to just
+      recompute it.
+
+    * If we prune rename/rename(1to1) entries from the cache, then we also
+      need to update dir_rename_counts to decrement the counts for the
+      involved directory and any relevant parent directories (to undo what
+      update_dir_rename_counts() in diffcore-rename.c incremented when the
+      rename was initially found).  If we instead just disable the
+      remembering renames optimization when the exceedingly rare
+      rename/rename(1to1) cases occur, then dir_rename_counts will get
+      re-computed the next time rename detection occurs, as noted above.
+
+  * the side with multiple commits to pick, is the side of history that we
+    do NOT cache renames for.  Thus, there are no additional commits to
+    change the number of renames in a directory, except for those done by
+    directory rename detection (which always pad the majority).
+
+  * the "renames" we cache are modified slightly by any directory rename,
+    as noted below.
+
+Now, with those notes out of the way, let's go through the four cases
+in order:
+
+Case 1: MERGE_SIDE1 renames old dir, MERGE_SIDE2 adds new file to old dir
+
+  This case looks like this:
+
+    MERGE_BASE:   E,   Has olddir/
+    MERGE_SIDE1:  G,   Renames olddir/ -> newdir/
+    MERGE_SIDE2:  A,   Adds olddir/newfile
+    => creates    A',  With newdir/newfile
+
+    MERGE_BASE:   A,   Has olddir/newfile
+    MERGE_SIDE1:  A',  Has newdir/newfile
+    MERGE_SIDE2:  B,   Modifies olddir/newfile
+    => expected   B',  with threeway-merged newdir/newfile from above
+
+  In this case, with the optimization, note that after the first commit:
+    * MERGE_SIDE1 remembers olddir/ -> newdir/
+    * MERGE_SIDE1 has cached olddir/newfile -> newdir/newfile
+  Given the cached rename noted above, the second merge can proceed as
+  expected without needing to perform rename detection from A -> A'.
+
+Case 2: MERGE_SIDE1 renames old dir, MERGE_SIDE2 renames  file into old dir
+
+  This case looks like this:
+    MERGE_BASE:   E    oldfile, olddir/
+    MERGE_SIDE1:  G    oldfile, olddir/ -> newdir/
+    MERGE_SIDE2:  A    oldfile -> olddir/newfile
+    => creates    A',  With newdir/newfile representing original oldfile
+
+    MERGE_BASE:   A    olddir/newfile
+    MERGE_SIDE1:  A'   newdir/newfile
+    MERGE_SIDE2:  B    modify olddir/newfile
+    => expected   B',  with threeway-merged newdir/newfile from above
+
+  In this case, with the optimization, note that after the first commit:
+    * MERGE_SIDE1 remembers olddir/ -> newdir/
+    * MERGE_SIDE1 has cached olddir/newfile -> newdir/newfile
+		  (NOT oldfile -> newdir/newfile; compare to case with
+		   (p->status == 'R' && new_path) in possibly_cache_new_pair())
+
+  Given the cached rename noted above, the second merge can proceed as
+  expected without needing to perform rename detection from A -> A'.
+
+Case 3: MERGE_SIDE1 adds new file to   old dir, MERGE_SIDE2 renames old dir
+
+  This case looks like this:
+
+    MERGE_BASE:   E,   Has olddir/
+    MERGE_SIDE1:  G,   Adds olddir/newfile
+    MERGE_SIDE2:  A,   Renames olddir/ -> newdir/
+    => creates    A',  With newdir/newfile
+
+    MERGE_BASE:   A,   Has newdir/, but no notion of newdir/newfile
+    MERGE_SIDE1:  A',  Has newdir/newfile
+    MERGE_SIDE2:  B,   Has newdir/, but no notion of newdir/newfile
+    => expected   B',  with newdir/newfile from A'
+
+  In this case, with the optimization, note that after the first commit there
+  were no renames on MERGE_SIDE1, and any renames on MERGE_SIDE2 are tossed.
+  But the second merge didn't need any renames so this is fine.
+
+Case 4: MERGE_SIDE1 renames  file into old dir, MERGE_SIDE2 renames old dir
+
+  This case looks like this:
+
+    MERGE_BASE:   E,   Has olddir/
+    MERGE_SIDE1:  G,   Renames oldfile -> olddir/newfile
+    MERGE_SIDE2:  A,   Renames olddir/ -> newdir/
+    => creates    A',  With newdir/newfile representing original oldfile
+
+    MERGE_BASE:   A,   Has oldfile
+    MERGE_SIDE1:  A',  Has newdir/newfile
+    MERGE_SIDE2:  B,   Modifies oldfile
+    => expected   B',  with threeway-merged newdir/newfile from above
+
+  In this case, with the optimization, note that after the first commit:
+    * MERGE_SIDE1 remembers oldfile -> newdir/newfile
+		  (NOT oldfile -> olddir/newfile; compare to case of second
+		   block under p->status == 'R' in possibly_cache_new_pair())
+    * MERGE_SIDE2 renames are tossed because only MERGE_SIDE1 is remembered
+
+  Given the cached rename noted above, the second merge can proceed as
+  expected without needing to perform rename detection from A -> A'.
+
+Finally, I'll just note here that interactions with the
+skip-irrelevant-renames optimization means we sometimes don't detect
+renames for any files within a directory that was renamed, in which
+case we will not have been able to detect any rename for the directory
+itself.  In such a case, we do not know whether the directory was
+renamed; we want to be careful to avoid cacheing some kind of "this
+directory was not renamed" statement.  If we did, then a subsequent
+commit being rebased could add a file to the old directory, and the
+user would expect it to end up in the correct directory -- something
+our erroneous "this directory was not renamed" cache would preclude.
diff --git a/Makefile b/Makefile
index c3565fc..502e0c9 100644
--- a/Makefile
+++ b/Makefile
@@ -398,6 +398,10 @@
 # with a different indexfile format version.  If it isn't set the index
 # file format used is index-v[23].
 #
+# Define GIT_TEST_UTF8_LOCALE to preferred utf-8 locale for testing.
+# If it isn't set, fallback to $LC_ALL, $LANG or use the first utf-8
+# locale returned by "locale -a".
+#
 # Define HAVE_CLOCK_GETTIME if your platform has clock_gettime.
 #
 # Define HAVE_CLOCK_MONOTONIC if your platform has CLOCK_MONOTONIC.
@@ -2802,6 +2806,9 @@
 ifdef GIT_TEST_CMP_USE_COPIED_CONTEXT
 	@echo GIT_TEST_CMP_USE_COPIED_CONTEXT=YesPlease >>$@+
 endif
+ifdef GIT_TEST_UTF8_LOCALE
+	@echo GIT_TEST_UTF8_LOCALE=\''$(subst ','\'',$(subst ','\'',$(GIT_TEST_UTF8_LOCALE)))'\' >>$@+
+endif
 	@echo NO_GETTEXT=\''$(subst ','\'',$(subst ','\'',$(NO_GETTEXT)))'\' >>$@+
 ifdef GIT_PERF_REPEAT_COUNT
 	@echo GIT_PERF_REPEAT_COUNT=\''$(subst ','\'',$(subst ','\'',$(GIT_PERF_REPEAT_COUNT)))'\' >>$@+
diff --git a/RelNotes b/RelNotes
index aece21e..f071367 120000
--- a/RelNotes
+++ b/RelNotes
@@ -1 +1 @@
-Documentation/RelNotes/2.32.0.txt
\ No newline at end of file
+Documentation/RelNotes/2.33.0.txt
\ No newline at end of file
diff --git a/builtin/clone.c b/builtin/clone.c
index eeb74c0..66fe666 100644
--- a/builtin/clone.c
+++ b/builtin/clone.c
@@ -1320,9 +1320,8 @@
 			}
 
 		if (!is_local && !complete_refs_before_fetch) {
-			err = transport_fetch_refs(transport, mapped_refs);
-			if (err)
-				goto cleanup;
+			if (transport_fetch_refs(transport, mapped_refs))
+				die(_("remote transport reported error"));
 		}
 
 		remote_head = find_ref_by_name(refs, "HEAD");
@@ -1380,9 +1379,8 @@
 	if (is_local)
 		clone_local(path, git_dir);
 	else if (refs && complete_refs_before_fetch) {
-		err = transport_fetch_refs(transport, mapped_refs);
-		if (err)
-			goto cleanup;
+		if (transport_fetch_refs(transport, mapped_refs))
+			die(_("remote transport reported error"));
 	}
 
 	update_remote_refs(refs, mapped_refs, remote_head_points_at,
@@ -1410,7 +1408,6 @@
 	junk_mode = JUNK_LEAVE_REPO;
 	err = checkout(submodule_progress);
 
-cleanup:
 	free(remote_name);
 	strbuf_release(&reflog_msg);
 	strbuf_release(&branch_top);
diff --git a/builtin/diff-index.c b/builtin/diff-index.c
index 176fe7f..cf09559 100644
--- a/builtin/diff-index.c
+++ b/builtin/diff-index.c
@@ -2,6 +2,7 @@
 #include "cache.h"
 #include "config.h"
 #include "diff.h"
+#include "diff-merges.h"
 #include "commit.h"
 #include "revision.h"
 #include "builtin.h"
@@ -27,6 +28,12 @@
 	rev.abbrev = 0;
 	prefix = precompose_argv_prefix(argc, argv, prefix);
 
+	/*
+	 * We need no diff for merges options, and we need to avoid conflict
+	 * with our own meaning of "-m".
+	 */
+	diff_merges_suppress_options_parsing();
+
 	argc = setup_revisions(argc, argv, &rev, NULL);
 	for (i = 1; i < argc; i++) {
 		const char *arg = argv[i];
@@ -35,6 +42,8 @@
 			option |= DIFF_INDEX_CACHED;
 		else if (!strcmp(arg, "--merge-base"))
 			option |= DIFF_INDEX_MERGE_BASE;
+		else if (!strcmp(arg, "-m"))
+			rev.match_missing = 1;
 		else
 			usage(diff_cache_usage);
 	}
diff --git a/builtin/fetch.c b/builtin/fetch.c
index dfde96a..9191620 100644
--- a/builtin/fetch.c
+++ b/builtin/fetch.c
@@ -1126,7 +1126,7 @@
 
 			if (rm->status == REF_STATUS_REJECT_SHALLOW) {
 				if (want_status == FETCH_HEAD_MERGE)
-					warning(_("reject %s because shallow roots are not allowed to be updated"),
+					warning(_("rejected %s because shallow roots are not allowed to be updated"),
 						rm->peer_ref ? rm->peer_ref->name : rm->name);
 				continue;
 			}
diff --git a/builtin/merge.c b/builtin/merge.c
index eddb8ae..a8a843b 100644
--- a/builtin/merge.c
+++ b/builtin/merge.c
@@ -56,8 +56,8 @@
 
 static const char * const builtin_merge_usage[] = {
 	N_("git merge [<options>] [<commit>...]"),
-	N_("git merge --abort"),
-	N_("git merge --continue"),
+	"git merge --abort",
+	"git merge --continue",
 	NULL
 };
 
diff --git a/builtin/rev-parse.c b/builtin/rev-parse.c
index 7af8dab..22c4e1a 100644
--- a/builtin/rev-parse.c
+++ b/builtin/rev-parse.c
@@ -435,11 +435,11 @@
 	/* get the usage up to the first line with a -- on it */
 	for (;;) {
 		if (strbuf_getline(&sb, stdin) == EOF)
-			die("premature end of input");
+			die(_("premature end of input"));
 		ALLOC_GROW(usage, unb + 1, usz);
 		if (!strcmp("--", sb.buf)) {
 			if (unb < 1)
-				die("no usage string given before the `--' separator");
+				die(_("no usage string given before the `--' separator"));
 			usage[unb] = NULL;
 			break;
 		}
@@ -545,7 +545,7 @@
 	if (quiet)
 		exit(1);
 	else
-		die("Needed a single revision");
+		die(_("Needed a single revision"));
 }
 
 static const char builtin_rev_parse_usage[] =
@@ -709,10 +709,10 @@
 			if (!strcmp(arg, "--resolve-git-dir")) {
 				const char *gitdir = argv[++i];
 				if (!gitdir)
-					die("--resolve-git-dir requires an argument");
+					die(_("--resolve-git-dir requires an argument"));
 				gitdir = resolve_gitdir(gitdir);
 				if (!gitdir)
-					die("not a gitdir '%s'", argv[i]);
+					die(_("not a gitdir '%s'"), argv[i]);
 				puts(gitdir);
 				continue;
 			}
@@ -736,7 +736,7 @@
 		if (!seen_end_of_options && *arg == '-') {
 			if (!strcmp(arg, "--git-path")) {
 				if (!argv[i + 1])
-					die("--git-path requires an argument");
+					die(_("--git-path requires an argument"));
 				strbuf_reset(&buf);
 				print_path(git_path("%s", argv[i + 1]), prefix,
 						format,
@@ -746,7 +746,7 @@
 			}
 			if (!strcmp(arg,"-n")) {
 				if (++i >= argc)
-					die("-n requires an argument");
+					die(_("-n requires an argument"));
 				if ((filter & DO_FLAGS) && (filter & DO_REVS)) {
 					show(arg);
 					show(argv[i]);
@@ -760,26 +760,26 @@
 			}
 			if (opt_with_value(arg, "--path-format", &arg)) {
 				if (!arg)
-					die("--path-format requires an argument");
+					die(_("--path-format requires an argument"));
 				if (!strcmp(arg, "absolute")) {
 					format = FORMAT_CANONICAL;
 				} else if (!strcmp(arg, "relative")) {
 					format = FORMAT_RELATIVE;
 				} else {
-					die("unknown argument to --path-format: %s", arg);
+					die(_("unknown argument to --path-format: %s"), arg);
 				}
 				continue;
 			}
 			if (!strcmp(arg, "--default")) {
 				def = argv[++i];
 				if (!def)
-					die("--default requires an argument");
+					die(_("--default requires an argument"));
 				continue;
 			}
 			if (!strcmp(arg, "--prefix")) {
 				prefix = argv[++i];
 				if (!prefix)
-					die("--prefix requires an argument");
+					die(_("--prefix requires an argument"));
 				startup_info->prefix = prefix;
 				output_prefix = 1;
 				continue;
@@ -848,7 +848,7 @@
 					else if (!strcmp(arg, "loose"))
 						abbrev_ref_strict = 0;
 					else
-						die("unknown mode for --abbrev-ref: %s",
+						die(_("unknown mode for --abbrev-ref: %s"),
 						    arg);
 				}
 				continue;
@@ -892,7 +892,7 @@
 				if (work_tree)
 					print_path(work_tree, prefix, format, DEFAULT_UNMODIFIED);
 				else
-					die("this operation must be run in a work tree");
+					die(_("this operation must be run in a work tree"));
 				continue;
 			}
 			if (!strcmp(arg, "--show-superproject-working-tree")) {
@@ -1020,7 +1020,7 @@
 				if (strcmp(val, "storage") &&
 				    strcmp(val, "input") &&
 				    strcmp(val, "output"))
-					die("unknown mode for --show-object-format: %s",
+					die(_("unknown mode for --show-object-format: %s"),
 					    arg);
 				puts(the_hash_algo->name);
 				continue;
@@ -1058,7 +1058,7 @@
 		if (verify)
 			die_no_single_rev(quiet);
 		if (has_dashdash)
-			die("bad revision '%s'", arg);
+			die(_("bad revision '%s'"), arg);
 		as_is = 1;
 		if (!show_file(arg, output_prefix))
 			continue;
diff --git a/builtin/stash.c b/builtin/stash.c
index 01066d7..9c72e4b 100644
--- a/builtin/stash.c
+++ b/builtin/stash.c
@@ -26,7 +26,7 @@
 	N_("git stash drop [-q|--quiet] [<stash>]"),
 	N_("git stash ( pop | apply ) [--index] [-q|--quiet] [<stash>]"),
 	N_("git stash branch <branchname> [<stash>]"),
-	N_("git stash clear"),
+	"git stash clear",
 	N_("git stash [push [-p|--patch] [-k|--[no-]keep-index] [-q|--quiet]\n"
 	   "          [-u|--include-untracked] [-a|--all] [-m|--message <message>]\n"
 	   "          [--pathspec-from-file=<file> [--pathspec-file-nul]]\n"
@@ -67,7 +67,7 @@
 };
 
 static const char * const git_stash_clear_usage[] = {
-	N_("git stash clear"),
+	"git stash clear",
 	NULL
 };
 
@@ -761,7 +761,7 @@
 
 	cp.git_cmd = 1;
 	strvec_pushl(&cp.args, "log", "--format=%gd: %gs", "-g",
-		     "--first-parent", "-m", NULL);
+		     "--first-parent", NULL);
 	strvec_pushv(&cp.args, argv);
 	strvec_push(&cp.args, ref_stash);
 	strvec_push(&cp.args, "--");
diff --git a/builtin/submodule--helper.c b/builtin/submodule--helper.c
index d55f626..ae6174a 100644
--- a/builtin/submodule--helper.c
+++ b/builtin/submodule--helper.c
@@ -1300,7 +1300,7 @@
 		OPT_BOOL(0, "cached", &cached,
 			 N_("use the commit stored in the index instead of the submodule HEAD")),
 		OPT_BOOL(0, "files", &files,
-			 N_("to compare the commit in the index with that in the submodule HEAD")),
+			 N_("compare the commit in the index with that in the submodule HEAD")),
 		OPT_BOOL(0, "for-status", &for_status,
 			 N_("skip submodules with 'ignore_config' value set to 'all'")),
 		OPT_INTEGER('n', "summary-limit", &summary_limit,
diff --git a/chunk-format.c b/chunk-format.c
index da191e5..1c3dca6 100644
--- a/chunk-format.c
+++ b/chunk-format.c
@@ -58,9 +58,11 @@
 
 int write_chunkfile(struct chunkfile *cf, void *data)
 {
-	int i;
+	int i, result = 0;
 	uint64_t cur_offset = hashfile_total(cf->f);
 
+	trace2_region_enter("chunkfile", "write", the_repository);
+
 	/* Add the table of contents to the current offset */
 	cur_offset += (cf->chunks_nr + 1) * CHUNK_TOC_ENTRY_SIZE;
 
@@ -77,10 +79,10 @@
 
 	for (i = 0; i < cf->chunks_nr; i++) {
 		off_t start_offset = hashfile_total(cf->f);
-		int result = cf->chunks[i].write_fn(cf->f, data);
+		result = cf->chunks[i].write_fn(cf->f, data);
 
 		if (result)
-			return result;
+			goto cleanup;
 
 		if (hashfile_total(cf->f) - start_offset != cf->chunks[i].size)
 			BUG("expected to write %"PRId64" bytes to chunk %"PRIx32", but wrote %"PRId64" instead",
@@ -88,7 +90,9 @@
 			    hashfile_total(cf->f) - start_offset);
 	}
 
-	return 0;
+cleanup:
+	trace2_region_leave("chunkfile", "write", the_repository);
+	return result;
 }
 
 int read_table_of_contents(struct chunkfile *cf,
diff --git a/ci/lib.sh b/ci/lib.sh
index d848c03..476c3f3 100755
--- a/ci/lib.sh
+++ b/ci/lib.sh
@@ -229,6 +229,7 @@
 	CC=gcc
 	MAKEFLAGS="$MAKEFLAGS PYTHON_PATH=/usr/bin/python3 USE_LIBPCRE2=Yes"
 	MAKEFLAGS="$MAKEFLAGS NO_REGEX=Yes ICONV_OMITS_BOM=Yes"
+	MAKEFLAGS="$MAKEFLAGS GIT_TEST_UTF8_LOCALE=C.UTF-8"
 	;;
 esac
 
diff --git a/contrib/completion/git-completion.bash b/contrib/completion/git-completion.bash
index b50c5d0..4073d67 100644
--- a/contrib/completion/git-completion.bash
+++ b/contrib/completion/git-completion.bash
@@ -1729,6 +1729,7 @@
 			--indent-heuristic --no-indent-heuristic
 			--textconv --no-textconv
 			--patch --no-patch
+			--anchored=
 "
 
