Merge branch 'fc/git-complete-helper' into fc/git-prompt-script

By Michael Haggerty (17) and others
via Junio C Hamano (36) and Jeff King (1)
* fc/git-complete-helper: (54 commits)
  completion: add new __git_complete helper
  Update draft release notes to 1.7.11 (11th batch)
  Git 1.7.10.2
  document submdule.$name.update=none option for gitmodules
  The tenth batch of topics
  Update draft release notes to 1.7.10.2
  checkout: do not corrupt HEAD on empty repo
  apply: remove lego in i18n string in gitdiff_verify_name
  dir: convert to strbuf
  status: refactor colopts handling
  status: respect "-b" for porcelain format
  status: fix null termination with "-b"
  status: refactor null_termination option
  commit: refactor option parsing
  Documentation/git-config: describe and clarify "--local <file>" option
  reflog-walk: tell explicit --date=default from not having --date at all
  clone: fix progress-regression
  grep.c: remove redundant line of code
  checkout (detached): truncate list of orphaned commits at the new HEAD
  t2020-checkout-detach: check for the number of orphaned commits
  ...

39 files changed, 1084 insertions, 527 deletions

diff --git a/Documentation/RelNotes/1.7.10.2.txt b/Documentation/RelNotes/1.7.10.2.txt
index 04fe29404c..7a7e9d6fd1 100644
--- a/Documentation/RelNotes/1.7.10.2.txt
+++ b/Documentation/RelNotes/1.7.10.2.txt
@@ -14,6 +14,9 @@ Fixes since v1.7.10.1
  * HTTP transport that requires authentication did not work correctly when
    multiple connections are used simultaneously.
 
+ * Minor memory leak during unpack_trees (hence "merge" and "checkout"
+   to check out another branch) has been plugged.
+
  * In the older days, the header "Conflicts:" in "cherry-pick" and "merge"
    was separated by a blank line from the list of paths that follow for
    readability, but when "merge" was rewritten in C, we lost it by
@@ -24,6 +27,9 @@ Fixes since v1.7.10.1
    be both revision name and a pathname, even though $name can never be a
    path in the context of the command.
 
+ * The "include.path" facility in the configuration mechanism added in
+   1.7.10 forgot to interpret "~/path" and "~user/path" as it should.
+
  * "git config --rename-section" to rename an existing section into a
    bogus one did not check the new name.
 
@@ -33,11 +39,44 @@ Fixes since v1.7.10.1
  * The report from "git fetch" said "new branch" even for a non branch
    ref.
 
+ * The http-backend (the server side of the smart http transfer) used
+   to overwrite GIT_COMMITTER_NAME and GIT_COMMITTER_EMAIL with the
+   value obtained from REMOTE_USER unconditionally, making it
+   impossible for the server side site-specific customization to use
+   different identity sources to affect the names logged. It now uses
+   REMOTE_USER only as a fallback value.
+
+ * "log --graph" was not very friendly with "--stat" option and its
+   output had line breaks at wrong places.
+
  * Octopus merge strategy did not reduce heads that are recorded in the
    final commit correctly.
 
+ * "git push" over smart-http lost progress output a few releases ago;
+   this release resurrects it.
+
+ * The error and advice messages given by "git push" when it fails due
+   to non-ff were not very helpful to new users; it has been broken
+   into three cases, and each is given a separate advice message.
+
+ * The insn sheet given by "rebase -i" did not make it clear that the
+   insn lines can be re-ordered to affect the order of the commits in
+   the resulting history.
+
+ * "git repack" used to write out unreachable objects as loose objects
+   when repacking, even if such loose objects will immediately pruned
+   due to its age.
+
+ * A contrib script "rerere-train" did not work out of the box unless
+   user futzed with her $PATH.
+
+ * "git rev-parse --show-prefix" used to emit nothing when run at the
+   top-level of the working tree, but now it gives a blank line.
+
  * The i18n of error message "git stash save" was not properly done.
 