 __git_diff_difftool_options="--cached --staged --pickaxe-all --pickaxe-regex
diff --git a/contrib/subtree/git-subtree.sh b/contrib/subtree/git-subtree.sh
index b06782b..7f767b5 100755
--- a/contrib/subtree/git-subtree.sh
+++ b/contrib/subtree/git-subtree.sh
@@ -5,8 +5,12 @@
 # Copyright (C) 2009 Avery Pennarun <apenwarr@gmail.com>
 #
 
-if test -z "$GIT_EXEC_PATH" || test "${PATH#"${GIT_EXEC_PATH}:"}" = "$PATH" || ! test -f "$GIT_EXEC_PATH/git-sh-setup"
+if test -z "$GIT_EXEC_PATH" || ! test -f "$GIT_EXEC_PATH/git-sh-setup" || {
+	test "${PATH#"${GIT_EXEC_PATH}:"}" = "$PATH" &&
+	test ! "$GIT_EXEC_PATH" -ef "${PATH%%:*}" 2>/dev/null
+}
 then
+	basename=${0##*[/\\]}
 	echo >&2 'It looks like either your git installation or your'
 	echo >&2 'git-subtree installation is broken.'
 	echo >&2
@@ -14,10 +18,10 @@
 	echo >&2 " - If \`git --exec-path\` does not print the correct path to"
 	echo >&2 "   your git install directory, then set the GIT_EXEC_PATH"
 	echo >&2 "   environment variable to the correct directory."
-	echo >&2 " - Make sure that your \`${0##*/}\` file is either in your"
+	echo >&2 " - Make sure that your \`$basename\` file is either in your"
 	echo >&2 "   PATH or in your git exec path (\`$(git --exec-path)\`)."
-	echo >&2 " - You should run git-subtree as \`git ${0##*/git-}\`,"
-	echo >&2 "   not as \`${0##*/}\`." >&2
+	echo >&2 " - You should run git-subtree as \`git ${basename#git-}\`,"
+	echo >&2 "   not as \`$basename\`." >&2
 	exit 126
 fi
 
diff --git a/csum-file.c b/csum-file.c
index 7510950..3487d28 100644
--- a/csum-file.c
+++ b/csum-file.c
@@ -11,35 +11,33 @@
 #include "progress.h"
 #include "csum-file.h"
 
+static void verify_buffer_or_die(struct hashfile *f,
+				 const void *buf,
+				 unsigned int count)
+{
+	ssize_t ret = read_in_full(f->check_fd, f->check_buffer, count);
+
+	if (ret < 0)
+		die_errno("%s: sha1 file read error", f->name);
+	if (ret != count)
+		die("%s: sha1 file truncated", f->name);
+	if (memcmp(buf, f->check_buffer, count))
+		die("sha1 file '%s' validation error", f->name);
+}
+
 static void flush(struct hashfile *f, const void *buf, unsigned int count)
 {
-	if (0 <= f->check_fd && count)  {
-		unsigned char check_buffer[8192];
-		ssize_t ret = read_in_full(f->check_fd, check_buffer, count);
+	if (0 <= f->check_fd && count)
+		verify_buffer_or_die(f, buf, count);
 
-		if (ret < 0)
-			die_errno("%s: sha1 file read error", f->name);
-		if (ret != count)
-			die("%s: sha1 file truncated", f->name);
-		if (memcmp(buf, check_buffer, count))
-			die("sha1 file '%s' validation error", f->name);
-	}
-
-	for (;;) {
-		int ret = xwrite(f->fd, buf, count);
-		if (ret > 0) {
-			f->total += ret;
-			display_throughput(f->tp, f->total);
-			buf = (char *) buf + ret;
-			count -= ret;
-			if (count)
-				continue;
-			return;
-		}
-		if (!ret)
+	if (write_in_full(f->fd, buf, count) < 0) {
+		if (errno == ENOSPC)
 			die("sha1 file '%s' write error. Out of diskspace", f->name);
 		die_errno("sha1 file '%s' write error", f->name);
 	}
+
+	f->total += count;
+	display_throughput(f->tp, f->total);
 }
 
 void hashflush(struct hashfile *f)
@@ -53,6 +51,13 @@
 	}
 }
 
+static void free_hashfile(struct hashfile *f)
+{
+	free(f->buffer);
+	free(f->check_buffer);
+	free(f);
+}
+
 int finalize_hashfile(struct hashfile *f, unsigned char *result, unsigned int flags)
 {
 	int fd;
@@ -82,20 +87,20 @@
 		if (close(f->check_fd))
 			die_errno("%s: sha1 file error on close", f->name);
 	}
-	free(f);
+	free_hashfile(f);
 	return fd;
 }
 
 void hashwrite(struct hashfile *f, const void *buf, unsigned int count)
 {
 	while (count) {
-		unsigned left = sizeof(f->buffer) - f->offset;
+		unsigned left = f->buffer_len - f->offset;
 		unsigned nr = count > left ? left : count;
 
 		if (f->do_crc)
 			f->crc32 = crc32(f->crc32, buf, nr);
 
-		if (nr == sizeof(f->buffer)) {
+		if (nr == f->buffer_len) {
 			/*
 			 * Flush a full batch worth of data directly
 			 * from the input, skipping the memcpy() to
@@ -121,11 +126,6 @@
 	}
 }
 
-struct hashfile *hashfd(int fd, const char *name)
-{
-	return hashfd_throughput(fd, name, NULL);
-}
-
 struct hashfile *hashfd_check(const char *name)
 {
 	int sink, check;
@@ -139,10 +139,14 @@
 		die_errno("unable to open '%s'", name);
 	f = hashfd(sink, name);
 	f->check_fd = check;
+	f->check_buffer = xmalloc(f->buffer_len);
+
 	return f;
 }
 
-struct hashfile *hashfd_throughput(int fd, const char *name, struct progress *tp)
+static struct hashfile *hashfd_internal(int fd, const char *name,
+					struct progress *tp,
+					size_t buffer_len)
 {
 	struct hashfile *f = xmalloc(sizeof(*f));
 	f->fd = fd;
@@ -153,9 +157,35 @@
 	f->name = name;
 	f->do_crc = 0;
 	the_hash_algo->init_fn(&f->ctx);
+
+	f->buffer_len = buffer_len;
+	f->buffer = xmalloc(buffer_len);
+	f->check_buffer = NULL;
+
 	return f;
 }
 
+struct hashfile *hashfd(int fd, const char *name)
+{
+	/*
+	 * Since we are not going to use a progress meter to
+	 * measure the rate of data passing through this hashfile,
+	 * use a larger buffer size to reduce fsync() calls.
+	 */
+	return hashfd_internal(fd, name, NULL, 128 * 1024);
+}
+
+struct hashfile *hashfd_throughput(int fd, const char *name, struct progress *tp)
+{
+	/*
+	 * Since we are expecting to report progress of the
+	 * write into this hashfile, use a smaller buffer
+	 * size so the progress indicators arrive at a more
+	 * frequent rate.
+	 */
+	return hashfd_internal(fd, name, tp, 8 * 1024);
+}
+
 void hashfile_checkpoint(struct hashfile *f, struct hashfile_checkpoint *checkpoint)
 {
 	hashflush(f);
diff --git a/csum-file.h b/csum-file.h
index e54d53d..3044bd1 100644
--- a/csum-file.h
+++ b/csum-file.h
@@ -16,7 +16,9 @@
 	const char *name;
 	int do_crc;
 	uint32_t crc32;
-	unsigned char buffer[8192];
+	size_t buffer_len;
+	unsigned char *buffer;
+	unsigned char *check_buffer;
 };
 
 /* Checkpoint */
diff --git a/diff-merges.c b/diff-merges.c
index f3a9dae..0dfcaa1 100644
--- a/diff-merges.c
+++ b/diff-merges.c
@@ -6,6 +6,7 @@
 static void set_separate(struct rev_info *revs);
 
 static diff_merges_setup_func_t set_to_default = set_separate;
+static int suppress_parsing;
 
 static void suppress(struct rev_info *revs)
 {
@@ -14,7 +15,7 @@
 	revs->combine_merges = 0;
 	revs->dense_combined_merges = 0;
 	revs->combined_all_paths = 0;
-	revs->combined_imply_patch = 0;
+	revs->merges_imply_patch = 0;
 	revs->merges_need_diff = 0;
 }
 
@@ -30,17 +31,6 @@
 	revs->first_parent_merges = 1;
 }
 
-static void set_m(struct rev_info *revs)
-{
-	/*
-	 * To "diff-index", "-m" means "match missing", and to the "log"
-	 * family of commands, it means "show default diff for merges". Set
-	 * both fields appropriately.
-	 */
-	set_to_default(revs);
-	revs->match_missing = 1;
-}
-
 static void set_combined(struct rev_info *revs)
 {
 	suppress(revs);
@@ -101,20 +91,29 @@
 	return 0;
 }
 
+void diff_merges_suppress_options_parsing(void)
+{
+	suppress_parsing = 1;
+}
+
 int diff_merges_parse_opts(struct rev_info *revs, const char **argv)
 {
 	int argcount = 1;
 	const char *optarg;
 	const char *arg = argv[0];
 
+	if (suppress_parsing)
+		return 0;
+
 	if (!strcmp(arg, "-m")) {
-		set_m(revs);
+		set_to_default(revs);
+		revs->merges_imply_patch = 1;
 	} else if (!strcmp(arg, "-c")) {
 		set_combined(revs);
-		revs->combined_imply_patch = 1;
+		revs->merges_imply_patch = 1;
 	} else if (!strcmp(arg, "--cc")) {
 		set_dense_combined(revs);
-		revs->combined_imply_patch = 1;
+		revs->merges_imply_patch = 1;
 	} else if (!strcmp(arg, "--no-diff-merges")) {
 		suppress(revs);
 	} else if (!strcmp(arg, "--combined-all-paths")) {
@@ -155,15 +154,18 @@
 
 void diff_merges_setup_revs(struct rev_info *revs)
 {
+	if (suppress_parsing)
+		return;
+
 	if (revs->combine_merges == 0)
 		revs->dense_combined_merges = 0;
 	if (revs->separate_merges == 0)
 		revs->first_parent_merges = 0;
 	if (revs->combined_all_paths && !revs->combine_merges)
 		die("--combined-all-paths makes no sense without -c or --cc");
-	if (revs->combined_imply_patch)
+	if (revs->merges_imply_patch)
 		revs->diff = 1;
-	if (revs->combined_imply_patch || revs->merges_need_diff) {
+	if (revs->merges_imply_patch || revs->merges_need_diff) {
 		if (!revs->diffopt.output_format)
 			revs->diffopt.output_format = DIFF_FORMAT_PATCH;
 	}
diff --git a/diff-merges.h b/diff-merges.h
index 09d9a6c..b5d57f6 100644
--- a/diff-merges.h
+++ b/diff-merges.h
@@ -11,6 +11,8 @@
 
 int diff_merges_config(const char *value);
 
+void diff_merges_suppress_options_parsing(void);
+
 int diff_merges_parse_opts(struct rev_info *revs, const char **argv);
 
 void diff_merges_suppress(struct rev_info *revs);
diff --git a/diffcore-rename.c b/diffcore-rename.c
index 963ca58..3375e24 100644
--- a/diffcore-rename.c
+++ b/diffcore-rename.c
@@ -568,7 +568,8 @@
 static void initialize_dir_rename_info(struct dir_rename_info *info,
 				       struct strintmap *relevant_sources,
 				       struct strintmap *dirs_removed,
-				       struct strmap *dir_rename_count)
+				       struct strmap *dir_rename_count,
+				       struct strmap *cached_pairs)
 {
 	struct hashmap_iter iter;
 	struct strmap_entry *entry;
@@ -633,6 +634,17 @@
 					 rename_dst[i].p->two->path);
 	}
 
+	/* Add cached_pairs to counts */
+	strmap_for_each_entry(cached_pairs, &iter, entry) {
+		const char *old_name = entry->key;
+		const char *new_name = entry->value;
+		if (!new_name)
+			/* known delete; ignore it */
+			continue;
+
+		update_dir_rename_counts(info, dirs_removed, old_name, new_name);
+	}
+
 	/*
 	 * Now we collapse
 	 *    dir_rename_count: old_directory -> {new_directory -> count}
@@ -1247,7 +1259,8 @@
 void diffcore_rename_extended(struct diff_options *options,
 			      struct strintmap *relevant_sources,
 			      struct strintmap *dirs_removed,
-			      struct strmap *dir_rename_count)
+			      struct strmap *dir_rename_count,
+			      struct strmap *cached_pairs)
 {
 	int detect_rename = options->detect_rename;
 	int minimum_score = options->rename_score;
@@ -1363,7 +1376,8 @@
 		/* Preparation for basename-driven matching. */
 		trace2_region_enter("diff", "dir rename setup", options->repo);
 		initialize_dir_rename_info(&info, relevant_sources,
-					   dirs_removed, dir_rename_count);
+					   dirs_removed, dir_rename_count,
+					   cached_pairs);
 		trace2_region_leave("diff", "dir rename setup", options->repo);
 
 		/* Utilize file basenames to quickly find renames. */
@@ -1560,5 +1574,5 @@
 
 void diffcore_rename(struct diff_options *options)
 {
-	diffcore_rename_extended(options, NULL, NULL, NULL);
+	diffcore_rename_extended(options, NULL, NULL, NULL, NULL);
 }
diff --git a/diffcore.h b/diffcore.h
index f5c6de4..533b30e 100644
--- a/diffcore.h
+++ b/diffcore.h
@@ -181,7 +181,8 @@
 void diffcore_rename_extended(struct diff_options *options,
 			      struct strintmap *relevant_sources,
 			      struct strintmap *dirs_removed,
-			      struct strmap *dir_rename_count);
+			      struct strmap *dir_rename_count,
+			      struct strmap *cached_pairs);
 void diffcore_merge_broken(void);
 void diffcore_pickaxe(struct diff_options *);
 void diffcore_order(const char *orderfile);
diff --git a/fetch-pack.c b/fetch-pack.c
index c135635..b0c7be7 100644
--- a/fetch-pack.c
+++ b/fetch-pack.c
@@ -1645,6 +1645,15 @@
 			if (process_section_header(&reader, "packfile-uris", 1))
 				receive_packfile_uris(&reader, &packfile_uris);
 			process_section_header(&reader, "packfile", 0);
+
+			/*
+			 * this is the final request we'll make of the server;
+			 * do a half-duplex shutdown to indicate that they can
+			 * hang up as soon as the pack is sent.
+			 */
+			close(fd[1]);
+			fd[1] = -1;
+
 			if (get_pack(args, fd, pack_lockfiles,
 				     packfile_uris.nr ? &index_pack_args : NULL,
 				     sought, nr_sought, &fsck_options.gitmodules_found))
diff --git a/git-compat-util.h b/git-compat-util.h
index a508dbe..fb6e9af 100644
--- a/git-compat-util.h
+++ b/git-compat-util.h
@@ -986,11 +986,9 @@
 
 static inline size_t xsize_t(off_t len)
 {
-	size_t size = (size_t) len;
-
-	if (len != (off_t) size)
+	if (len < 0 || (uintmax_t) len > SIZE_MAX)
 		die("Cannot handle files this big");
-	return size;
+	return (size_t) len;
 }
 
 __attribute__((format (printf, 3, 4)))
diff --git a/git-send-email.perl b/git-send-email.perl
index 25be2eb..7ba0b34 100755
--- a/git-send-email.perl
+++ b/git-send-email.perl
@@ -70,6 +70,7 @@
 
   Sending:
     --envelope-sender       <str>  * Email envelope sender.
+    --sendmail-cmd          <str>  * Command to run to send email.
     --smtp-server       <str:int>  * Outgoing SMTP server to use. The port
                                      is optional. Default 'localhost'.
     --smtp-server-option    <str>  * Outgoing SMTP server option to use.
@@ -262,6 +263,7 @@
 my (@suppress_cc);
 my ($auto_8bit_encoding);
 my ($compose_encoding);
+my ($sendmail_cmd);
 # Variables with corresponding config settings & hardcoded defaults
 my ($debug_net_smtp) = 0;		# Net::SMTP, see send_message()
 my $thread = 1;
@@ -309,6 +311,7 @@
     "assume8bitencoding" => \$auto_8bit_encoding,
     "composeencoding" => \$compose_encoding,
     "transferencoding" => \$target_xfer_encoding,
+    "sendmailcmd" => \$sendmail_cmd,
 );
 
 my %config_path_settings = (
@@ -442,6 +445,7 @@
 		    "no-bcc" => \$no_bcc,
 		    "chain-reply-to!" => \$chain_reply_to,
 		    "no-chain-reply-to" => sub {$chain_reply_to = 0},
+		    "sendmail-cmd=s" => \$sendmail_cmd,
 		    "smtp-server=s" => \$smtp_server,
 		    "smtp-server-option=s" => \@smtp_server_options,
 		    "smtp-server-port=s" => \$smtp_server_port,
@@ -1013,16 +1017,19 @@
 	$reply_to = sanitize_address($reply_to);
 }
 
-if (!defined $smtp_server) {
+if (!defined $sendmail_cmd && !defined $smtp_server) {
 	my @sendmail_paths = qw( /usr/sbin/sendmail /usr/lib/sendmail );
 	push @sendmail_paths, map {"$_/sendmail"} split /:/, $ENV{PATH};
 	foreach (@sendmail_paths) {
 		if (-x $_) {
-			$smtp_server = $_;
+			$sendmail_cmd = $_;
 			last;
 		}
 	}
-	$smtp_server ||= 'localhost'; # could be 127.0.0.1, too... *shrug*
+
+	if (!defined $sendmail_cmd) {
+		$smtp_server = 'localhost'; # could be 127.0.0.1, too... *shrug*
+	}
 }
 
 if ($compose && $compose > 0) {
@@ -1502,11 +1509,17 @@
 
 	if ($dry_run) {
 		# We don't want to send the email.
-	} elsif (file_name_is_absolute($smtp_server)) {
+	} elsif (defined $sendmail_cmd || file_name_is_absolute($smtp_server)) {
 		my $pid = open my $sm, '|-';
 		defined $pid or die $!;
 		if (!$pid) {
-			exec($smtp_server, @sendmail_parameters) or die $!;
+			if (defined $sendmail_cmd) {
+				exec ("sh", "-c", "$sendmail_cmd \"\$@\"", "-", @sendmail_parameters)
+					or die $!;
+			} else {
+				exec ($smtp_server, @sendmail_parameters)
+					or die $!;
+			}
 		}
 		print $sm "$header\n$message";
 		close $sm or die $!;
@@ -1602,14 +1615,21 @@
 		printf($dry_run ? __("Dry-Sent %s\n") : __("Sent %s\n"), $subject);
 	} else {
 		print($dry_run ? __("Dry-OK. Log says:\n") : __("OK. Log says:\n"));
-		if (!file_name_is_absolute($smtp_server)) {
+		if (!defined $sendmail_cmd && !file_name_is_absolute($smtp_server)) {
 			print "Server: $smtp_server\n";
 			print "MAIL FROM:<$raw_from>\n";
 			foreach my $entry (@recipients) {
 			    print "RCPT TO:<$entry>\n";
 			}
 		} else {
-			print "Sendmail: $smtp_server ".join(' ',@sendmail_parameters)."\n";
+			my $sm;
+			if (defined $sendmail_cmd) {
+				$sm = $sendmail_cmd;
+			} else {
+				$sm = $smtp_server;
+			}
+
+			print "Sendmail: $sm ".join(' ',@sendmail_parameters)."\n";
 		}
 		print $header, "\n";
 		if ($smtp) {
diff --git a/hash.h b/hash.h
index 2986f99..9c6df4d 100644
--- a/hash.h
+++ b/hash.h
@@ -263,6 +263,22 @@
 	dst->algo = src->algo;
 }
 
+/* Like oidcpy() but zero-pads the unused bytes in dst's hash array. */
+static inline void oidcpy_with_padding(struct object_id *dst,
+				       struct object_id *src)
+{
+	size_t hashsz;
+
+	if (!src->algo)
+		hashsz = the_hash_algo->rawsz;
+	else
+		hashsz = hash_algos[src->algo].rawsz;
+
+	memcpy(dst->hash, src->hash, hashsz);
+	memset(dst->hash + hashsz, 0, GIT_MAX_RAWSZ - hashsz);
+	dst->algo = src->algo;
+}
+
 static inline struct object_id *oiddup(const struct object_id *src)
 {
 	struct object_id *dst = xmalloc(sizeof(struct object_id));
diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 96a605c..fd8d59f 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -102,7 +102,7 @@
 	} else if (skip_prefix(arg, "object:type=", &v0)) {
 		int type = type_from_string_gently(v0, strlen(v0), 1);
 		if (type < 0) {
-			strbuf_addf(errbuf, _("'%s' for 'object:type=<type>' is"
+			strbuf_addf(errbuf, _("'%s' for 'object:type=<type>' is "
 					      "not a valid object type"), v0);
 			return 1;
 		}
diff --git a/mailinfo.c b/mailinfo.c
index ccc6beb..184ed8d 100644
--- a/mailinfo.c
+++ b/mailinfo.c
@@ -19,7 +19,7 @@
 static void get_sane_name(struct strbuf *out, struct strbuf *name, struct strbuf *email)
 {
 	struct strbuf *src = name;
-	if (name->len < 3 || 60 < name->len || strpbrk(name->buf, "@<>"))
+	if (!name->len || 60 < name->len || strpbrk(name->buf, "@<>"))
 		src = email;
 	else if (name == out)
 		return;
diff --git a/merge-ort.c b/merge-ort.c
index 4a9ce2a..b954f71 100644
--- a/merge-ort.c
+++ b/merge-ort.c
@@ -53,6 +53,8 @@
 	MERGE_SIDE2 = 2
 };
 
+static unsigned RESULT_INITIALIZED = 0x1abe11ed; /* unlikely accidental value */
+
 struct traversal_callback_data {
 	unsigned long mask;
 	unsigned long dirmask;
@@ -141,6 +143,72 @@
 	char *callback_data_traverse_path;
 
 	/*
+	 * merge_trees: trees passed to the merge algorithm for the merge
+	 *
+	 * merge_trees records the trees passed to the merge algorithm.  But,
+	 * this data also is stored in merge_result->priv.  If a sequence of
+	 * merges are being done (such as when cherry-picking or rebasing),
+	 * the next merge can look at this and re-use information from
+	 * previous merges under certain circumstances.
+	 *
+	 * See also all the cached_* variables.
+	 */
+	struct tree *merge_trees[3];
+
+	/*
+	 * cached_pairs_valid_side: which side's cached info can be reused
+	 *
+	 * See the description for merge_trees.  For repeated merges, at most
+	 * only one side's cached information can be used.  Valid values:
+	 *   MERGE_SIDE2: cached data from side2 can be reused
+	 *   MERGE_SIDE1: cached data from side1 can be reused
+	 *   0:           no cached data can be reused
+	 */
+	int cached_pairs_valid_side;
+
+	/*
+	 * cached_pairs: Caching of renames and deletions.
+	 *
+	 * These are mappings recording renames and deletions of individual
+	 * files (not directories).  They are thus a map from an old
+	 * filename to either NULL (for deletions) or a new filename (for
+	 * renames).
+	 */
+	struct strmap cached_pairs[3];
+
+	/*
+	 * cached_target_names: just the destinations from cached_pairs
+	 *
+	 * We sometimes want a fast lookup to determine if a given filename
+	 * is one of the destinations in cached_pairs.  cached_target_names
+	 * is thus duplicative information, but it provides a fast lookup.
+	 */
+	struct strset cached_target_names[3];
+
+	/*
+	 * cached_irrelevant: Caching of rename_sources that aren't relevant.
+	 *
+	 * If we try to detect a rename for a source path and succeed, it's
+	 * part of a rename.  If we try to detect a rename for a source path
+	 * and fail, then it's a delete.  If we do not try to detect a rename
+	 * for a path, then we don't know if it's a rename or a delete.  If
+	 * merge-ort doesn't think the path is relevant, then we just won't
+	 * cache anything for that path.  But there's a slight problem in
+	 * that merge-ort can think a path is RELEVANT_LOCATION, but due to
+	 * commit 9bd342137e ("diffcore-rename: determine which
+	 * relevant_sources are no longer relevant", 2021-03-13),
+	 * diffcore-rename can downgrade the path to RELEVANT_NO_MORE.  To
+	 * avoid excessive calls to diffcore_rename_extended() we still need
+	 * to cache such paths, though we cannot record them as either
+	 * renames or deletes.  So we cache them here as a "turned out to be
+	 * irrelevant *for this commit*" as they are often also irrelevant
+	 * for subsequent commits, though we will have to do some extra
+	 * checking to see whether such paths become relevant for rename
+	 * detection when cherry-picking/rebasing subsequent commits.
+	 */
+	struct strset cached_irrelevant[3];
+
+	/*
 	 * needed_limit: value needed for inexact rename detection to run
 	 *
 	 * If the current rename limit wasn't high enough for inexact
@@ -382,6 +450,8 @@
 		reinitialize ? strmap_partial_clear : strmap_clear;
 	void (*strintmap_func)(struct strintmap *) =
 		reinitialize ? strintmap_partial_clear : strintmap_clear;
+	void (*strset_func)(struct strset *) =
+		reinitialize ? strset_partial_clear : strset_clear;
 
 	/*
 	 * We marked opti->paths with strdup_strings = 0, so that we
@@ -417,15 +487,21 @@
 	/* Free memory used by various renames maps */
 	for (i = MERGE_SIDE1; i <= MERGE_SIDE2; ++i) {
 		strintmap_func(&renames->dirs_removed[i]);
-
-		partial_clear_dir_rename_count(&renames->dir_rename_count[i]);
-		if (!reinitialize)
-			strmap_clear(&renames->dir_rename_count[i], 1);
-
 		strmap_func(&renames->dir_renames[i], 0);
-
 		strintmap_func(&renames->relevant_sources[i]);
+		if (!reinitialize)
+			assert(renames->cached_pairs_valid_side == 0);
+		if (i != renames->cached_pairs_valid_side) {
+			strset_func(&renames->cached_target_names[i]);
+			strmap_func(&renames->cached_pairs[i], 1);
+			strset_func(&renames->cached_irrelevant[i]);
+			partial_clear_dir_rename_count(&renames->dir_rename_count[i]);
+			if (!reinitialize)
+				strmap_clear(&renames->dir_rename_count[i], 1);
+		}
 	}
+	renames->cached_pairs_valid_side = 0;
+	renames->dir_rename_mask = 0;
 
 	if (!reinitialize) {
 		struct hashmap_iter iter;
@@ -448,8 +524,6 @@
 		strmap_clear(&opti->output, 0);
 	}
 
-	renames->dir_rename_mask = 0;
-
 	/* Clean out callback_data as well. */
 	FREE_AND_NULL(renames->callback_data);
 	renames->callback_data_nr = renames->callback_data_alloc = 0;
@@ -690,15 +764,48 @@
 	struct rename_info *renames = &opt->priv->renames;
 	int names_idx = is_add ? side : 0;
 
-	if (!is_add) {
+	if (is_add) {
+		if (strset_contains(&renames->cached_target_names[side],
+				    pathname))
+			return;
+	} else {
 		unsigned content_relevant = (match_mask == 0);
 		unsigned location_relevant = (dir_rename_mask == 0x07);
 
+		/*
+		 * If pathname is found in cached_irrelevant[side] due to
+		 * previous pick but for this commit content is relevant,
+		 * then we need to remove it from cached_irrelevant.
+		 */
+		if (content_relevant)
+			/* strset_remove is no-op if strset doesn't have key */
+			strset_remove(&renames->cached_irrelevant[side],
+				      pathname);
+
+		/*
+		 * We do not need to re-detect renames for paths that we already
+		 * know the pairing, i.e. for cached_pairs (or
+		 * cached_irrelevant).  However, handle_deferred_entries() needs
+		 * to loop over the union of keys from relevant_sources[side] and
+		 * cached_pairs[side], so for simplicity we set relevant_sources
+		 * for all the cached_pairs too and then strip them back out in
+		 * prune_cached_from_relevant() at the beginning of
+		 * detect_regular_renames().
+		 */
 		if (content_relevant || location_relevant) {
 			/* content_relevant trumps location_relevant */
 			strintmap_set(&renames->relevant_sources[side], pathname,
 				      content_relevant ? RELEVANT_CONTENT : RELEVANT_LOCATION);
 		}
+
+		/*
+		 * Avoid creating pair if we've already cached rename results.
+		 * Note that we do this after setting relevant_sources[side]
+		 * as noted in the comment above.
+		 */
+		if (strmap_contains(&renames->cached_pairs[side], pathname) ||
+		    strset_contains(&renames->cached_irrelevant[side], pathname))
+			return;
 	}
 
 	one = alloc_filespec(pathname);
@@ -2037,6 +2144,9 @@
 			VERIFY_CI(side2);
 
 			if (!strcmp(pathnames[1], pathnames[2])) {
+				struct rename_info *ri = &opt->priv->renames;
+				int j;
+
 				/* Both sides renamed the same way */
 				assert(side1 == side2);
 				memcpy(&side1->stages[0], &base->stages[0],
@@ -2046,6 +2156,16 @@
 				base->merged.is_null = 1;
 				base->merged.clean = 1;
 
+				/*
+				 * Disable remembering renames optimization;
+				 * rename/rename(1to1) is incredibly rare, and
+				 * just disabling the optimization is easier
+				 * than purging cached_pairs,
+				 * cached_target_names, and dir_rename_counts.
+				 */
+				for (j = 0; j < 3; j++)
+					ri->merge_trees[j] = NULL;
+
 				/* We handled both renames, i.e. i+1 handled */
 				i++;
 				/* Move to next rename */
@@ -2273,7 +2393,9 @@
 static inline int possible_renames(struct rename_info *renames)
 {
 	return possible_side_renames(renames, 1) ||
-	       possible_side_renames(renames, 2);
+	       possible_side_renames(renames, 2) ||
+	       !strmap_empty(&renames->cached_pairs[1]) ||
+	       !strmap_empty(&renames->cached_pairs[2]);
 }
 
 static void resolve_diffpair_statuses(struct diff_queue_struct *q)
@@ -2297,6 +2419,112 @@
 	}
 }
 
+static void prune_cached_from_relevant(struct rename_info *renames,
+				       unsigned side)
+{
+	/* Reason for this function described in add_pair() */
+	struct hashmap_iter iter;
+	struct strmap_entry *entry;
+
+	/* Remove from relevant_sources all entries in cached_pairs[side] */
+	strmap_for_each_entry(&renames->cached_pairs[side], &iter, entry) {
+		strintmap_remove(&renames->relevant_sources[side],
+				 entry->key);
+	}
+	/* Remove from relevant_sources all entries in cached_irrelevant[side] */
+	strset_for_each_entry(&renames->cached_irrelevant[side], &iter, entry) {
+		strintmap_remove(&renames->relevant_sources[side],
+				 entry->key);
+	}
+}
+
+static void use_cached_pairs(struct merge_options *opt,
+			     struct strmap *cached_pairs,
+			     struct diff_queue_struct *pairs)
+{
+	struct hashmap_iter iter;
+	struct strmap_entry *entry;
+
+	/*
+	 * Add to side_pairs all entries from renames->cached_pairs[side_index].
+	 * (Info in cached_irrelevant[side_index] is not relevant here.)
+	 */
+	strmap_for_each_entry(cached_pairs, &iter, entry) {
+		struct diff_filespec *one, *two;
+		const char *old_name = entry->key;
+		const char *new_name = entry->value;
+		if (!new_name)
+			new_name = old_name;
+
+		/* We don't care about oid/mode, only filenames and status */
+		one = alloc_filespec(old_name);
+		two = alloc_filespec(new_name);
+		diff_queue(pairs, one, two);
+		pairs->queue[pairs->nr-1]->status = entry->value ? 'R' : 'D';
+	}
+}
+
+static void cache_new_pair(struct rename_info *renames,
+			   int side,
+			   char *old_path,
+			   char *new_path,
+			   int free_old_value)
+{
+	char *old_value;
+	new_path = xstrdup(new_path);
+	old_value = strmap_put(&renames->cached_pairs[side],
+			       old_path, new_path);
+	strset_add(&renames->cached_target_names[side], new_path);
+	if (free_old_value)
+		free(old_value);
+	else
+		assert(!old_value);
+}
+
+static void possibly_cache_new_pair(struct rename_info *renames,
+				    struct diff_filepair *p,
+				    unsigned side,
+				    char *new_path)
+{
+	int dir_renamed_side = 0;
+
+	if (new_path) {
+		/*
+		 * Directory renames happen on the other side of history from
+		 * the side that adds new files to the old directory.
+		 */
+		dir_renamed_side = 3 - side;
+	} else {
+		int val = strintmap_get(&renames->relevant_sources[side],
+					p->one->path);
+		if (val == RELEVANT_NO_MORE) {
+			assert(p->status == 'D');
+			strset_add(&renames->cached_irrelevant[side],
+				   p->one->path);
+		}
+		if (val <= 0)
+			return;
+	}
+
+	if (p->status == 'D') {
+		/*
+		 * If we already had this delete, we'll just set it's value
+		 * to NULL again, so no harm.
+		 */
+		strmap_put(&renames->cached_pairs[side], p->one->path, NULL);
+	} else if (p->status == 'R') {
+		if (!new_path)
+			new_path = p->two->path;
+		else
+			cache_new_pair(renames, dir_renamed_side,
+				       p->two->path, new_path, 0);
+		cache_new_pair(renames, side, p->one->path, new_path, 1);
+	} else if (p->status == 'A' && new_path) {
+		cache_new_pair(renames, dir_renamed_side,
+			       p->two->path, new_path, 0);
+	}
+}
+
 static int compare_pairs(const void *a_, const void *b_)
 {
 	const struct diff_filepair *a = *((const struct diff_filepair **)a_);
@@ -2312,6 +2540,7 @@
 	struct diff_options diff_opts;
 	struct rename_info *renames = &opt->priv->renames;
 
+	prune_cached_from_relevant(renames, side_index);
 	if (!possible_side_renames(renames, side_index)) {
 		/*
 		 * No rename detection needed for this side, but we still need
@@ -2322,6 +2551,7 @@
 		return;
 	}
 
+	partial_clear_dir_rename_count(&renames->dir_rename_count[side_index]);
 	repo_diff_setup(opt->repo, &diff_opts);
 	diff_opts.flags.recursive = 1;
 	diff_opts.flags.rename_empty = 0;
@@ -2339,7 +2569,8 @@
 	diffcore_rename_extended(&diff_opts,
 				 &renames->relevant_sources[side_index],
 				 &renames->dirs_removed[side_index],
-				 &renames->dir_rename_count[side_index]);
+				 &renames->dir_rename_count[side_index],
+				 &renames->cached_pairs[side_index]);
 	trace2_region_leave("diff", "diffcore_rename", opt->repo);
 	resolve_diffpair_statuses(&diff_queued_diff);
 
@@ -2379,6 +2610,7 @@
 		char *new_path; /* non-NULL only with directory renames */
 
 		if (p->status != 'A' && p->status != 'R') {
+			possibly_cache_new_pair(renames, p, side_index, NULL);
 			diff_free_filepair(p);
 			continue;
 		}
@@ -2390,6 +2622,7 @@
 						      &collisions,
 						      &clean);
 
+		possibly_cache_new_pair(renames, p, side_index, new_path);
 		if (p->status != 'R' && !new_path) {
 			diff_free_filepair(p);
 			continue;
@@ -2445,6 +2678,8 @@
 	trace2_region_enter("merge", "regular renames", opt->repo);
 	detect_regular_renames(opt, MERGE_SIDE1);
 	detect_regular_renames(opt, MERGE_SIDE2);
+	use_cached_pairs(opt, &renames->cached_pairs[1], &renames->pairs[1]);
+	use_cached_pairs(opt, &renames->cached_pairs[2], &renames->pairs[2]);
 	trace2_region_leave("merge", "regular renames", opt->repo);
 
 	trace2_region_enter("merge", "directory renames", opt->repo);
@@ -3635,6 +3870,10 @@
 	assert(opt->obuf.len == 0);
 
 	assert(opt->priv == NULL);
+	if (result->_properly_initialized != 0 &&
+	    result->_properly_initialized != RESULT_INITIALIZED)
+		BUG("struct merge_result passed to merge_incore_*recursive() must be zeroed or filled with values from a previous run");
+	assert(!!result->priv == !!result->_properly_initialized);
 	if (result->priv) {
 		opt->priv = result->priv;
 		result->priv = NULL;
@@ -3674,8 +3913,22 @@
 					 NULL, 1);
 		strmap_init_with_options(&renames->dir_renames[i],
 					 NULL, 0);
+		/*
+		 * relevant_sources uses -1 for the default, because we need
+		 * to be able to distinguish not-in-strintmap from valid
+		 * relevant_source values from enum file_rename_relevance.
+		 * In particular, possibly_cache_new_pair() expects a negative
+		 * value for not-found entries.
+		 */
 		strintmap_init_with_options(&renames->relevant_sources[i],
-					    0, NULL, 0);
+					    -1 /* explicitly invalid */,
+					    NULL, 0);
+		strmap_init_with_options(&renames->cached_pairs[i],
+					 NULL, 1);
+		strset_init_with_options(&renames->cached_irrelevant[i],
+					 NULL, 1);
+		strset_init_with_options(&renames->cached_target_names[i],
+					 NULL, 0);
 	}
 
 	/*
@@ -3701,6 +3954,50 @@
 	trace2_region_leave("merge", "allocate/init", opt->repo);
 }
 
+static void merge_check_renames_reusable(struct merge_options *opt,
+					 struct merge_result *result,
+					 struct tree *merge_base,
+					 struct tree *side1,
+					 struct tree *side2)
+{
+	struct rename_info *renames;
+	struct tree **merge_trees;
+	struct merge_options_internal *opti = result->priv;
+
+	if (!opti)
+		return;
+
+	renames = &opti->renames;
+	merge_trees = renames->merge_trees;
+
+	/*
+	 * Handle case where previous merge operation did not want cache to
+	 * take effect, e.g. because rename/rename(1to1) makes it invalid.
+	 */
+	if (!merge_trees[0]) {
+		assert(!merge_trees[0] && !merge_trees[1] && !merge_trees[2]);
+		renames->cached_pairs_valid_side = 0; /* neither side valid */
+		return;
+	}
+
+	/*
+	 * Handle other cases; note that merge_trees[0..2] will only
+	 * be NULL if opti is, or if all three were manually set to
+	 * NULL by e.g. rename/rename(1to1) handling.
+	 */
+	assert(merge_trees[0] && merge_trees[1] && merge_trees[2]);
+
+	/* Check if we meet a condition for re-using cached_pairs */
+	if (oideq(&merge_base->object.oid, &merge_trees[2]->object.oid) &&
+	    oideq(&side1->object.oid, &result->tree->object.oid))
+		renames->cached_pairs_valid_side = MERGE_SIDE1;
+	else if (oideq(&merge_base->object.oid, &merge_trees[1]->object.oid) &&
+		 oideq(&side2->object.oid, &result->tree->object.oid))
+		renames->cached_pairs_valid_side = MERGE_SIDE2;
+	else
+		renames->cached_pairs_valid_side = 0; /* neither side valid */
+}
+
 /*** Function Grouping: merge_incore_*() and their internal variants ***/
 