+ * "git submodule" used a sed script that some platforms mishandled.
+
  * When using a Perl script on a system where "perl" found on user's
    $PATH could be ancient or otherwise broken, we allow builders to
    specify the path to a good copy of Perl with $PERL_PATH.  The
diff --git a/Documentation/RelNotes/1.7.11.txt b/Documentation/RelNotes/1.7.11.txt
index 8dfeddcd68..9ef7e18ca5 100644
--- a/Documentation/RelNotes/1.7.11.txt
+++ b/Documentation/RelNotes/1.7.11.txt
@@ -25,13 +25,15 @@ UI, Workflows & Features
    tracking.  Also "branch" learned the "-q"uiet option to squelch
    informational message.
 
+ * Your build platform may support hardlinks but you may prefer not to
+   use them, e.g. when installing to DESTDIR to make a tarball and
+   untarring on a filesystem that has poor support for hardlinks.
+   There is a Makefile option NO_INSTALL_HARDLINKS for you.
+
  * The smart-http backend used to always override GIT_COMMITTER_*
    variables with REMOTE_USER and REMOTE_ADDR, but these variables are
    now preserved when set.
 
- * "include.path" mechanism of the configuration files learned to
-   understand "~/path" and "~user/path".
-
  * "git am" learned the "--include" option, which is an opposite of
    existing the "--exclude" option.
 
@@ -51,9 +53,6 @@ UI, Workflows & Features
  * The "fmt-merge-msg" command learns to list the primary contributors
    involved in the side topic you are merging.
 
- * The cases "git push" fails due to non-ff can be broken into three
-   categories; each case is given a separate advise message.
-
  * "git rebase" learned to optionally keep commits that do not
    introduce any change in the original history.
 
@@ -83,27 +82,20 @@ Performance and Internal Implementation (please report possible regressions)
  * An experimental "version 4" format of the index file has been
    introduced to reduce on-disk footprint and I/O overhead.
 
+ * "git archive" learned to produce its output without reading the
+   blob object it writes out in memory in its entirety.
+
  * The code to compute hash values for lines used by the internal diff
    engine was optimized on little-endian machines, using the same
    trick the kernel folks came up with.
 
  * "git apply" had some memory leaks plugged.
 
- * "git repack" used to write out unreachable objects as loose objects
-   when repacking, even if such loose objects will immediately pruned
-   due to its age.
-
  * Setting up a revision traversal with many starting points was
    inefficient as these were placed in a date-order priority queue
    one-by-one.  Now they are collected in the queue unordered first,
    and sorted immediately before getting used.
 
- * "git rev-parse --show-prefix" used to emit nothing when run at the
-   top-level of the working tree, but now it gives a blank line.
-
- * Minor memory leak during unpack_trees (hence "merge" and "checkout"
-   to check out another branch) has been plugged.
-
  * More lower-level commands learned to use the streaming API to read
    from the object store without keeping everything in core.
 
@@ -124,6 +116,30 @@ Unless otherwise noted, all the fixes since v1.7.10 in the maintenance
 releases are contained in this release (see release notes to them for
 details).
 