 /*
@@ -3751,6 +4048,7 @@
 	result->clean &= strmap_empty(&opt->priv->conflicted);
 	if (!opt->priv->call_depth) {
 		result->priv = opt->priv;
+		result->_properly_initialized = RESULT_INITIALIZED;
 		opt->priv = NULL;
 	}
 }
@@ -3848,7 +4146,16 @@
 
 	trace2_region_enter("merge", "merge_start", opt->repo);
 	assert(opt->ancestor != NULL);
+	merge_check_renames_reusable(opt, result, merge_base, side1, side2);
 	merge_start(opt, result);
+	/*
+	 * Record the trees used in this merge, so if there's a next merge in
+	 * a cherry-pick or rebase sequence it might be able to take advantage
+	 * of the cached_pairs in that next merge.
+	 */
+	opt->priv->renames.merge_trees[0] = merge_base;
+	opt->priv->renames.merge_trees[1] = side1;
+	opt->priv->renames.merge_trees[2] = side2;
 	trace2_region_leave("merge", "merge_start", opt->repo);
 
 	merge_ort_nonrecursive_internal(opt, merge_base, side1, side2, result);
diff --git a/merge-ort.h b/merge-ort.h
index d53a0a3..c011864 100644
--- a/merge-ort.h
+++ b/merge-ort.h
@@ -29,6 +29,8 @@
 	 * !clean) and to print "CONFLICT" messages.  Not for external use.
 	 */
 	void *priv;
+	/* Also private */
+	unsigned _properly_initialized;
 };
 
 /*
diff --git a/parallel-checkout.c b/parallel-checkout.c
index 6b1af32..ddc0ff3 100644
--- a/parallel-checkout.c
+++ b/parallel-checkout.c
@@ -411,7 +411,7 @@
 	len_data = sizeof(struct pc_item_fixed_portion) + name_len +
 		   working_tree_encoding_len;
 
-	data = xcalloc(1, len_data);
+	data = xmalloc(len_data);
 
 	fixed_portion = (struct pc_item_fixed_portion *)data;
 	fixed_portion->id = pc_item->id;
@@ -421,13 +421,12 @@
 	fixed_portion->name_len = name_len;
 	fixed_portion->working_tree_encoding_len = working_tree_encoding_len;
 	/*
-	 * We use hashcpy() instead of oidcpy() because the hash[] positions
-	 * after `the_hash_algo->rawsz` might not be initialized. And Valgrind
-	 * would complain about passing uninitialized bytes to a syscall
-	 * (write(2)). There is no real harm in this case, but the warning could
-	 * hinder the detection of actual errors.
+	 * We pad the unused bytes in the hash array because, otherwise,
+	 * Valgrind would complain about passing uninitialized bytes to a
+	 * write() syscall. The warning doesn't represent any real risk here,
+	 * but it could hinder the detection of actual errors.
 	 */
-	hashcpy(fixed_portion->oid.hash, pc_item->ce->oid.hash);
+	oidcpy_with_padding(&fixed_portion->oid, &pc_item->ce->oid);
 
 	variant = data + sizeof(*fixed_portion);
 	if (working_tree_encoding_len) {
diff --git a/perl/Git/SVN.pm b/perl/Git/SVN.pm
index f6f1dc0..35ff5a6 100644
--- a/perl/Git/SVN.pm
+++ b/perl/Git/SVN.pm
@@ -1636,7 +1636,7 @@
 	my $commit = shift;
 
 	my @revs = split / /, command_oneline(
-		qw(rev-list --parents -1 -m), $commit);
+		qw(rev-list --parents -1), $commit);
 
 	# Commits with no parents, e.g. the start of a partial branch,
 	# have changes by definition.
diff --git a/read-cache.c b/read-cache.c
index 1b3c2eb..77961a3 100644
--- a/read-cache.c
+++ b/read-cache.c
@@ -26,6 +26,7 @@
 #include "thread-utils.h"
 #include "progress.h"
 #include "sparse-index.h"
+#include "csum-file.h"
 
 /* Mask for the name length in ce_flags in the on-disk index */
 
@@ -2521,80 +2522,23 @@
 	}
 }
 
-#define WRITE_BUFFER_SIZE (128 * 1024)
-static unsigned char write_buffer[WRITE_BUFFER_SIZE];
-static unsigned long write_buffer_len;
-
-static int ce_write_flush(git_hash_ctx *context, int fd)
+static int write_index_ext_header(struct hashfile *f,
+				  git_hash_ctx *eoie_f,
+				  unsigned int ext,
+				  unsigned int sz)
 {
-	unsigned int buffered = write_buffer_len;
-	if (buffered) {
-		the_hash_algo->update_fn(context, write_buffer, buffered);
-		if (write_in_full(fd, write_buffer, buffered) < 0)
-			return -1;
-		write_buffer_len = 0;
+	hashwrite_be32(f, ext);
+	hashwrite_be32(f, sz);
+
+	if (eoie_f) {
+		ext = htonl(ext);
+		sz = htonl(sz);
+		the_hash_algo->update_fn(eoie_f, &ext, sizeof(ext));
+		the_hash_algo->update_fn(eoie_f, &sz, sizeof(sz));
 	}
 	return 0;
 }
 
-static int ce_write(git_hash_ctx *context, int fd, void *data, unsigned int len)
-{
-	while (len) {
-		unsigned int buffered = write_buffer_len;
-		unsigned int partial = WRITE_BUFFER_SIZE - buffered;
-		if (partial > len)
-			partial = len;
-		memcpy(write_buffer + buffered, data, partial);
-		buffered += partial;
-		if (buffered == WRITE_BUFFER_SIZE) {
-			write_buffer_len = buffered;
-			if (ce_write_flush(context, fd))
-				return -1;
-			buffered = 0;
-		}
-		write_buffer_len = buffered;
-		len -= partial;
-		data = (char *) data + partial;
-	}
-	return 0;
-}
-
-static int write_index_ext_header(git_hash_ctx *context, git_hash_ctx *eoie_context,
-				  int fd, unsigned int ext, unsigned int sz)
-{
-	ext = htonl(ext);
-	sz = htonl(sz);
-	if (eoie_context) {
-		the_hash_algo->update_fn(eoie_context, &ext, 4);
-		the_hash_algo->update_fn(eoie_context, &sz, 4);
-	}
-	return ((ce_write(context, fd, &ext, 4) < 0) ||
-		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
-}
-
-static int ce_flush(git_hash_ctx *context, int fd, unsigned char *hash)
-{
-	unsigned int left = write_buffer_len;
-
-	if (left) {
-		write_buffer_len = 0;
-		the_hash_algo->update_fn(context, write_buffer, left);
-	}
-
-	/* Flush first if not enough space for hash signature */
-	if (left + the_hash_algo->rawsz > WRITE_BUFFER_SIZE) {
-		if (write_in_full(fd, write_buffer, left) < 0)
-			return -1;
-		left = 0;
-	}
-
-	/* Append the hash signature at the end */
-	the_hash_algo->final_fn(write_buffer + left, context);
-	hashcpy(hash, write_buffer + left);
-	left += the_hash_algo->rawsz;
-	return (write_in_full(fd, write_buffer, left) < 0) ? -1 : 0;
-}
-
 static void ce_smudge_racily_clean_entry(struct index_state *istate,
 					 struct cache_entry *ce)
 {
@@ -2673,11 +2617,10 @@
 	}
 }
 
-static int ce_write_entry(git_hash_ctx *c, int fd, struct cache_entry *ce,
+static int ce_write_entry(struct hashfile *f, struct cache_entry *ce,
 			  struct strbuf *previous_name, struct ondisk_cache_entry *ondisk)
 {
 	int size;
-	int result;
 	unsigned int saved_namelen;
 	int stripped_name = 0;
 	static unsigned char padding[8] = { 0x00 };
@@ -2693,11 +2636,9 @@
 	if (!previous_name) {
 		int len = ce_namelen(ce);
 		copy_cache_entry_to_ondisk(ondisk, ce);
-		result = ce_write(c, fd, ondisk, size);
-		if (!result)
-			result = ce_write(c, fd, ce->name, len);
-		if (!result)
-			result = ce_write(c, fd, padding, align_padding_size(size, len));
+		hashwrite(f, ondisk, size);
+		hashwrite(f, ce->name, len);
+		hashwrite(f, padding, align_padding_size(size, len));
 	} else {
 		int common, to_remove, prefix_size;
 		unsigned char to_remove_vi[16];
@@ -2711,13 +2652,10 @@
 		prefix_size = encode_varint(to_remove, to_remove_vi);
 
 		copy_cache_entry_to_ondisk(ondisk, ce);
-		result = ce_write(c, fd, ondisk, size);
-		if (!result)
-			result = ce_write(c, fd, to_remove_vi, prefix_size);
-		if (!result)
-			result = ce_write(c, fd, ce->name + common, ce_namelen(ce) - common);
-		if (!result)
-			result = ce_write(c, fd, padding, 1);
+		hashwrite(f, ondisk, size);
+		hashwrite(f, to_remove_vi, prefix_size);
+		hashwrite(f, ce->name + common, ce_namelen(ce) - common);
+		hashwrite(f, padding, 1);
 
 		strbuf_splice(previous_name, common, to_remove,
 			      ce->name + common, ce_namelen(ce) - common);
@@ -2727,7 +2665,7 @@
 		ce->ce_flags &= ~CE_STRIP_NAME;
 	}
 
-	return result;
+	return 0;
 }
 
 /*
@@ -2839,8 +2777,8 @@
 			  int strip_extensions)
 {
 	uint64_t start = getnanotime();
-	int newfd = tempfile->fd;
-	git_hash_ctx c, eoie_c;
+	struct hashfile *f;
+	git_hash_ctx *eoie_c = NULL;
 	struct cache_header hdr;
 	int i, err = 0, removed, extended, hdr_version;
 	struct cache_entry **cache = istate->cache;
@@ -2854,6 +2792,8 @@
 	struct index_entry_offset_table *ieot = NULL;
 	int nr, nr_threads;
 
+	f = hashfd(tempfile->fd, tempfile->filename.buf);
+
 	for (i = removed = extended = 0; i < entries; i++) {
 		if (cache[i]->ce_flags & CE_REMOVE)
 			removed++;
@@ -2882,9 +2822,7 @@
 	hdr.hdr_version = htonl(hdr_version);
 	hdr.hdr_entries = htonl(entries - removed);
 
-	the_hash_algo->init_fn(&c);
-	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
-		return -1;
+	hashwrite(f, &hdr, sizeof(hdr));
 
 	if (!HAVE_THREADS || git_config_get_index_threads(&nr_threads))
 		nr_threads = 1;
@@ -2919,12 +2857,8 @@
 		}
 	}
 
-	offset = lseek(newfd, 0, SEEK_CUR);
-	if (offset < 0) {
-		free(ieot);
-		return -1;
-	}
-	offset += write_buffer_len;
+	offset = hashfile_total(f);
+
 	nr = 0;
 	previous_name = (hdr_version == 4) ? &previous_name_buf : NULL;
 
@@ -2959,14 +2893,10 @@
 			if (previous_name)
 				previous_name->buf[0] = 0;
 			nr = 0;
-			offset = lseek(newfd, 0, SEEK_CUR);
-			if (offset < 0) {
-				free(ieot);
-				return -1;
-			}
-			offset += write_buffer_len;
+
+			offset = hashfile_total(f);
 		}
-		if (ce_write_entry(&c, newfd, ce, previous_name, (struct ondisk_cache_entry *)&ondisk) < 0)
+		if (ce_write_entry(f, ce, previous_name, (struct ondisk_cache_entry *)&ondisk) < 0)
 			err = -1;
 
 		if (err)
@@ -2985,14 +2915,16 @@
 		return err;
 	}
 
-	/* Write extension data here */
-	offset = lseek(newfd, 0, SEEK_CUR);
-	if (offset < 0) {
-		free(ieot);
-		return -1;
+	offset = hashfile_total(f);
+
+	/*
+	 * The extension headers must be hashed on their own for the
+	 * EOIE extension. Create a hashfile here to compute that hash.
+	 */
+	if (offset && record_eoie()) {
+		CALLOC_ARRAY(eoie_c, 1);
+		the_hash_algo->init_fn(eoie_c);
 	}
-	offset += write_buffer_len;
-	the_hash_algo->init_fn(&eoie_c);
 
 	/*
 	 * Lets write out CACHE_EXT_INDEXENTRYOFFSETTABLE first so that we
@@ -3005,8 +2937,8 @@
 		struct strbuf sb = STRBUF_INIT;
 
 		write_ieot_extension(&sb, ieot);
-		err = write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_INDEXENTRYOFFSETTABLE, sb.len) < 0
-			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
+		err = write_index_ext_header(f, eoie_c, CACHE_EXT_INDEXENTRYOFFSETTABLE, sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		free(ieot);
 		if (err)
@@ -3018,9 +2950,9 @@
 		struct strbuf sb = STRBUF_INIT;
 
 		err = write_link_extension(&sb, istate) < 0 ||
-			write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_LINK,
-					       sb.len) < 0 ||
-			ce_write(&c, newfd, sb.buf, sb.len) < 0;
+			write_index_ext_header(f, eoie_c, CACHE_EXT_LINK,
+					       sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		if (err)
 			return -1;
@@ -3029,8 +2961,8 @@
 		struct strbuf sb = STRBUF_INIT;
 
 		cache_tree_write(&sb, istate->cache_tree);
-		err = write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_TREE, sb.len) < 0
-			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
+		err = write_index_ext_header(f, eoie_c, CACHE_EXT_TREE, sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		if (err)
 			return -1;
@@ -3039,9 +2971,9 @@
 		struct strbuf sb = STRBUF_INIT;
 
 		resolve_undo_write(&sb, istate->resolve_undo);
-		err = write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_RESOLVE_UNDO,
-					     sb.len) < 0
-			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
+		err = write_index_ext_header(f, eoie_c, CACHE_EXT_RESOLVE_UNDO,
+					     sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		if (err)
 			return -1;
@@ -3050,9 +2982,9 @@
 		struct strbuf sb = STRBUF_INIT;
 
 		write_untracked_extension(&sb, istate->untracked);
-		err = write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_UNTRACKED,
-					     sb.len) < 0 ||
-			ce_write(&c, newfd, sb.buf, sb.len) < 0;
+		err = write_index_ext_header(f, eoie_c, CACHE_EXT_UNTRACKED,
+					     sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		if (err)
 			return -1;
@@ -3061,14 +2993,14 @@
 		struct strbuf sb = STRBUF_INIT;
 
 		write_fsmonitor_extension(&sb, istate);
-		err = write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_FSMONITOR, sb.len) < 0
-			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
+		err = write_index_ext_header(f, eoie_c, CACHE_EXT_FSMONITOR, sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		if (err)
 			return -1;
 	}
 	if (istate->sparse_index) {
-		if (write_index_ext_header(&c, &eoie_c, newfd, CACHE_EXT_SPARSE_DIRECTORIES, 0) < 0)
+		if (write_index_ext_header(f, eoie_c, CACHE_EXT_SPARSE_DIRECTORIES, 0) < 0)
 			return -1;
 	}
 
@@ -3078,19 +3010,18 @@
 	 * read.  Write it out regardless of the strip_extensions parameter as we need it
 	 * when loading the shared index.
 	 */
-	if (offset && record_eoie()) {
+	if (eoie_c) {
 		struct strbuf sb = STRBUF_INIT;
 
-		write_eoie_extension(&sb, &eoie_c, offset);
-		err = write_index_ext_header(&c, NULL, newfd, CACHE_EXT_ENDOFINDEXENTRIES, sb.len) < 0
-			|| ce_write(&c, newfd, sb.buf, sb.len) < 0;
+		write_eoie_extension(&sb, eoie_c, offset);
+		err = write_index_ext_header(f, NULL, CACHE_EXT_ENDOFINDEXENTRIES, sb.len) < 0;
+		hashwrite(f, sb.buf, sb.len);
 		strbuf_release(&sb);
 		if (err)
 			return -1;
 	}
 
-	if (ce_flush(&c, newfd, istate->oid.hash))
-		return -1;
+	finalize_hashfile(f, istate->oid.hash, CSUM_HASH_IN_STREAM);
 	if (close_tempfile_gently(tempfile)) {
 		error(_("could not close '%s'"), get_tempfile_path(tempfile));
 		return -1;
diff --git a/ref-filter.c b/ref-filter.c
index 97116e1..4db0e40 100644
--- a/ref-filter.c
+++ b/ref-filter.c
@@ -109,6 +109,56 @@
 } ref_to_worktree_map;
 
 /*
+ * The enum atom_type is used as the index of valid_atom array.
+ * In the atom parsing stage, it will be passed to used_atom.atom_type
+ * as the identifier of the atom type. We can check the type of used_atom
+ * entry by `if (used_atom[i].atom_type == ATOM_*)`.
+ */
+enum atom_type {
+	ATOM_REFNAME,
+	ATOM_OBJECTTYPE,
+	ATOM_OBJECTSIZE,
+	ATOM_OBJECTNAME,
+	ATOM_DELTABASE,
+	ATOM_TREE,
+	ATOM_PARENT,
+	ATOM_NUMPARENT,
+	ATOM_OBJECT,
+	ATOM_TYPE,
+	ATOM_TAG,
+	ATOM_AUTHOR,
+	ATOM_AUTHORNAME,
+	ATOM_AUTHOREMAIL,
+	ATOM_AUTHORDATE,
+	ATOM_COMMITTER,
+	ATOM_COMMITTERNAME,
+	ATOM_COMMITTEREMAIL,
+	ATOM_COMMITTERDATE,
+	ATOM_TAGGER,
+	ATOM_TAGGERNAME,
+	ATOM_TAGGEREMAIL,
+	ATOM_TAGGERDATE,
+	ATOM_CREATOR,
+	ATOM_CREATORDATE,
+	ATOM_SUBJECT,
+	ATOM_BODY,
+	ATOM_TRAILERS,
+	ATOM_CONTENTS,
+	ATOM_UPSTREAM,
+	ATOM_PUSH,
+	ATOM_SYMREF,
+	ATOM_FLAG,
+	ATOM_HEAD,
+	ATOM_COLOR,
+	ATOM_WORKTREEPATH,
+	ATOM_ALIGN,
+	ATOM_END,
+	ATOM_IF,
+	ATOM_THEN,
+	ATOM_ELSE,
+};
+
+/*
  * An atom is a valid field atom listed below, possibly prefixed with
  * a "*" to denote deref_tag().
  *
@@ -119,6 +169,7 @@
  * array.
  */
 static struct used_atom {
+	enum atom_type atom_type;
 	const char *name;
 	cmp_type type;
 	info_source source;
@@ -146,6 +197,9 @@
 			enum { O_FULL, O_LENGTH, O_SHORT } option;
 			unsigned int length;
 		} oid;
+		struct {
+			enum { O_SIZE, O_SIZE_DISK } option;
+		} objectsize;
 		struct email_option {
 			enum { EO_RAW, EO_TRIM, EO_LOCALPART } option;
 		} email_option;
@@ -269,11 +323,13 @@
 				  const char *arg, struct strbuf *err)
 {
 	if (!arg) {
+		atom->u.objectsize.option = O_SIZE;
 		if (*atom->name == '*')
 			oi_deref.info.sizep = &oi_deref.size;
 		else
 			oi.info.sizep = &oi.size;
 	} else if (!strcmp(arg, "disk")) {
+		atom->u.objectsize.option = O_SIZE_DISK;
 		if (*atom->name == '*')
 			oi_deref.info.disk_sizep = &oi_deref.disk_size;
 		else
@@ -501,47 +557,47 @@
 	int (*parser)(const struct ref_format *format, struct used_atom *atom,
 		      const char *arg, struct strbuf *err);
 } valid_atom[] = {
-	{ "refname", SOURCE_NONE, FIELD_STR, refname_atom_parser },
-	{ "objecttype", SOURCE_OTHER, FIELD_STR, objecttype_atom_parser },
-	{ "objectsize", SOURCE_OTHER, FIELD_ULONG, objectsize_atom_parser },
-	{ "objectname", SOURCE_OTHER, FIELD_STR, oid_atom_parser },
-	{ "deltabase", SOURCE_OTHER, FIELD_STR, deltabase_atom_parser },
-	{ "tree", SOURCE_OBJ, FIELD_STR, oid_atom_parser },
-	{ "parent", SOURCE_OBJ, FIELD_STR, oid_atom_parser },
-	{ "numparent", SOURCE_OBJ, FIELD_ULONG },
-	{ "object", SOURCE_OBJ },
-	{ "type", SOURCE_OBJ },
-	{ "tag", SOURCE_OBJ },
-	{ "author", SOURCE_OBJ },
-	{ "authorname", SOURCE_OBJ },
-	{ "authoremail", SOURCE_OBJ, FIELD_STR, person_email_atom_parser },
-	{ "authordate", SOURCE_OBJ, FIELD_TIME },
-	{ "committer", SOURCE_OBJ },
-	{ "committername", SOURCE_OBJ },
-	{ "committeremail", SOURCE_OBJ, FIELD_STR, person_email_atom_parser },
-	{ "committerdate", SOURCE_OBJ, FIELD_TIME },
-	{ "tagger", SOURCE_OBJ },
-	{ "taggername", SOURCE_OBJ },
-	{ "taggeremail", SOURCE_OBJ, FIELD_STR, person_email_atom_parser },
-	{ "taggerdate", SOURCE_OBJ, FIELD_TIME },
-	{ "creator", SOURCE_OBJ },
-	{ "creatordate", SOURCE_OBJ, FIELD_TIME },
-	{ "subject", SOURCE_OBJ, FIELD_STR, subject_atom_parser },
-	{ "body", SOURCE_OBJ, FIELD_STR, body_atom_parser },
-	{ "trailers", SOURCE_OBJ, FIELD_STR, trailers_atom_parser },
-	{ "contents", SOURCE_OBJ, FIELD_STR, contents_atom_parser },
-	{ "upstream", SOURCE_NONE, FIELD_STR, remote_ref_atom_parser },
-	{ "push", SOURCE_NONE, FIELD_STR, remote_ref_atom_parser },
-	{ "symref", SOURCE_NONE, FIELD_STR, refname_atom_parser },
-	{ "flag", SOURCE_NONE },
-	{ "HEAD", SOURCE_NONE, FIELD_STR, head_atom_parser },
-	{ "color", SOURCE_NONE, FIELD_STR, color_atom_parser },
-	{ "worktreepath", SOURCE_NONE },
-	{ "align", SOURCE_NONE, FIELD_STR, align_atom_parser },
-	{ "end", SOURCE_NONE },
-	{ "if", SOURCE_NONE, FIELD_STR, if_atom_parser },
-	{ "then", SOURCE_NONE },
-	{ "else", SOURCE_NONE },
+	[ATOM_REFNAME] = { "refname", SOURCE_NONE, FIELD_STR, refname_atom_parser },
+	[ATOM_OBJECTTYPE] = { "objecttype", SOURCE_OTHER, FIELD_STR, objecttype_atom_parser },
+	[ATOM_OBJECTSIZE] = { "objectsize", SOURCE_OTHER, FIELD_ULONG, objectsize_atom_parser },
+	[ATOM_OBJECTNAME] = { "objectname", SOURCE_OTHER, FIELD_STR, oid_atom_parser },
+	[ATOM_DELTABASE] = { "deltabase", SOURCE_OTHER, FIELD_STR, deltabase_atom_parser },
+	[ATOM_TREE] = { "tree", SOURCE_OBJ, FIELD_STR, oid_atom_parser },
+	[ATOM_PARENT] = { "parent", SOURCE_OBJ, FIELD_STR, oid_atom_parser },
+	[ATOM_NUMPARENT] = { "numparent", SOURCE_OBJ, FIELD_ULONG },
+	[ATOM_OBJECT] = { "object", SOURCE_OBJ },
+	[ATOM_TYPE] = { "type", SOURCE_OBJ },
+	[ATOM_TAG] = { "tag", SOURCE_OBJ },
+	[ATOM_AUTHOR] = { "author", SOURCE_OBJ },
+	[ATOM_AUTHORNAME] = { "authorname", SOURCE_OBJ },
+	[ATOM_AUTHOREMAIL] = { "authoremail", SOURCE_OBJ, FIELD_STR, person_email_atom_parser },
+	[ATOM_AUTHORDATE] = { "authordate", SOURCE_OBJ, FIELD_TIME },
+	[ATOM_COMMITTER] = { "committer", SOURCE_OBJ },
+	[ATOM_COMMITTERNAME] = { "committername", SOURCE_OBJ },
+	[ATOM_COMMITTEREMAIL] = { "committeremail", SOURCE_OBJ, FIELD_STR, person_email_atom_parser },
+	[ATOM_COMMITTERDATE] = { "committerdate", SOURCE_OBJ, FIELD_TIME },
+	[ATOM_TAGGER] = { "tagger", SOURCE_OBJ },
+	[ATOM_TAGGERNAME] = { "taggername", SOURCE_OBJ },
+	[ATOM_TAGGEREMAIL] = { "taggeremail", SOURCE_OBJ, FIELD_STR, person_email_atom_parser },
+	[ATOM_TAGGERDATE] = { "taggerdate", SOURCE_OBJ, FIELD_TIME },
+	[ATOM_CREATOR] = { "creator", SOURCE_OBJ },
+	[ATOM_CREATORDATE] = { "creatordate", SOURCE_OBJ, FIELD_TIME },
+	[ATOM_SUBJECT] = { "subject", SOURCE_OBJ, FIELD_STR, subject_atom_parser },
+	[ATOM_BODY] = { "body", SOURCE_OBJ, FIELD_STR, body_atom_parser },
+	[ATOM_TRAILERS] = { "trailers", SOURCE_OBJ, FIELD_STR, trailers_atom_parser },
+	[ATOM_CONTENTS] = { "contents", SOURCE_OBJ, FIELD_STR, contents_atom_parser },
+	[ATOM_UPSTREAM] = { "upstream", SOURCE_NONE, FIELD_STR, remote_ref_atom_parser },
+	[ATOM_PUSH] = { "push", SOURCE_NONE, FIELD_STR, remote_ref_atom_parser },
+	[ATOM_SYMREF] = { "symref", SOURCE_NONE, FIELD_STR, refname_atom_parser },
+	[ATOM_FLAG] = { "flag", SOURCE_NONE },
+	[ATOM_HEAD] = { "HEAD", SOURCE_NONE, FIELD_STR, head_atom_parser },
+	[ATOM_COLOR] = { "color", SOURCE_NONE, FIELD_STR, color_atom_parser },
+	[ATOM_WORKTREEPATH] = { "worktreepath", SOURCE_NONE },
+	[ATOM_ALIGN] = { "align", SOURCE_NONE, FIELD_STR, align_atom_parser },
+	[ATOM_END] = { "end", SOURCE_NONE },
+	[ATOM_IF] = { "if", SOURCE_NONE, FIELD_STR, if_atom_parser },
+	[ATOM_THEN] = { "then", SOURCE_NONE },
+	[ATOM_ELSE] = { "else", SOURCE_NONE },
 	/*
 	 * Please update $__git_ref_fieldlist in git-completion.bash
 	 * when you add new atoms
@@ -623,6 +679,7 @@
 	at = used_atom_cnt;
 	used_atom_cnt++;
 	REALLOC_ARRAY(used_atom, used_atom_cnt);
+	used_atom[at].atom_type = i;
 	used_atom[at].name = xmemdupz(atom, ep - atom);
 	used_atom[at].type = valid_atom[i].cmp_type;
 	used_atom[at].source = valid_atom[i].source;
@@ -647,7 +704,7 @@
 		return -1;
 	if (*atom == '*')
 		need_tagged = 1;
-	if (!strcmp(valid_atom[i].name, "symref"))
+	if (i == ATOM_SYMREF)
 		need_symref = 1;
 	return at;
 }
@@ -960,22 +1017,25 @@
 
 	for (i = 0; i < used_atom_cnt; i++) {
 		const char *name = used_atom[i].name;
+		enum atom_type atom_type = used_atom[i].atom_type;
 		struct atom_value *v = &val[i];
 		if (!!deref != (*name == '*'))
 			continue;
 		if (deref)
 			name++;
-		if (!strcmp(name, "objecttype"))
+		if (atom_type == ATOM_OBJECTTYPE)
 			v->s = xstrdup(type_name(oi->type));
-		else if (!strcmp(name, "objectsize:disk")) {
-			v->value = oi->disk_size;
-			v->s = xstrfmt("%"PRIuMAX, (uintmax_t)oi->disk_size);
-		} else if (!strcmp(name, "objectsize")) {
-			v->value = oi->size;
-			v->s = xstrfmt("%"PRIuMAX , (uintmax_t)oi->size);
-		} else if (!strcmp(name, "deltabase"))
+		else if (atom_type == ATOM_OBJECTSIZE) {
+			if (used_atom[i].u.objectsize.option == O_SIZE_DISK) {
+				v->value = oi->disk_size;
+				v->s = xstrfmt("%"PRIuMAX, (uintmax_t)oi->disk_size);
+			} else if (used_atom[i].u.objectsize.option == O_SIZE) {
+				v->value = oi->size;
+				v->s = xstrfmt("%"PRIuMAX , (uintmax_t)oi->size);
+			}
+		} else if (atom_type == ATOM_DELTABASE)
 			v->s = xstrdup(oid_to_hex(&oi->delta_base_oid));
-		else if (deref)
+		else if (atom_type == ATOM_OBJECTNAME && deref)
 			grab_oid(name, "objectname", &oi->oid, v, &used_atom[i]);
 	}
 }
@@ -988,16 +1048,17 @@
 
 	for (i = 0; i < used_atom_cnt; i++) {
 		const char *name = used_atom[i].name;
+		enum atom_type atom_type = used_atom[i].atom_type;
 		struct atom_value *v = &val[i];
 		if (!!deref != (*name == '*'))
 			continue;
 		if (deref)
 			name++;
-		if (!strcmp(name, "tag"))
+		if (atom_type == ATOM_TAG)
 			v->s = xstrdup(tag->tag);
-		else if (!strcmp(name, "type") && tag->tagged)
+		else if (atom_type == ATOM_TYPE && tag->tagged)
 			v->s = xstrdup(type_name(tag->tagged->type));
-		else if (!strcmp(name, "object") && tag->tagged)
+		else if (atom_type == ATOM_OBJECT && tag->tagged)
 			v->s = xstrdup(oid_to_hex(&tag->tagged->oid));
 	}
 }
@@ -1010,18 +1071,20 @@
 
 	for (i = 0; i < used_atom_cnt; i++) {
 		const char *name = used_atom[i].name;
+		enum atom_type atom_type = used_atom[i].atom_type;
 		struct atom_value *v = &val[i];
 		if (!!deref != (*name == '*'))
 			continue;
 		if (deref)
 			name++;
-		if (grab_oid(name, "tree", get_commit_tree_oid(commit), v, &used_atom[i]))
+		if (atom_type == ATOM_TREE &&
+		    grab_oid(name, "tree", get_commit_tree_oid(commit), v, &used_atom[i]))
 			continue;
-		if (!strcmp(name, "numparent")) {
+		if (atom_type == ATOM_NUMPARENT) {
 			v->value = commit_list_count(commit->parents);
 			v->s = xstrfmt("%lu", (unsigned long)v->value);
 		}
-		else if (starts_with(name, "parent")) {
+		else if (atom_type == ATOM_PARENT) {
 			struct commit_list *parents;
 			struct strbuf s = STRBUF_INIT;
 			for (parents = commit->parents; parents; parents = parents->next) {
@@ -1201,15 +1264,16 @@
 		return;
 	for (i = 0; i < used_atom_cnt; i++) {
 		const char *name = used_atom[i].name;
+		enum atom_type atom_type = used_atom[i].atom_type;
 		struct atom_value *v = &val[i];
 		if (!!deref != (*name == '*'))
 			continue;
 		if (deref)
 			name++;
 
-		if (starts_with(name, "creatordate"))
+		if (atom_type == ATOM_CREATORDATE)
 			grab_date(wholine, v, name);
-		else if (!strcmp(name, "creator"))
+		else if (atom_type == ATOM_CREATOR)
 			v->s = copy_line(wholine);
 	}
 }
@@ -1689,6 +1753,7 @@
 	/* Fill in specials first */
 	for (i = 0; i < used_atom_cnt; i++) {
 		struct used_atom *atom = &used_atom[i];
+		enum atom_type atom_type = atom->atom_type;
 		const char *name = used_atom[i].name;
 		struct atom_value *v = &ref->value[i];
 		int deref = 0;
@@ -1703,18 +1768,18 @@
 			name++;
 		}
 
-		if (starts_with(name, "refname"))
+		if (atom_type == ATOM_REFNAME)
 			refname = get_refname(atom, ref);
-		else if (!strcmp(name, "worktreepath")) {
+		else if (atom_type == ATOM_WORKTREEPATH) {
 			if (ref->kind == FILTER_REFS_BRANCHES)
 				v->s = get_worktree_path(atom, ref);
 			else
 				v->s = xstrdup("");
 			continue;
 		}
-		else if (starts_with(name, "symref"))
+		else if (atom_type == ATOM_SYMREF)
 			refname = get_symref(atom, ref);
-		else if (starts_with(name, "upstream")) {
+		else if (atom_type == ATOM_UPSTREAM) {
 			const char *branch_name;
 			/* only local branches may have an upstream */
 			if (!skip_prefix(ref->refname, "refs/heads/",
@@ -1730,7 +1795,7 @@
 			else
 				v->s = xstrdup("");
 			continue;
-		} else if (!strcmp(atom->name, "push") || starts_with(atom->name, "push:")) {
+		} else if (atom_type == ATOM_PUSH && atom->u.remote_ref.push) {
 			const char *branch_name;
 			v->s = xstrdup("");
 			if (!skip_prefix(ref->refname, "refs/heads/",
@@ -1749,10 +1814,10 @@
 			free((char *)v->s);
 			fill_remote_ref_details(atom, refname, branch, &v->s);
 			continue;
-		} else if (starts_with(name, "color:")) {
+		} else if (atom_type == ATOM_COLOR) {
 			v->s = xstrdup(atom->u.color);
 			continue;
-		} else if (!strcmp(name, "flag")) {
+		} else if (atom_type == ATOM_FLAG) {
 			char buf[256], *cp = buf;
 			if (ref->flag & REF_ISSYMREF)
 				cp = copy_advance(cp, ",symref");
@@ -1765,23 +1830,24 @@
 				v->s = xstrdup(buf + 1);
 			}
 			continue;
-		} else if (!deref && grab_oid(name, "objectname", &ref->objectname, v, atom)) {
-			continue;
-		} else if (!strcmp(name, "HEAD")) {
+		} else if (!deref && atom_type == ATOM_OBJECTNAME &&
+			   grab_oid(name, "objectname", &ref->objectname, v, atom)) {
+				continue;
+		} else if (atom_type == ATOM_HEAD) {
 			if (atom->u.head && !strcmp(ref->refname, atom->u.head))
 				v->s = xstrdup("*");
 			else
 				v->s = xstrdup(" ");
 			continue;
-		} else if (starts_with(name, "align")) {
+		} else if (atom_type == ATOM_ALIGN) {
 			v->handler = align_atom_handler;
 			v->s = xstrdup("");
 			continue;
-		} else if (!strcmp(name, "end")) {
+		} else if (atom_type == ATOM_END) {
 			v->handler = end_atom_handler;
 			v->s = xstrdup("");
 			continue;
-		} else if (starts_with(name, "if")) {
+		} else if (atom_type == ATOM_IF) {
 			const char *s;
 			if (skip_prefix(name, "if:", &s))
 				v->s = xstrdup(s);
@@ -1789,11 +1855,11 @@
 				v->s = xstrdup("");
 			v->handler = if_atom_handler;
 			continue;
-		} else if (!strcmp(name, "then")) {
+		} else if (atom_type == ATOM_THEN) {
 			v->handler = then_atom_handler;
 			v->s = xstrdup("");
 			continue;
-		} else if (!strcmp(name, "else")) {
+		} else if (atom_type == ATOM_ELSE) {
 			v->handler = else_atom_handler;
 			v->s = xstrdup("");
 			continue;
diff --git a/revision.h b/revision.h
index 93aa012..17698cb 100644
--- a/revision.h
+++ b/revision.h
@@ -193,10 +193,10 @@
 			/* Diff-merge flags */
 			explicit_diff_merges: 1,
 			merges_need_diff: 1,
+			merges_imply_patch:1,
 			separate_merges: 1,
 			combine_merges:1,
 			combined_all_paths:1,
-			combined_imply_patch:1,
 			dense_combined_merges:1,
 			first_parent_merges:1;
 
diff --git a/setup.c b/setup.c
index 59e2fac..ead2f80 100644
--- a/setup.c
+++ b/setup.c
@@ -666,7 +666,9 @@
 	if (format->version >= 1 && format->unknown_extensions.nr) {
 		int i;
 
-		strbuf_addstr(err, _("unknown repository extensions found:"));
+		strbuf_addstr(err, Q_("unknown repository extension found:",
+				      "unknown repository extensions found:",
+				      format->unknown_extensions.nr));
 
 		for (i = 0; i < format->unknown_extensions.nr; i++)
 			strbuf_addf(err, "\n\t%s",
@@ -678,7 +680,9 @@
 		int i;
 
 		strbuf_addstr(err,
-			      _("repo version is 0, but v1-only extensions found:"));
+			      Q_("repo version is 0, but v1-only extension found:",
+				 "repo version is 0, but v1-only extensions found:",
+				 format->v1_only_extensions.nr));
 
 		for (i = 0; i < format->v1_only_extensions.nr; i++)
 			strbuf_addf(err, "\n\t%s",
diff --git a/t/helper/test-fast-rebase.c b/t/helper/test-fast-rebase.c
index 3732122..fc2d460 100644
--- a/t/helper/test-fast-rebase.c
+++ b/t/helper/test-fast-rebase.c
@@ -91,7 +91,6 @@
 	struct commit *last_commit = NULL, *last_picked_commit = NULL;
 	struct object_id head;
 	struct lock_file lock = LOCK_INIT;
-	int clean = 1;
 	struct strvec rev_walk_args = STRVEC_INIT;
 	struct rev_info revs;
 	struct commit *commit;
@@ -124,7 +123,8 @@
 	assert(oideq(&onto->object.oid, &head));
 
 	hold_locked_index(&lock, LOCK_DIE_ON_ERROR);
-	assert(repo_read_index(the_repository) >= 0);
+	if (repo_read_index(the_repository) < 0)
+		BUG("Could not read index");
 
 	repo_init_revisions(the_repository, &revs, NULL);
 	revs.verbose_header = 1;
@@ -175,11 +175,10 @@
 		free((char*)merge_opt.ancestor);
 		merge_opt.ancestor = NULL;
 		if (!result.clean)
-			die("Aborting: Hit a conflict and restarting is not implemented.");
+			break;
 		last_picked_commit = commit;
 		last_commit = create_commit(result.tree, commit, last_commit);
 	}
-	fprintf(stderr, "\nDone.\n");
 	/* TODO: There should be some kind of rev_info_free(&revs) call... */
 	memset(&revs, 0, sizeof(revs));
 
@@ -188,24 +187,39 @@
 	if (result.clean < 0)
 		exit(128);
 
-	strbuf_addf(&reflog_msg, "finish rebase %s onto %s",
-		    oid_to_hex(&last_picked_commit->object.oid),
-		    oid_to_hex(&last_commit->object.oid));
-	if (update_ref(reflog_msg.buf, branch_name.buf,
-		       &last_commit->object.oid,
-		       &last_picked_commit->object.oid,
-		       REF_NO_DEREF, UPDATE_REFS_MSG_ON_ERR)) {
-		error(_("could not update %s"), argv[4]);
-		die("Failed to update %s", argv[4]);
-	}
-	if (create_symref("HEAD", branch_name.buf, reflog_msg.buf) < 0)
-		die(_("unable to update HEAD"));
-	strbuf_release(&reflog_msg);
-	strbuf_release(&branch_name);
+	if (result.clean) {
+		fprintf(stderr, "\nDone.\n");
+		strbuf_addf(&reflog_msg, "finish rebase %s onto %s",
+			    oid_to_hex(&last_picked_commit->object.oid),
+			    oid_to_hex(&last_commit->object.oid));
+		if (update_ref(reflog_msg.buf, branch_name.buf,
+			       &last_commit->object.oid,
+			       &last_picked_commit->object.oid,
+			       REF_NO_DEREF, UPDATE_REFS_MSG_ON_ERR)) {
+			error(_("could not update %s"), argv[4]);
+			die("Failed to update %s", argv[4]);
+		}
+		if (create_symref("HEAD", branch_name.buf, reflog_msg.buf) < 0)
+			die(_("unable to update HEAD"));
+		strbuf_release(&reflog_msg);
+		strbuf_release(&branch_name);
 