+ * The DWIM behaviour for "log --pretty=format:%gd -g" was somewhat
+   broken and gave undue precedence to configured log.date, causing
+   "git stash list" to show "stash@{time stamp string}".
+   (merge 55ccf85 jk/maint-reflog-walk-count-vs-time later to maint).
+
+ * Running "git checkout" on an unborn branch used to corrupt HEAD.
+   (merge 8338f77 ef/checkout-empty later to maint).
+
+ * When checking out another commit from an already detached state, we
+   used to report all commits that are not reachable from any of the
+   refs as lossage, but some of them might be reachable from the new
+   HEAD, and there is no need to warn about them.
+   (merge 5d88639 js/checkout-detach-count later to maint).
+
+ * Some time ago, "git clone" lost the progress output for its
+   "checkout" phase; when run without any "--quiet" option, it should
+   give progress to the lengthy operation.
+   (merge 8f63da1 ef/maint-clone-progress-fix later to maint).
+
+ * "git status --porcelain" ignored "--branch" option by mistake.  The
+   output for "git status --branch -z" was also incorrect and did not
+   terminate the record for the current branch name with NUL as asked.
+   (merge d4a6bf1 jk/maint-status-porcelain-z-b later to maint).
+
  * "git diff --stat" used to fully count a binary file with modified
    execution bits whose contents is unmodified, which was not quite
    right.
@@ -132,14 +148,3 @@ details).
    NUL.  The fix is not entirely correct when the output also asks for
    --patch and/or --stat, though.
    (merge fafd382 jk/maint-tformat-with-z later to maint).
-
- * "git push" over smart-http lost progress output a few releases ago.
-   (merge e304aeb jk/maint-push-progress later to maint).
-
- * A contrib script "rerere-train" did not work out of the box unless
-   user futzed with her $PATH.
-   (merge 53876fc jc/rerere-train later to maint).
-
- * "log --graph" was not very friendly with "--stat" option and its
-   output had line breaks at wrong places.
-   (merge bafa16e lp/diffstat-with-graph later to maint).
diff --git a/Documentation/git-config.txt b/Documentation/git-config.txt
index 81b03982e3..3f5d216a09 100644
--- a/Documentation/git-config.txt
+++ b/Documentation/git-config.txt
@@ -44,11 +44,15 @@ a "true" or "false" string for bool), or '--path', which does some
 path expansion (see '--path' below).  If no type specifier is passed, no
 checks or transformations are performed on the value.
 
-The file-option can be one of '--system', '--global' or '--file'
-which specify where the values will be read from or written to.
-The default is to assume the config file of the current repository,
-.git/config unless defined otherwise with GIT_DIR and GIT_CONFIG
-(see <<FILES>>).
+When reading, the values are read from the system, global and
+repository local configuration files by default, and options
+'--system', '--global', '--local' and '--file <filename>' can be
+used to tell the command to read from only that location (see <<FILES>>).
+
+When writing, the new value is written to the repository local
+configuration file by default, and options '--system', '--global',
+'--file <filename>' can be used to tell the command to write to
+that location (you can say '--local' but that is the default).
 
 This command will fail (with exit code ret) if:
 
diff --git a/Documentation/git-status.txt b/Documentation/git-status.txt
index 2883a285ba..67e5f53a9e 100644
--- a/Documentation/git-status.txt
+++ b/Documentation/git-status.txt
@@ -184,7 +184,7 @@ order is reversed (e.g 'from \-> to' becomes 'to from'). Second, a NUL
 and the terminating newline (but a space still separates the status
 field from the first filename).  Third, filenames containing special
 characters are not specially formatted; no quoting or
-backslash-escaping is performed. Fourth, there is no branch line.
+backslash-escaping is performed.
 
 CONFIGURATION
 -------------
diff --git a/Documentation/git-submodule.txt b/Documentation/git-submodule.txt
index c243ee552b..c83a856574 100644
--- a/Documentation/git-submodule.txt
+++ b/Documentation/git-submodule.txt
@@ -140,7 +140,8 @@ update::
 	checkout the commit specified in the index of the containing repository.
 	This will make the submodules HEAD be detached unless `--rebase` or
 	`--merge` is specified or the key `submodule.$name.update` is set to
-	`rebase`, `merge` or `none`.
+	`rebase`, `merge` or `none`. `none` can be overriden by specifying
+	`--checkout`.
 +
 If the submodule is not yet initialized, and you just want to use the
 setting as stored in .gitmodules, you can automatically initialize the
@@ -148,10 +149,6 @@ submodule with the `--init` option.
 +
 If `--recursive` is specified, this command will recurse into the
 registered submodules, and update any nested submodules within.
-+
-If the configuration key `submodule.$name.update` is set to `none` the
-submodule with name `$name` will not be updated by default. This can be
-overriden by adding `--checkout` to the command.
 