-	prime_cache_tree(the_repository, the_repository->index, result.tree);
+		prime_cache_tree(the_repository, the_repository->index,
+				 result.tree);
+	} else {
+		fprintf(stderr, "\nAborting: Hit a conflict.\n");
+		strbuf_addf(&reflog_msg, "rebase progress up to %s",
+			    oid_to_hex(&last_picked_commit->object.oid));
+		if (update_ref(reflog_msg.buf, "HEAD",
+			       &last_commit->object.oid,
+			       &head,
+			       REF_NO_DEREF, UPDATE_REFS_MSG_ON_ERR)) {
+			error(_("could not update %s"), argv[4]);
+			die("Failed to update %s", argv[4]);
+		}
+	}
 	if (write_locked_index(&the_index, &lock,
 			       COMMIT_LOCK | SKIP_IF_UNCHANGED))
 		die(_("unable to write %s"), get_index_file());
-	return (clean == 0);
+	return (result.clean == 0);
 }
diff --git a/t/lib-git-svn.sh b/t/lib-git-svn.sh
index 547eb3c..2fde235 100644
--- a/t/lib-git-svn.sh
+++ b/t/lib-git-svn.sh
@@ -121,12 +121,22 @@
 		 --listen-host 127.0.0.1 &
 }
 
-prepare_a_utf8_locale () {
-	a_utf8_locale=$(locale -a | sed -n '/\.[uU][tT][fF]-*8$/{
-	p
-	q
-}')
-	if test -n "$a_utf8_locale"
+prepare_utf8_locale () {
+	if test -z "$GIT_TEST_UTF8_LOCALE"
+	then
+		case "${LC_ALL:-$LANG}" in
+		*.[Uu][Tt][Ff]8 | *.[Uu][Tt][Ff]-8)
+			GIT_TEST_UTF8_LOCALE="${LC_ALL:-$LANG}"
+			;;
+		*)
+			GIT_TEST_UTF8_LOCALE=$(locale -a | sed -n '/\.[uU][tT][fF]-*8$/{
+				p
+				q
+			}')
+			;;
+		esac
+	fi
+	if test -n "$GIT_TEST_UTF8_LOCALE"
 	then
 		test_set_prereq UTF8
 	else
diff --git a/t/lib-submodule-update.sh b/t/lib-submodule-update.sh
index 4b714e9..f7c7df0 100644
--- a/t/lib-submodule-update.sh
+++ b/t/lib-submodule-update.sh
@@ -63,6 +63,7 @@
 	git init submodule_update_repo &&
 	(
 		cd submodule_update_repo &&
+		branch=$(git symbolic-ref --short HEAD) &&
 		echo "expect" >>.gitignore &&
 		echo "actual" >>.gitignore &&
 		echo "x" >file1 &&
@@ -144,7 +145,7 @@
 		git checkout -b valid_sub1 &&
 		git revert HEAD &&
 
-		git checkout "${GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME-master}"
+		git checkout "$branch"
 	)
 }
 
diff --git a/t/t0000-basic.sh b/t/t0000-basic.sh
index 705d62c..2c6e34b 100755
--- a/t/t0000-basic.sh
+++ b/t/t0000-basic.sh
@@ -84,10 +84,6 @@
 		passing metrics
 		'
 
-		# Tell the framework that we are self-testing to make sure
-		# it yields a stable result.
-		GIT_TEST_FRAMEWORK_SELFTEST=t &&
-
 		# Point to the t/test-lib.sh, which isn't in ../ as usual
 		. "\$TEST_DIRECTORY"/test-lib.sh
 		EOF
diff --git a/t/t1307-config-blob.sh b/t/t1307-config-blob.sh
index 002e6d3..930dce0 100755
--- a/t/t1307-config-blob.sh
+++ b/t/t1307-config-blob.sh
@@ -65,9 +65,7 @@
 '
 
 test_expect_success 'can parse blob ending with CR' '
-	printf "[some]key = value\\r" >config &&
-	git add config &&
-	git commit -m CR &&
+	test_commit --printf CR config "[some]key = value\\r" &&
 	echo value >expect &&
 	git config --blob=HEAD:config some.key >actual &&
 	test_cmp expect actual
diff --git a/t/t1403-show-ref.sh b/t/t1403-show-ref.sh
index 6ce62f8..17d3cc1 100755
--- a/t/t1403-show-ref.sh
+++ b/t/t1403-show-ref.sh
@@ -7,11 +7,9 @@
 . ./test-lib.sh
 
 test_expect_success setup '
-	test_commit A &&
-	git tag -f -a -m "annotated A" A &&
+	test_commit --annotate A &&
 	git checkout -b side &&
-	test_commit B &&
-	git tag -f -a -m "annotated B" B &&
+	test_commit --annotate B &&
 	git checkout main &&
 	test_commit C &&
 	git branch B A^0
diff --git a/t/t2030-unresolve-info.sh b/t/t2030-unresolve-info.sh
index be6c84c..f691e6d 100755
--- a/t/t2030-unresolve-info.sh
+++ b/t/t2030-unresolve-info.sh
@@ -179,8 +179,7 @@
 
 test_expect_success 'rerere forget (binary)' '
 	git checkout -f side &&
-	printf "a\0c" >binary &&
-	git commit -a -m binary &&
+	test_commit --printf binary binary "a\0c" &&
 	test_must_fail git merge second &&
 	git rerere forget binary
 '
diff --git a/t/t3513-revert-submodule.sh b/t/t3513-revert-submodule.sh
index 74cd96e..8bfe3ed 100755
--- a/t/t3513-revert-submodule.sh
+++ b/t/t3513-revert-submodule.sh
@@ -14,7 +14,7 @@
 git_revert () {
 	git status -su >expect &&
 	ls -1pR * >>expect &&
-	tar cf "$TRASH_DIRECTORY/tmp.tar" * &&
+	"$TAR" cf "$TRASH_DIRECTORY/tmp.tar" * &&
 	may_only_be_test_must_fail "$2" &&
 	$2 git checkout "$1" &&
 	if test -n "$2"
@@ -23,7 +23,7 @@
 	fi &&
 	git revert HEAD &&
 	rm -rf * &&
-	tar xf "$TRASH_DIRECTORY/tmp.tar" &&
+	"$TAR" xf "$TRASH_DIRECTORY/tmp.tar" &&
 	git status -su >actual &&
 	ls -1pR * >>actual &&
 	test_cmp expect actual &&
diff --git a/t/t3903-stash.sh b/t/t3903-stash.sh
index 5f282ec..873aa56 100755
--- a/t/t3903-stash.sh
+++ b/t/t3903-stash.sh
@@ -859,7 +859,7 @@
 	git stash
 '
 
-test_expect_success 'stash list implies --first-parent -m' '
+test_expect_success 'stash list -p shows simple diff' '
 	cat >expect <<-EOF &&
 	stash@{0}
 
diff --git a/t/t4006-diff-mode.sh b/t/t4006-diff-mode.sh
index 275ce5f..6cdee2a 100755
--- a/t/t4006-diff-mode.sh
+++ b/t/t4006-diff-mode.sh
@@ -26,10 +26,8 @@
 '
 
 test_expect_success 'prepare binary file' '
-	git commit -m rezrov &&
-	printf "\00\01\02\03\04\05\06" >binbin &&
-	git add binbin &&
-	git commit -m binbin
+	git commit -m one &&
+	test_commit --printf two binbin "\00\01\02\03\04\05\06"
 '
 
 test_expect_success '--stat output after text chmod' '
diff --git a/t/t4013-diff-various.sh b/t/t4013-diff-various.sh
index 87def81..7fadc98 100755
--- a/t/t4013-diff-various.sh
+++ b/t/t4013-diff-various.sh
@@ -293,6 +293,7 @@
 diff-tree --summary initial mode
 
 diff-tree master
+diff-tree -m master
 diff-tree -p master
 diff-tree -p -m master
 diff-tree -c master
@@ -337,6 +338,8 @@
 log -m -p master
 log --cc -m -p master
 log -c -m -p master
+log -m --raw master
+log -m --stat master
 log -SF master
 log -S F master
 log -SF -p master
@@ -452,6 +455,14 @@
 diff-tree -R --stat --compact-summary initial mode
 EOF
 
+test_expect_success 'log -m matches log -m -p' '
+	git log -m -p master >result &&
+	process_diffs result >expected &&
+	git log -m >result &&
+	process_diffs result >actual &&
+	test_cmp expected actual
+'
+
 test_expect_success 'log --diff-merges=on matches --diff-merges=separate' '
 	git log -p --diff-merges=separate master >result &&
 	process_diffs result >expected &&
@@ -483,6 +494,19 @@
 	test_cmp expected actual
 '
 
+# -m in "git diff-index" means "match missing", that differs
+# from its meaning in "git diff". Let's check it in diff-index.
+# The line in the output for removed file should disappear when
+# we provide -m in diff-index.
+test_expect_success 'git diff-index -m' '
+	rm -f file1 &&
+	git diff-index HEAD >without-m &&
+	lines_count=$(wc -l <without-m) &&
+	git diff-index -m HEAD >with-m &&
+	git restore file1 &&
+	test_line_count = $((lines_count - 1)) with-m
+'
+
 test_expect_success 'log -S requires an argument' '
 	test_must_fail git log -S
 '
diff --git a/t/t4013/diff.diff-tree_-m_master b/t/t4013/diff.diff-tree_-m_master
new file mode 100644
index 0000000..6d0a220
--- /dev/null
+++ b/t/t4013/diff.diff-tree_-m_master
@@ -0,0 +1,11 @@
+$ git diff-tree -m master
+59d314ad6f356dd08601a4cd5e530381da3e3c64
+:040000 040000 65f5c9dd60ce3b2b3324b618ac7accf8d912c113 0564e026437809817a64fff393079714b6dd4628 M	dir
+:100644 100644 b414108e81e5091fe0974a1858b4d0d22b107f70 10a8a9f3657f91a156b9f0184ed79a20adef9f7f M	file0
+59d314ad6f356dd08601a4cd5e530381da3e3c64
+:040000 040000 f977ed46ae6873c1c30ab878e15a4accedc3618b 0564e026437809817a64fff393079714b6dd4628 M	dir
+:100644 100644 f4615da674c09df322d6ba8d6b21ecfb1b1ba510 10a8a9f3657f91a156b9f0184ed79a20adef9f7f M	file0
+:000000 100644 0000000000000000000000000000000000000000 b1e67221afe8461efd244b487afca22d46b95eb8 A	file1
+:100644 000000 01e79c32a8c99c557f0757da7cb6d65b3414466d 0000000000000000000000000000000000000000 D	file2
+:100644 000000 7289e35bff32727c08dda207511bec138fdb9ea5 0000000000000000000000000000000000000000 D	file3
+$
diff --git a/t/t4013/diff.log_-m_--raw_master b/t/t4013/diff.log_-m_--raw_master
new file mode 100644
index 0000000..cd2ecc4
--- /dev/null
+++ b/t/t4013/diff.log_-m_--raw_master
@@ -0,0 +1,61 @@
+$ git log -m --raw master
+commit 59d314ad6f356dd08601a4cd5e530381da3e3c64 (from 9a6d4949b6b76956d9d5e26f2791ec2ceff5fdc0)
+Merge: 9a6d494 c7a2ab9
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:04:00 2006 +0000
+
+    Merge branch 'side'
+
+:100644 100644 cead32e... 992913c... M	dir/sub
+:100644 100644 b414108... 10a8a9f... M	file0
+
+commit 59d314ad6f356dd08601a4cd5e530381da3e3c64 (from c7a2ab9e8eac7b117442a607d5a9b3950ae34d5a)
+Merge: 9a6d494 c7a2ab9
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:04:00 2006 +0000
+
+    Merge branch 'side'
+
+:100644 100644 7289e35... 992913c... M	dir/sub
+:100644 100644 f4615da... 10a8a9f... M	file0
+:000000 100644 0000000... b1e6722... A	file1
+:100644 000000 01e79c3... 0000000... D	file2
+:100644 000000 7289e35... 0000000... D	file3
+
+commit c7a2ab9e8eac7b117442a607d5a9b3950ae34d5a
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:03:00 2006 +0000
+
+    Side
+
+:100644 100644 35d242b... 7289e35... M	dir/sub
+:100644 100644 01e79c3... f4615da... M	file0
+:000000 100644 0000000... 7289e35... A	file3
+
+commit 9a6d4949b6b76956d9d5e26f2791ec2ceff5fdc0
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:02:00 2006 +0000
+
+    Third
+
+:100644 100644 8422d40... cead32e... M	dir/sub
+:000000 100644 0000000... b1e6722... A	file1
+
+commit 1bde4ae5f36c8d9abe3a0fce0c6aab3c4a12fe44
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:01:00 2006 +0000
+
+    Second
+    
+    This is the second commit.
+
+:100644 100644 35d242b... 8422d40... M	dir/sub
+:100644 100644 01e79c3... b414108... M	file0
+:100644 000000 01e79c3... 0000000... D	file2
+
+commit 444ac553ac7612cc88969031b02b3767fb8a353a
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:00:00 2006 +0000
+
+    Initial
+$
diff --git a/t/t4013/diff.log_-m_--stat_master b/t/t4013/diff.log_-m_--stat_master
new file mode 100644
index 0000000..c7db084
--- /dev/null
+++ b/t/t4013/diff.log_-m_--stat_master
@@ -0,0 +1,66 @@
+$ git log -m --stat master
+commit 59d314ad6f356dd08601a4cd5e530381da3e3c64 (from 9a6d4949b6b76956d9d5e26f2791ec2ceff5fdc0)
+Merge: 9a6d494 c7a2ab9
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:04:00 2006 +0000
+
+    Merge branch 'side'
+
+ dir/sub | 2 ++
+ file0   | 3 +++
+ 2 files changed, 5 insertions(+)
+
+commit 59d314ad6f356dd08601a4cd5e530381da3e3c64 (from c7a2ab9e8eac7b117442a607d5a9b3950ae34d5a)
+Merge: 9a6d494 c7a2ab9
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:04:00 2006 +0000
+
+    Merge branch 'side'
+
+ dir/sub | 4 ++++
+ file0   | 3 +++
+ file1   | 3 +++
+ file2   | 3 ---
+ file3   | 4 ----
+ 5 files changed, 10 insertions(+), 7 deletions(-)
+
+commit c7a2ab9e8eac7b117442a607d5a9b3950ae34d5a
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:03:00 2006 +0000
+
+    Side
+
+ dir/sub | 2 ++
+ file0   | 3 +++
+ file3   | 4 ++++
+ 3 files changed, 9 insertions(+)
+
+commit 9a6d4949b6b76956d9d5e26f2791ec2ceff5fdc0
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:02:00 2006 +0000
+
+    Third
+
+ dir/sub | 2 ++
+ file1   | 3 +++
+ 2 files changed, 5 insertions(+)
+
+commit 1bde4ae5f36c8d9abe3a0fce0c6aab3c4a12fe44
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:01:00 2006 +0000
+
+    Second
+    
+    This is the second commit.
+
+ dir/sub | 2 ++
+ file0   | 3 +++
+ file2   | 3 ---
+ 3 files changed, 5 insertions(+), 3 deletions(-)
+
+commit 444ac553ac7612cc88969031b02b3767fb8a353a
+Author: A U Thor <author@example.com>
+Date:   Mon Jun 26 00:00:00 2006 +0000
+
+    Initial
+$
diff --git a/t/t4030-diff-textconv.sh b/t/t4030-diff-textconv.sh
index c906320..a39a626 100755
--- a/t/t4030-diff-textconv.sh
+++ b/t/t4030-diff-textconv.sh
@@ -26,12 +26,8 @@
 chmod +x hexdump
 
 test_expect_success 'setup binary file with history' '
-	printf "\\0\\n" >file &&
-	git add file &&
-	git commit -m one &&
-	printf "\\01\\n" >>file &&
-	git add file &&
-	git commit -m two
+	test_commit --printf one file "\\0\\n" &&
+	test_commit --printf --append two file "\\01\\n"
 '
 
 test_expect_success 'file is considered binary by porcelain' '
diff --git a/t/t5406-remote-rejects.sh b/t/t5406-remote-rejects.sh
index ff06f99..5c509db 100755
--- a/t/t5406-remote-rejects.sh
+++ b/t/t5406-remote-rejects.sh
@@ -5,7 +5,6 @@
 . ./test-lib.sh
 
 test_expect_success 'setup' '
-	mkdir .git/hooks &&
 	write_script .git/hooks/update <<-\EOF &&
 	exit 1
 	EOF
diff --git a/t/t5407-post-rewrite-hook.sh b/t/t5407-post-rewrite-hook.sh
index 5bb23cc..6da8d76 100755
--- a/t/t5407-post-rewrite-hook.sh
+++ b/t/t5407-post-rewrite-hook.sh
@@ -20,8 +20,6 @@
 	git checkout main
 '
 
-mkdir .git/hooks
-
 cat >.git/hooks/post-rewrite <<EOF
 #!/bin/sh
 echo \$@ > "$TRASH_DIRECTORY"/post-rewrite.args
diff --git a/t/t5409-colorize-remote-messages.sh b/t/t5409-colorize-remote-messages.sh
index 5d8f401..9f1a483 100755
--- a/t/t5409-colorize-remote-messages.sh
+++ b/t/t5409-colorize-remote-messages.sh
@@ -5,7 +5,6 @@
 . ./test-lib.sh
 
 test_expect_success 'setup' '
-	mkdir .git/hooks &&
 	write_script .git/hooks/update <<-\EOF &&
 	echo error: error
 	echo ERROR: also highlighted
diff --git a/t/t5520-pull.sh b/t/t5520-pull.sh
index a094113..e2c0c51 100755
--- a/t/t5520-pull.sh
+++ b/t/t5520-pull.sh
@@ -746,14 +746,8 @@
 '
 
 test_expect_success 'setup for detecting upstreamed changes' '
-	mkdir src &&
-	(
-		cd src &&
-		git init &&
-		printf "1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n" > stuff &&
-		git add stuff &&
-		git commit -m "Initial revision"
-	) &&
+	test_create_repo src &&
+	test_commit -C src --printf one stuff "1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n" &&
 	git clone src dst &&
 	(
 		cd src &&
diff --git a/t/t5600-clone-fail-cleanup.sh b/t/t5600-clone-fail-cleanup.sh
index 4a1a912..5bf1026 100755
--- a/t/t5600-clone-fail-cleanup.sh
+++ b/t/t5600-clone-fail-cleanup.sh
@@ -97,4 +97,11 @@
 	test_dir_is_empty empty-wt
 '
 