 summary::
 	Show commit summary between the given commit (defaults to HEAD) and
diff --git a/Documentation/git.txt b/Documentation/git.txt
index 8527775988..c543213126 100644
--- a/Documentation/git.txt
+++ b/Documentation/git.txt
@@ -44,9 +44,10 @@ unreleased) version of git, that is available from 'master'
 branch of the `git.git` repository.
 Documentation for older releases are available here:
 
-* link:v1.7.10.1/git.html[documentation for release 1.7.10.1]
+* link:v1.7.10.2/git.html[documentation for release 1.7.10.2]
 
 * release notes for
+  link:RelNotes/1.7.10.2.txt[1.7.10.2],
   link:RelNotes/1.7.10.1.txt[1.7.10.1],
   link:RelNotes/1.7.10.txt[1.7.10].
 
diff --git a/Documentation/gitmodules.txt b/Documentation/gitmodules.txt
index 4e1fd52e7d..4effd78902 100644
--- a/Documentation/gitmodules.txt
+++ b/Documentation/gitmodules.txt
@@ -41,8 +41,11 @@ submodule.<name>.update::
 	the commit specified in the superproject. If 'merge', the commit
 	specified in the superproject will be merged into the current branch
 	in the submodule.
+	If 'none', the submodule with name `$name` will not be updated
+	by default.
+
 	This config option is overridden if 'git submodule update' is given
-	the '--merge' or '--rebase' options.
+	the '--merge', '--rebase' or '--checkout' options.
 
 submodule.<name>.fetchRecurseSubmodules::
 	This option can be used to control recursive fetching of this
diff --git a/Makefile b/Makefile
index 2fa7211055..3eda1e8c68 100644
--- a/Makefile
+++ b/Makefile
@@ -247,6 +247,9 @@ all::
 # Define NO_CROSS_DIRECTORY_HARDLINKS if you plan to distribute the installed
 # programs as a tar, where bin/ and libexec/ might be on different file systems.
 #
+# Define NO_INSTALL_HARDLINKS if you prefer to use either symbolic links or
+# copies to install built-in git commands e.g. git-cat-file.
+#
 # Define USE_NED_ALLOCATOR if you want to replace the platforms default
 # memory allocators with the nedmalloc allocator written by Niall Douglas.
 #
@@ -1835,6 +1838,10 @@ ifdef ASCIIDOC7
 	export ASCIIDOC7
 endif
 
+ifdef NO_INSTALL_HARDLINKS
+	export NO_INSTALL_HARDLINKS
+endif
+
 ### profile feedback build
 #
 
@@ -2574,19 +2581,21 @@ endif
 	{ test "$$bindir/" = "$$execdir/" || \
 	  for p in git$X $(filter $(install_bindir_programs),$(ALL_PROGRAMS)); do \
 		$(RM) "$$execdir/$$p" && \
-		test -z "$(NO_CROSS_DIRECTORY_HARDLINKS)" && \
+		test -z "$(NO_INSTALL_HARDLINKS)$(NO_CROSS_DIRECTORY_HARDLINKS)" && \
 		ln "$$bindir/$$p" "$$execdir/$$p" 2>/dev/null || \
 		cp "$$bindir/$$p" "$$execdir/$$p" || exit; \
 	  done; \
 	} && \
 	for p in $(filter $(install_bindir_programs),$(BUILT_INS)); do \
 		$(RM) "$$bindir/$$p" && \
+		test -z "$(NO_INSTALL_HARDLINKS)" && \
 		ln "$$bindir/git$X" "$$bindir/$$p" 2>/dev/null || \
 		ln -s "git$X" "$$bindir/$$p" 2>/dev/null || \
 		cp "$$bindir/git$X" "$$bindir/$$p" || exit; \
 	done && \
 	for p in $(BUILT_INS); do \
 		$(RM) "$$execdir/$$p" && \
+		test -z "$(NO_INSTALL_HARDLINKS)" && \
 		ln "$$execdir/git$X" "$$execdir/$$p" 2>/dev/null || \
 		ln -s "git$X" "$$execdir/$$p" 2>/dev/null || \
 		cp "$$execdir/git$X" "$$execdir/$$p" || exit; \
@@ -2594,6 +2603,7 @@ endif
 	remote_curl_aliases="$(REMOTE_CURL_ALIASES)" && \
 	for p in $$remote_curl_aliases; do \
 		$(RM) "$$execdir/$$p" && \
+		test -z "$(NO_INSTALL_HARDLINKS)" && \
 		ln "$$execdir/git-remote-http$X" "$$execdir/$$p" 2>/dev/null || \
 		ln -s "git-remote-http$X" "$$execdir/$$p" 2>/dev/null || \
 		cp "$$execdir/git-remote-http$X" "$$execdir/$$p" || exit; \
diff --git a/archive-tar.c b/archive-tar.c
index 20af0051a3..93387ea336 100644
--- a/archive-tar.c
+++ b/archive-tar.c
@@ -4,6 +4,7 @@
 #include "cache.h"
 #include "tar.h"
 #include "archive.h"
+#include "streaming.h"
 #include "run-command.h"
 