+test_expect_success 'transport failure cleans up directory' '
+	test_must_fail git clone --no-local \
+		-u "f() { git-upload-pack \"\$@\"; return 1; }; f" \
+		foo broken-clone &&
+	test_path_is_missing broken-clone
+'
+
 test_done
diff --git a/t/t6041-bisect-submodule.sh b/t/t6041-bisect-submodule.sh
index df1eff0..82013fc 100755
--- a/t/t6041-bisect-submodule.sh
+++ b/t/t6041-bisect-submodule.sh
@@ -8,7 +8,7 @@
 git_bisect () {
 	git status -su >expect &&
 	ls -1pR * >>expect &&
-	tar cf "$TRASH_DIRECTORY/tmp.tar" * &&
+	"$TAR" cf "$TRASH_DIRECTORY/tmp.tar" * &&
 	GOOD=$(git rev-parse --verify HEAD) &&
 	may_only_be_test_must_fail "$2" &&
 	$2 git checkout "$1" &&
@@ -25,7 +25,7 @@
 	git bisect start &&
 	git bisect good $GOOD &&
 	rm -rf * &&
-	tar xf "$TRASH_DIRECTORY/tmp.tar" &&
+	"$TAR" xf "$TRASH_DIRECTORY/tmp.tar" &&
 	git status -su >actual &&
 	ls -1pR * >>actual &&
 	test_cmp expect actual &&
diff --git a/t/t6120-describe.sh b/t/t6120-describe.sh
index e89b674..88fddc9 100755
--- a/t/t6120-describe.sh
+++ b/t/t6120-describe.sh
@@ -31,64 +31,32 @@
 }
 
 test_expect_success setup '
+	test_commit initial file one &&
+	test_commit second file two &&
+	test_commit third file three &&
+	test_commit --annotate A file A &&
+	test_commit c file c &&
 
-	test_tick &&
-	echo one >file && git add file && git commit -m initial &&
-	one=$(git rev-parse HEAD) &&
-
-	git describe --always HEAD &&
-
-	test_tick &&
-	echo two >file && git add file && git commit -m second &&
-	two=$(git rev-parse HEAD) &&
-
-	test_tick &&
-	echo three >file && git add file && git commit -m third &&
-
-	test_tick &&
-	echo A >file && git add file && git commit -m A &&
-	test_tick &&
-	git tag -a -m A A &&
-
-	test_tick &&
-	echo c >file && git add file && git commit -m c &&
-	test_tick &&
-	git tag c &&
-
-	git reset --hard $two &&
-	test_tick &&
-	echo B >side && git add side && git commit -m B &&
-	test_tick &&
-	git tag -a -m B B &&
+	git reset --hard second &&
+	test_commit --annotate B side B &&
 
 	test_tick &&
 	git merge -m Merged c &&
 	merged=$(git rev-parse HEAD) &&
 
-	git reset --hard $two &&
-	test_tick &&
-	echo D >another && git add another && git commit -m D &&
-	test_tick &&
-	git tag -a -m D D &&
+	git reset --hard second &&
+	test_commit --no-tag D another D &&
+
 	test_tick &&
 	git tag -a -m R R &&
 
-	test_tick &&
-	echo DD >another && git commit -a -m another &&
-
-	test_tick &&
-	git tag e &&
-
-	test_tick &&
-	echo DDD >another && git commit -a -m "yet another" &&
+	test_commit e another DD &&
+	test_commit --no-tag "yet another" another DDD &&
 
 	test_tick &&
 	git merge -m Merged $merged &&
 
-	test_tick &&
-	echo X >file && echo X >side && git add file side &&
-	git commit -m x
-
+	test_commit --no-tag x file
 '
 
 check_describe A-* HEAD
diff --git a/t/t6423-merge-rename-directories.sh b/t/t6423-merge-rename-directories.sh
index 7134769..be84d22 100755
--- a/t/t6423-merge-rename-directories.sh
+++ b/t/t6423-merge-rename-directories.sh
@@ -4966,6 +4966,64 @@
 	)
 '
 
+# Testcase 12h, Testcase with two kinds of "relevant" renames
+#   Commit O: olddir/{a_1, b}
+#   Commit A: newdir/{a_2, b}
+#   Commit B: olddir/{alpha_1, b}
+#   Expected: newdir/{alpha_2, b}
+
+test_setup_12h () {
+	test_create_repo 12h &&
+	(
+		cd 12h &&
+
+		mkdir olddir &&
+		test_seq 3 8 >olddir/a &&
+		>olddir/b &&
+		git add olddir &&
+		git commit -m orig &&
+
+		git branch O &&
+		git branch A &&
+		git branch B &&
+
+		git switch A &&
+		test_seq 3 10 >olddir/a &&
+		git add olddir/a &&
+		git mv olddir newdir &&
+		git commit -m A &&
+
+		git switch B &&
+
+		git mv olddir/a olddir/alpha &&
+		git commit -m B
+	)
+}
+
+test_expect_failure '12h: renaming a file within a renamed directory' '
+	test_setup_12h &&
+	(
+		cd 12h &&
+
+		git checkout A^0 &&
+
+		test_might_fail git -c merge.directoryRenames=true merge -s recursive B^0 &&
+
+		git ls-files >tracked &&
+		test_line_count = 2 tracked &&
+
+		test_path_is_missing olddir/a &&
+		test_path_is_file newdir/alpha &&
+		test_path_is_file newdir/b &&
+
+		git rev-parse >actual \
+			HEAD:newdir/alpha  HEAD:newdir/b &&
+		git rev-parse >expect \
+			A:newdir/a         O:oldir/b &&
+		test_cmp expect actual
+	)
+'
+
 ###########################################################################
 # SECTION 13: Checking informational and conflict messages
 #
diff --git a/t/t6429-merge-sequence-rename-caching.sh b/t/t6429-merge-sequence-rename-caching.sh
new file mode 100755
index 0000000..035edc4
--- /dev/null
+++ b/t/t6429-merge-sequence-rename-caching.sh
@@ -0,0 +1,700 @@
+#!/bin/sh
+
+test_description="remember regular & dir renames in sequence of merges"
+
+. ./test-lib.sh
+
+#
+# NOTE 1: this testfile tends to not only rename files, but modify on both
+#         sides; without modifying on both sides, optimizations can kick in
+#         which make rename detection irrelevant or trivial.  We want to make
+#         sure that we are triggering rename caching rather than rename
+#         bypassing.
+#
+# NOTE 2: this testfile uses 'test-tool fast-rebase' instead of either
+#         cherry-pick or rebase.  sequencer.c is only superficially
+#         integrated with merge-ort; it calls merge_switch_to_result()
+#         after EACH merge, which updates the index and working copy AND
+#         throws away the cached results (because merge_switch_to_result()
+#         is only supposed to be called at the end of the sequence).
+#         Integrating them more deeply is a big task, so for now the tests
+#         use 'test-tool fast-rebase'.
+#
+
+
+#
+# In the following simple testcase:
+#   Base:     numbers_1, values_1
+#   Upstream: numbers_2, values_2
+#   Topic_1:  sequence_3
+#   Topic_2:  scruples_3
+# or, in english, rename numbers -> sequence in the first commit, and rename
+# values -> scruples in the second commit.
+#
+# This shouldn't be a challenge, it's just verifying that cached renames isn't
+# preventing us from finding new renames.
+#
+test_expect_success 'caching renames does not preclude finding new ones' '
+	test_create_repo caching-renames-and-new-renames &&
+	(
+		cd caching-renames-and-new-renames &&
+
+		test_seq 2 10 >numbers &&
+		test_seq 2 10 >values &&
+		git add numbers values &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch upstream &&
+		test_seq 1 10 >numbers &&
+		test_seq 1 10 >values &&
+		git add numbers values &&
+		git commit -m "Tweaked both files" &&
+
+		git switch topic &&
+
+		test_seq 2 12 >numbers &&
+		git add numbers &&
+		git mv numbers sequence &&
+		git commit -m A &&
+
+		test_seq 2 12 >values &&
+		git add values &&
+		git mv values scruples &&
+		git commit -m B &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream~1..topic
+
+		git ls-files >tracked-files &&
+		test_line_count = 2 tracked-files &&
+		test_seq 1 12 >expect &&
+		test_cmp expect sequence &&
+		test_cmp expect scruples
+	)
+'
+
+#
+# In the following testcase:
+#   Base:     numbers_1
+#   Upstream: rename numbers_1 -> sequence_2
+#   Topic_1:  numbers_3
+#   Topic_2:  numbers_1
+# or, in english, the first commit on the topic branch modifies numbers by
+# shrinking it (dramatically) and the second commit on topic reverts its
+# parent.
+#
+# Can git apply both patches?
+#
+# Traditional cherry-pick/rebase will fail to apply the second commit, the
+# one that reverted its parent, because despite detecting the rename from
+# 'numbers' to 'sequence' for the first commit, it fails to detect that
+# rename when picking the second commit.  That's "reasonable" given the
+# dramatic change in size of the file, but remembering the rename and
+# reusing it is reasonable too.
+#
+# We do test here that we expect rename detection to only be run once total
+# (the topic side of history doesn't need renames, and with caching we
+# should be able to only run rename detection on the upstream side one
+# time.)
+test_expect_success 'cherry-pick both a commit and its immediate revert' '
+	test_create_repo pick-commit-and-its-immediate-revert &&
+	(
+		cd pick-commit-and-its-immediate-revert &&
+
+		test_seq 11 30 >numbers &&
+		git add numbers &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch upstream &&
+		test_seq 1 30 >numbers &&
+		git add numbers &&
+		git mv numbers sequence &&
+		git commit -m "Renamed (and modified) numbers -> sequence" &&
+
+		git switch topic &&
+
+		test_seq 11 13 >numbers &&
+		git add numbers &&
+		git commit -m A &&
+
+		git revert HEAD &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream~1..topic &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 1 calls
+	)
+'
+
+#
+# In the following testcase:
+#   Base:     sequence_1
+#   Upstream: rename sequence_1 -> values_2
+#   Topic_1:  rename sequence_1 -> values_3
+#   Topic_2:  add unrelated sequence_4
+# or, in english, both sides rename sequence -> values, and then the second
+# commit on the topic branch adds an unrelated file called sequence.
+#
+# This testcase presents no problems for git traditionally, but having both
+# sides do the same rename in effect "uses it up" and if it remains cached,
+# could cause a spurious rename/add conflict.
+#
+test_expect_success 'rename same file identically, then reintroduce it' '
+	test_create_repo rename-rename-1to1-then-add-old-filename &&
+	(
+		cd rename-rename-1to1-then-add-old-filename &&
+
+		test_seq 3 8 >sequence &&
+		git add sequence &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch upstream &&
+		test_seq 1 8 >sequence &&
+		git add sequence &&
+		git mv sequence values &&
+		git commit -m "Renamed (and modified) sequence -> values" &&
+
+		git switch topic &&
+
+		test_seq 3 10 >sequence &&
+		git add sequence &&
+		git mv sequence values &&
+		git commit -m A &&
+
+		test_write_lines A B C D E F G H I J >sequence &&
+		git add sequence &&
+		git commit -m B &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream~1..topic &&
+
+		git ls-files >tracked &&
+		test_line_count = 2 tracked &&
+		test_path_is_file values &&
+		test_path_is_file sequence &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 2 calls
+	)
+'
+
+#
+# In the following testcase:
+#   Base:     olddir/{valuesZ_1, valuesY_1, valuesX_1}
+#   Upstream: rename olddir/valuesZ_1 -> dirA/valuesZ_2
+#             rename olddir/valuesY_1 -> dirA/valuesY_2
+#             rename olddir/valuesX_1 -> dirB/valuesX_2
+#   Topic_1:  rename olddir/valuesZ_1 -> dirA/valuesZ_3
+#             rename olddir/valuesY_1 -> dirA/valuesY_3
+#   Topic_2:  add olddir/newfile
+#   Expected Pick1: dirA/{valuesZ, valuesY}, dirB/valuesX
+#   Expected Pick2: dirA/{valuesZ, valuesY}, dirB/{valuesX, newfile}
+#
+# This testcase presents no problems for git traditionally, but having both
+# sides do the same renames in effect "use it up" but if the renames remain
+# cached, the directory rename could put newfile in the wrong directory.
+#
+test_expect_success 'rename same file identically, then add file to old dir' '
+	test_create_repo rename-rename-1to1-then-add-file-to-old-dir &&
+	(
+		cd rename-rename-1to1-then-add-file-to-old-dir &&
+
+		mkdir olddir/ &&
+		test_seq 3 8 >olddir/valuesZ &&
+		test_seq 3 8 >olddir/valuesY &&
+		test_seq 3 8 >olddir/valuesX &&
+		git add olddir &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch upstream &&
+		test_seq 1 8 >olddir/valuesZ &&
+		test_seq 1 8 >olddir/valuesY &&
+		test_seq 1 8 >olddir/valuesX &&
+		git add olddir &&
+		mkdir dirA &&
+		git mv olddir/valuesZ olddir/valuesY dirA &&
+		git mv olddir/ dirB/ &&
+		git commit -m "Renamed (and modified) values*" &&
+
+		git switch topic &&
+
+		test_seq 3 10 >olddir/valuesZ &&
+		test_seq 3 10 >olddir/valuesY &&
+		git add olddir &&
+		mkdir dirA &&
+		git mv olddir/valuesZ olddir/valuesY dirA &&
+		git commit -m A &&
+
+		>olddir/newfile &&
+		git add olddir/newfile &&
+		git commit -m B &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+		git config merge.directoryRenames true &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream~1..topic &&
+
+		git ls-files >tracked &&
+		test_line_count = 4 tracked &&
+		test_path_is_file dirA/valuesZ &&
+		test_path_is_file dirA/valuesY &&
+		test_path_is_file dirB/valuesX &&
+		test_path_is_file dirB/newfile &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 3 calls
+	)
+'
+
+#
+# In the following testcase, upstream renames a directory, and the topic branch
+# first adds a file to the directory, then later renames the directory
+# differently:
+#   Base:     olddir/a
+#             olddir/b
+#   Upstream: rename olddir/ -> newdir/
+#   Topic_1:  add olddir/newfile
+#   Topic_2:  rename olddir/ -> otherdir/
+#
+# Here we are just concerned that cached renames might prevent us from seeing
+# the rename conflict, and we want to ensure that we do get a conflict.
+#
+# While at it, though, we do test that we only try to detect renames 2
+# times and not three.  (The first merge needs to detect renames on the
+# upstream side.  Traditionally, the second merge would need to detect
+# renames on both sides of history, but our caching of upstream renames
+# should avoid the need to re-detect upstream renames.)
+#
+test_expect_success 'cached dir rename does not prevent noticing later conflict' '
+	test_create_repo dir-rename-cache-not-occluding-later-conflict &&
+	(
+		cd dir-rename-cache-not-occluding-later-conflict &&
+
+		mkdir olddir &&
+		test_seq 3 10 >olddir/a &&
+		test_seq 3 10 >olddir/b &&
+		git add olddir &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch upstream &&
+		test_seq 3 10 >olddir/a &&
+		test_seq 3 10 >olddir/b &&
+		git add olddir &&
+		git mv olddir newdir &&
+		git commit -m "Dir renamed" &&
+
+		git switch topic &&
+
+		>olddir/newfile &&
+		git add olddir/newfile &&
+		git commit -m A &&
+
+		test_seq 1 8 >olddir/a &&
+		test_seq 1 8 >olddir/b &&
+		git add olddir &&
+		git mv olddir otherdir &&
+		git commit -m B &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+		git config merge.directoryRenames true &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test_must_fail test-tool fast-rebase --onto HEAD upstream~1 topic >output &&
+		#git cherry-pick upstream..topic &&
+
+		grep CONFLICT..rename/rename output &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 2 calls
+	)
+'
+
+# Helper for the next two tests
+test_setup_upstream_rename () {
+	test_create_repo $1 &&
+	(
+		cd $1 &&
+
+		test_seq 3 8 >somefile &&
+		test_seq 3 8 >relevant-rename &&
+		git add somefile relevant-rename &&
+		mkdir olddir &&
+		test_write_lines a b c d e f g >olddir/a &&
+		test_write_lines z y x w v u t >olddir/b &&
+		git add olddir &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch upstream &&
+		test_seq 1 8 >somefile &&
+		test_seq 1 8 >relevant-rename &&
+		git add somefile relevant-rename &&
+		git mv relevant-rename renamed &&
+		echo h >>olddir/a &&
+		echo s >>olddir/b &&
+		git add olddir &&
+		git mv olddir newdir &&
+		git commit -m "Dir renamed"
+	)
+}
+
+#
+# In the following testcase, upstream renames a file in the toplevel directory
+# as well as its only directory:
+#   Base:     relevant-rename_1
+#             somefile
+#             olddir/a
+#             olddir/b
+#   Upstream: rename relevant-rename_1 -> renamed_2
+#             rename olddir/           -> newdir/
+#   Topic_1:  relevant-rename_3
+#   Topic_2:  olddir/newfile_1
+#   Topic_3:  olddir/newfile_2
+#
+# In this testcase, since the first commit being picked only modifies a
+# file in the toplevel directory, the directory rename is irrelevant for
+# that first merge.  However, we need to notice the directory rename for
+# the merge that picks the second commit, and we don't want the third
+# commit to mess up its location either.  We want to make sure that
+# olddir/newfile doesn't exist in the result and that newdir/newfile does.
+#
+# We also test that we only do rename detection twice.  We never need
+# rename detection on the topic side of history, but we do need it twice on
+# the upstream side of history.  For the first topic commit, we only need
+# the
+#   relevant-rename -> renamed
+# rename, because olddir is unmodified by Topic_1.  For Topic_2, however,
+# the new file being added to olddir means files that were previously
+# irrelevant for rename detection are now relevant, forcing us to repeat
+# rename detection for the paths we don't already have cached.  Topic_3 also
+# tweaks olddir/newfile, but the renames in olddir/ will have been cached
+# from the second rename detection run.
+#
+test_expect_success 'dir rename unneeded, then add new file to old dir' '
+	test_setup_upstream_rename dir-rename-unneeded-until-new-file &&
+	(
+		cd dir-rename-unneeded-until-new-file &&
+
+		git switch topic &&
+
+		test_seq 3 10 >relevant-rename &&
+		git add relevant-rename &&
+		git commit -m A &&
+
+		echo foo >olddir/newfile &&
+		git add olddir/newfile &&
+		git commit -m B &&
+
+		echo bar >>olddir/newfile &&
+		git add olddir/newfile &&
+		git commit -m C &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+		git config merge.directoryRenames true &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream..topic &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 2 calls &&
+
+		git ls-files >tracked &&
+		test_line_count = 5 tracked &&
+		test_path_is_missing olddir/newfile &&
+		test_path_is_file newdir/newfile
+	)
+'
+
+#
+# The following testcase is *very* similar to the last one, but instead of
+# adding a new olddir/newfile, it renames somefile -> olddir/newfile:
+#   Base:     relevant-rename_1
+#             somefile_1
+#             olddir/a
+#             olddir/b
+#   Upstream: rename relevant-rename_1 -> renamed_2
+#             rename olddir/           -> newdir/
+#   Topic_1:  relevant-rename_3
+#   Topic_2:  rename somefile -> olddir/newfile_2
+#   Topic_3:  modify olddir/newfile_3
+#
+# In this testcase, since the first commit being picked only modifies a
+# file in the toplevel directory, the directory rename is irrelevant for
+# that first merge.  However, we need to notice the directory rename for
+# the merge that picks the second commit, and we don't want the third
+# commit to mess up its location either.  We want to make sure that
+# neither somefile or olddir/newfile exists in the result and that
+# newdir/newfile does.
+#
+# This testcase needs one more call to rename detection than the last
+# testcase, because of the somefile -> olddir/newfile rename in Topic_2.
+test_expect_success 'dir rename unneeded, then rename existing file into old dir' '
+	test_setup_upstream_rename dir-rename-unneeded-until-file-moved-inside &&
+	(
+		cd dir-rename-unneeded-until-file-moved-inside &&
+
+		git switch topic &&
+
+		test_seq 3 10 >relevant-rename &&
+		git add relevant-rename &&
+		git commit -m A &&
+
+		test_seq 1 10 >somefile &&
+		git add somefile &&
+		git mv somefile olddir/newfile &&
+		git commit -m B &&
+
+		test_seq 1 12 >olddir/newfile &&
+		git add olddir/newfile &&
+		git commit -m C &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+		git config merge.directoryRenames true &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream..topic &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 3 calls &&
+
+		test_path_is_missing somefile &&
+		test_path_is_missing olddir/newfile &&
+		test_path_is_file newdir/newfile &&
+		git ls-files >tracked &&
+		test_line_count = 4 tracked
+	)
+'
+
+# Helper for the next two tests
+test_setup_topic_rename () {
+	test_create_repo $1 &&
+	(
+		cd $1 &&
+
+		test_seq 3 8 >somefile &&
+		mkdir olddir &&
+		test_seq 3 8 >olddir/a &&
+		echo b >olddir/b &&
+		git add olddir somefile &&
+		git commit -m orig &&
+
+		git branch upstream &&
+		git branch topic &&
+
+		git switch topic &&
+		test_seq 1 8 >somefile &&
+		test_seq 1 8 >olddir/a &&
+		git add somefile olddir/a &&
+		git mv olddir newdir &&
+		git commit -m "Dir renamed" &&
+
+		test_seq 1 10 >somefile &&
+		git add somefile &&
+		mkdir olddir &&
+		>olddir/unrelated-file &&
+		git add olddir &&
+		git commit -m "Unrelated file in recreated old dir"
+	)
+}
+
+#
+# In the following testcase, the first commit on the topic branch renames
+# a directory, while the second recreates the old directory and places a
+# file into it:
+#   Base:     somefile
+#             olddir/a
+#             olddir/b
+#   Upstream: olddir/newfile
+#   Topic_1:  somefile_2
+#             rename olddir/ -> newdir/
+#   Topic_2:  olddir/unrelated-file
+#
+# Note that the first pick should merge:
+#   Base:     somefile
+#             olddir/{a,b}
+#   Upstream: olddir/newfile
+#   Topic_1:  rename olddir/ -> newdir/
+# For which the expected result (assuming merge.directoryRenames=true) is
+# clearly:
+#   Result:   somefile
+#             newdir/{a, b, newfile}
+#
+# While the second pick does the following three-way merge:
+#   Base (Topic_1):           somefile
+#                             newdir/{a,b}
+#   Upstream (Result from 1): same files as base, but adds newdir/newfile
+#   Topic_2:                  same files as base, but adds olddir/unrelated-file
+#
+# The second merge is pretty trivial; upstream adds newdir/newfile, and
+# topic_2 adds olddir/unrelated-file.  We're just testing that we don't
+# accidentally cache directory renames somehow and rename
+# olddir/unrelated-file to newdir/unrelated-file.
+#
+# This testcase should only need one call to diffcore_rename_extended().
+test_expect_success 'caching renames only on upstream side, part 1' '
+	test_setup_topic_rename cache-renames-only-upstream-add-file &&
+	(
+		cd cache-renames-only-upstream-add-file &&
+
+		git switch upstream &&
+
+		>olddir/newfile &&
+		git add olddir/newfile &&
+		git commit -m "Add newfile" &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+
+		git config merge.directoryRenames true &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream..topic &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 1 calls &&
+
+		git ls-files >tracked &&
+		test_line_count = 5 tracked &&
+		test_path_is_missing newdir/unrelated-file &&
+		test_path_is_file olddir/unrelated-file &&
+		test_path_is_file newdir/newfile &&
+		test_path_is_file newdir/b &&
+		test_path_is_file newdir/a &&
+		test_path_is_file somefile
+	)
+'
+
+#
+# The following testcase is *very* similar to the last one, but instead of
+# adding a new olddir/newfile, it renames somefile -> olddir/newfile:
+#   Base:     somefile
+#             olddir/a
+#             olddir/b
+#   Upstream: somefile_1 -> olddir/newfile
+#   Topic_1:  rename olddir/ -> newdir/
+#             somefile_2
+#   Topic_2:  olddir/unrelated-file
+#             somefile_3
+#
+# Much like the previous test, this case is actually trivial and we are just
+# making sure there isn't some spurious directory rename caching going on
+# for the wrong side of history.
+#
+#
+# This testcase should only need two calls to diffcore_rename_extended(),
+# both for the first merge, one for each side of history.
+#
+test_expect_success 'caching renames only on upstream side, part 2' '
+	test_setup_topic_rename cache-renames-only-upstream-rename-file &&
+	(
+		cd cache-renames-only-upstream-rename-file &&
+
+		git switch upstream &&
+
+		git mv somefile olddir/newfile &&
+		git commit -m "Add newfile" &&
+
+		#
+		# Actual testing
+		#
+
+		git switch upstream &&
+
+		git config merge.directoryRenames true &&
+
+		GIT_TRACE2_PERF="$(pwd)/trace.output" &&
+		export GIT_TRACE2_PERF &&
+
+		test-tool fast-rebase --onto HEAD upstream~1 topic &&
+		#git cherry-pick upstream..topic &&
+
+		grep region_enter.*diffcore_rename trace.output >calls &&
+		test_line_count = 2 calls &&
+
+		git ls-files >tracked &&
+		test_line_count = 4 tracked &&
+		test_path_is_missing newdir/unrelated-file &&
+		test_path_is_file olddir/unrelated-file &&
+		test_path_is_file newdir/newfile &&
+		test_path_is_file newdir/b &&
+		test_path_is_file newdir/a
+	)
+'
+
+test_done
diff --git a/t/t9001-send-email.sh b/t/t9001-send-email.sh
index 3b75400..30eff72 100755
--- a/t/t9001-send-email.sh
+++ b/t/t9001-send-email.sh
@@ -2167,6 +2167,37 @@
 	test_cmp expected-list actual-list
 '
 
+test_expect_success $PREREQ 'test using command name with --sendmail-cmd' '
+	clean_fake_sendmail &&
+	PATH="$(pwd):$PATH" \
+	git send-email \
+		--from="Example <nobody@example.com>" \
+		--to=nobody@example.com \
+		--sendmail-cmd="fake.sendmail" \
+		HEAD^ &&
+	test_path_is_file commandline1
+'
+
+test_expect_success $PREREQ 'test using arguments with --sendmail-cmd' '
+	clean_fake_sendmail &&
+	git send-email \
+		--from="Example <nobody@example.com>" \
+		--to=nobody@example.com \
+		--sendmail-cmd='\''"$(pwd)/fake.sendmail" -f nobody@example.com'\'' \
+		HEAD^ &&
+	test_path_is_file commandline1
+'
+
+test_expect_success $PREREQ 'test shell expression with --sendmail-cmd' '
+	clean_fake_sendmail &&
+	git send-email \
+		--from="Example <nobody@example.com>" \
+		--to=nobody@example.com \
+		--sendmail-cmd='\''f() { "$(pwd)/fake.sendmail" "$@"; };f'\'' \
+		HEAD^ &&
+	test_path_is_file commandline1
+'
+
 test_expect_success $PREREQ 'invoke hook' '
 	mkdir -p .git/hooks &&
 
diff --git a/t/t9100-git-svn-basic.sh b/t/t9100-git-svn-basic.sh
index 1d3fdcc..d5563ec 100755
--- a/t/t9100-git-svn-basic.sh
+++ b/t/t9100-git-svn-basic.sh
@@ -4,21 +4,13 @@
 #
 
 test_description='git svn basic tests'
-GIT_SVN_LC_ALL=${LC_ALL:-$LANG}
 
 GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=main
 export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME
 
 . ./lib-git-svn.sh
 
-case "$GIT_SVN_LC_ALL" in
-*.UTF-8)
-	test_set_prereq UTF8
-	;;
-*)
-	say "# UTF-8 locale not set, some tests skipped ($GIT_SVN_LC_ALL)"
-	;;
-esac
+prepare_utf8_locale
 
 test_expect_success 'git svn --version works anywhere' '
 	nongit git svn --version
@@ -187,8 +179,8 @@
 	test ! -h "$SVN_TREE"/exec-2.sh &&
 	test_cmp help "$SVN_TREE"/exec-2.sh'
 
-name="commit with UTF-8 message: locale: $GIT_SVN_LC_ALL"
-LC_ALL="$GIT_SVN_LC_ALL"
+name="commit with UTF-8 message: locale: $GIT_TEST_UTF8_LOCALE"
+LC_ALL="$GIT_TEST_UTF8_LOCALE"
 export LC_ALL
 # This test relies on the previous test, hence requires POSIXPERM,SYMLINKS
 test_expect_success UTF8,POSIXPERM,SYMLINKS "$name" "
diff --git a/t/t9115-git-svn-dcommit-funky-renames.sh b/t/t9115-git-svn-dcommit-funky-renames.sh
index 9b44a44..743fbe1 100755
--- a/t/t9115-git-svn-dcommit-funky-renames.sh
+++ b/t/t9115-git-svn-dcommit-funky-renames.sh
@@ -93,9 +93,9 @@
 # > ... All of the above characters, except for the backslash, are converted
 # > to special UNICODE characters in the range 0xf000 to 0xf0ff (the
 # > "Private use area") when creating or accessing files.
-prepare_a_utf8_locale
+prepare_utf8_locale
 test_expect_success UTF8,!MINGW,!UTF8_NFD_TO_NFC 'svn.pathnameencoding=cp932 new file on dcommit' '
-	LC_ALL=$a_utf8_locale &&
+	LC_ALL=$GIT_TEST_UTF8_LOCALE &&
 	export LC_ALL &&
 	neq=$(printf "\201\202") &&
 	git config svn.pathnameencoding cp932 &&
@@ -107,7 +107,7 @@
 
 # See the comment on the above test for setting of LC_ALL.
 test_expect_success !MINGW,!UTF8_NFD_TO_NFC 'svn.pathnameencoding=cp932 rename on dcommit' '
-	LC_ALL=$a_utf8_locale &&
+	LC_ALL=$GIT_TEST_UTF8_LOCALE &&
 	export LC_ALL &&
 	inf=$(printf "\201\207") &&
 	git config svn.pathnameencoding cp932 &&
diff --git a/t/t9129-git-svn-i18n-commitencoding.sh b/t/t9129-git-svn-i18n-commitencoding.sh
index 2c213ae..01e1e8a 100755
--- a/t/t9129-git-svn-i18n-commitencoding.sh
+++ b/t/t9129-git-svn-i18n-commitencoding.sh
@@ -14,12 +14,12 @@
 	test_cmp current "$1"
 }
 
-prepare_a_utf8_locale
+prepare_utf8_locale
 
 compare_svn_head_with () {
 	# extract just the log message and strip out committer info.
 	# don't use --limit here since svn 1.1.x doesn't have it,
-	LC_ALL="$a_utf8_locale" svn log $(git svn info --url) | perl -w -e '
+	LC_ALL="$GIT_TEST_UTF8_LOCALE" svn log $(git svn info --url) | perl -w -e '
 		use bytes;
 		$/ = ("-"x72) . "\n";
 		my @x = <STDIN>;
diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
index b823c14..f0448da 100644
--- a/t/test-lib-functions.sh
+++ b/t/test-lib-functions.sh
@@ -172,12 +172,23 @@
 #   --notick
 #	Do not call test_tick before making a commit
 #   --append
-#	Use "echo >>" instead of "echo >" when writing "<contents>" to
-#	"<file>"
+#	Use ">>" instead of ">" when writing "<contents>" to "<file>"
+#   --printf
+#       Use "printf" instead of "echo" when writing "<contents>" to
+#       "<file>", use this to write escape sequences such as "\0", a
+#       trailing "\n" won't be added automatically. This option
+#       supports nothing but the FORMAT of printf(1), i.e. no custom
+#       ARGUMENT(s).
 #   --signoff
 #	Invoke "git commit" with --signoff
 #   --author <author>
 #	Invoke "git commit" with --author <author>
+#   --no-tag
+#	Do not tag the resulting commit
+#   --annotate
+#	Create an annotated tag with "--annotate -m <message>". Calls
+#	test_tick between making the commit and tag, unless --notick
+#	is given.
 #
 # This will commit a file with the given contents and the given commit
 # message, and tag the resulting commit with the given tag name.
@@ -186,17 +197,21 @@
 
 test_commit () {
 	notick= &&
+	echo=echo &&
 	append= &&
 	author= &&
 	signoff= &&
 	indir= &&
-	no_tag= &&
+	tag=light &&
 	while test $# != 0
 	do
 		case "$1" in
 		--notick)
 			notick=yes
 			;;
+		--printf)
+			echo=printf
+			;;
 		--append)
 			append=yes
 			;;
@@ -218,7 +233,10 @@
 			shift
 			;;
 		--no-tag)
-			no_tag=yes
+			tag=none
+			;;
+		--annotate)
+			tag=annotate
 			;;
 		*)
 			break
@@ -230,9 +248,9 @@
 	file=${2:-"$1.t"} &&
 	if test -n "$append"
 	then
-		echo "${3-$1}" >>"$indir$file"
+		$echo "${3-$1}" >>"$indir$file"
 	else
-		echo "${3-$1}" >"$indir$file"
+		$echo "${3-$1}" >"$indir$file"
 	fi &&
 	git ${indir:+ -C "$indir"} add "$file" &&
 	if test -z "$notick"
@@ -242,10 +260,20 @@
 	git ${indir:+ -C "$indir"} commit \
 	    ${author:+ --author "$author"} \
 	    $signoff -m "$1" &&
-	if test -z "$no_tag"
-	then
+	case "$tag" in
+	none)
+		;;
+	light)
 		git ${indir:+ -C "$indir"} tag "${4:-$1}"
-	fi
+		;;
+	annotate)
+		if test -z "$notick"
+		then
+			test_tick
+		fi &&
+		git ${indir:+ -C "$indir"} tag -a -m "$1" "${4:-$1}"
+		;;
+	esac
 }
 
 # Call test_merge with the arguments "<message> <commit>", where <commit>
@@ -1215,22 +1243,10 @@
 		} && (exit \"\$eval_ret\"); eval_ret=\$?; $test_atexit_cleanup"
 }
 
-# Most tests can use the created repository, but some may need to create more.
+# Deprecated wrapper for "git init", use "git init" directly instead
 # Usage: test_create_repo <directory>
 test_create_repo () {
-	test "$#" = 1 ||
-	BUG "not 1 parameter to test-create-repo"
-	repo="$1"
-	mkdir -p "$repo"
-	(
-		cd "$repo" || error "Cannot setup test environment"
-		"${GIT_TEST_INSTALLED:-$GIT_EXEC_PATH}/git$X" -c \
-			init.defaultBranch="${GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME-master}" \
-			init \
-			"--template=$GIT_BUILD_DIR/templates/blt/" >&3 2>&4 ||
-		error "cannot run git init -- have you built things yet?"
-		mv .git/hooks .git/hooks-disabled
-	) || exit
+	git init "$@"
 }
 
 # This function helps on symlink challenged file systems when it is not
diff --git a/t/test-lib.sh b/t/test-lib.sh
index adaf035..54938c6 100644
--- a/t/test-lib.sh
+++ b/t/test-lib.sh
@@ -64,6 +64,11 @@
 	export GIT_TEST_DISALLOW_ABBREVIATED_OPTIONS
 fi
 
+# Explicitly set the default branch name for testing, to avoid the
+# transitory "git init" warning under --verbose.
+: ${GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME:=master}
+export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME
+
 ################################################################
 # It appears that people try to run tests without building...
 "${GIT_TEST_INSTALLED:-$GIT_BUILD_DIR}/git$X" >/dev/null
@@ -1172,7 +1177,7 @@
 			esac
 		fi
 
-		if test -z "$debug"
+		if test -z "$debug" && test -n "$remove_trash"
 		then
 			test -d "$TRASH_DIRECTORY" ||
 			error "Tests passed but trash directory already removed before test cleanup; aborting"
@@ -1337,28 +1342,8 @@
 	exit 1
 fi
 
-# Test repository
-rm -fr "$TRASH_DIRECTORY" || {
-	GIT_EXIT_OK=t
-	echo >&5 "FATAL: Cannot prepare test area"
-	exit 1
-}
-
-HOME="$TRASH_DIRECTORY"
-GNUPGHOME="$HOME/gnupg-home-not-used"
-export HOME GNUPGHOME
-
-if test -z "$TEST_NO_CREATE_REPO"
-then
-	test_create_repo "$TRASH_DIRECTORY"
-else
-	mkdir -p "$TRASH_DIRECTORY"
-fi
-
-# Use -P to resolve symlinks in our working directory so that the cwd
-# in subprocesses like git equals our $PWD (for pathname comparisons).
-cd -P "$TRASH_DIRECTORY" || exit 1
-
+# Are we running this test at all?
+remove_trash=
 this_test=${0##*/}
 this_test=${this_test%%-*}
 if match_pattern_list "$this_test" $GIT_SKIP_TESTS
@@ -1368,6 +1353,31 @@
 	test_done
 fi
 
+# Last-minute variable setup
+HOME="$TRASH_DIRECTORY"
+GNUPGHOME="$HOME/gnupg-home-not-used"
+export HOME GNUPGHOME
+
+# Test repository
+rm -fr "$TRASH_DIRECTORY" || {
+	GIT_EXIT_OK=t
+	echo >&5 "FATAL: Cannot prepare test area"
+	exit 1
+}
+
+remove_trash=t
+if test -z "$TEST_NO_CREATE_REPO"
+then
+	git init "$TRASH_DIRECTORY" >&3 2>&4 ||
+	error "cannot run git init"
+else
+	mkdir -p "$TRASH_DIRECTORY"
+fi
+
+# Use -P to resolve symlinks in our working directory so that the cwd
+# in subprocesses like git equals our $PWD (for pathname comparisons).
+cd -P "$TRASH_DIRECTORY" || exit 1
+
 if test -n "$write_junit_xml"
 then
 	junit_xml_dir="$TEST_OUTPUT_DIRECTORY/out"
diff --git a/trace2/tr2_dst.c b/trace2/tr2_dst.c
index ae052a0..bda283e 100644
--- a/trace2/tr2_dst.c
+++ b/trace2/tr2_dst.c
@@ -204,15 +204,16 @@
 
 	fd = socket(AF_UNIX, sock_type, 0);
 	if (fd == -1)
-		return errno;
+		return -1;
 
 	sa.sun_family = AF_UNIX;
 	strlcpy(sa.sun_path, path, sizeof(sa.sun_path));
 
 	if (connect(fd, (struct sockaddr *)&sa, sizeof(sa)) == -1) {
-		int e = errno;
+		int saved_errno = errno;
 		close(fd);
-		return e;
+		errno = saved_errno;
+		return -1;
 	}
 
 	*out_fd = fd;
@@ -227,7 +228,6 @@
 {
 	unsigned int uds_try = 0;
 	int fd;
-	int e;
 	const char *path = NULL;
 
 	/*
@@ -271,15 +271,13 @@
 	}
 
 	if (uds_try & TR2_DST_UDS_TRY_STREAM) {
-		e = tr2_dst_try_uds_connect(path, SOCK_STREAM, &fd);
-		if (!e)
+		if (!tr2_dst_try_uds_connect(path, SOCK_STREAM, &fd))
 			goto connected;
-		if (e != EPROTOTYPE)
+		if (errno != EPROTOTYPE)
 			goto error;
 	}
 	if (uds_try & TR2_DST_UDS_TRY_DGRAM) {
-		e = tr2_dst_try_uds_connect(path, SOCK_DGRAM, &fd);
-		if (!e)
+		if (!tr2_dst_try_uds_connect(path, SOCK_DGRAM, &fd))
 			goto connected;
 	}
 
@@ -287,7 +285,7 @@
 	if (tr2_dst_want_warning())
 		warning("trace2: could not connect to socket '%s' for '%s' tracing: %s",
 			path, tr2_sysenv_display_name(dst->sysenv_var),
-			strerror(e));
+			strerror(errno));
 
 	tr2_dst_trace_disable(dst);
 	return 0;
diff --git a/transport.c b/transport.c
index 6cf3da1..50f5830 100644
--- a/transport.c
+++ b/transport.c
@@ -427,7 +427,8 @@
 
 cleanup:
 	close(data->fd[0]);
-	close(data->fd[1]);
+	if (data->fd[1] >= 0)
+		close(data->fd[1]);
 	if (finish_connect(data->conn))
 		ret = -1;
 	data->conn = NULL;
@@ -869,7 +870,8 @@
 		if (data->got_remote_heads && !transport->stateless_rpc)
 			packet_flush(data->fd[1]);
 		close(data->fd[0]);
-		close(data->fd[1]);
+		if (data->fd[1] >= 0)
+			close(data->fd[1]);
 		finish_connect(data->conn);
 	}