 #define RECORDSIZE	(512)
@@ -30,10 +31,9 @@ static void write_if_needed(void)
  * queues up writes, so that all our write(2) calls write exactly one
  * full block; pads writes to RECORDSIZE
  */
-static void write_blocked(const void *data, unsigned long size)
+static void do_write_blocked(const void *data, unsigned long size)
 {
 	const char *buf = data;
-	unsigned long tail;
 
 	if (offset) {
 		unsigned long chunk = BLOCKSIZE - offset;
@@ -54,6 +54,11 @@ static void write_blocked(const void *data, unsigned long size)
 		memcpy(block + offset, buf, size);
 		offset += size;
 	}
+}
+
+static void finish_record(void)
+{
+	unsigned long tail;
 	tail = offset % RECORDSIZE;
 	if (tail)  {
 		memset(block + offset, 0, RECORDSIZE - tail);
@@ -62,6 +67,12 @@ static void write_blocked(const void *data, unsigned long size)
 	write_if_needed();
 }
 
+static void write_blocked(const void *data, unsigned long size)
+{
+	do_write_blocked(data, size);
+	finish_record();
+}
+
 /*
  * The end of tar archives is marked by 2*512 nul bytes and after that
  * follows the rest of the block (if any).
@@ -78,6 +89,33 @@ static void write_trailer(void)
 }
 
 /*
+ * queues up writes, so that all our write(2) calls write exactly one
+ * full block; pads writes to RECORDSIZE
+ */
+static int stream_blocked(const unsigned char *sha1)
+{
+	struct git_istream *st;
+	enum object_type type;
+	unsigned long sz;
+	char buf[BLOCKSIZE];
+	ssize_t readlen;
+
+	st = open_istream(sha1, &type, &sz, NULL);
+	if (!st)
+		return error("cannot stream blob %s", sha1_to_hex(sha1));
+	for (;;) {
+		readlen = read_istream(st, buf, sizeof(buf));
+		if (readlen <= 0)
+			break;
+		do_write_blocked(buf, readlen);
+	}
+	close_istream(st);
+	if (!readlen)
+		finish_record();
+	return readlen;
+}
+
+/*
  * pax extended header records have the format "%u %s=%s\n".  %u contains
  * the size of the whole string (including the %u), the first %s is the
  * keyword, the second one is the value.  This function constructs such a
@@ -123,56 +161,101 @@ static size_t get_path_prefix(const char *path, size_t pathlen, size_t maxlen)
 	return i;
 }
 
+static void prepare_header(struct archiver_args *args,
+			   struct ustar_header *header,
+			   unsigned int mode, unsigned long size)
+{
+	sprintf(header->mode, "%07o", mode & 07777);
+	sprintf(header->size, "%011lo", S_ISREG(mode) ? size : 0);
+	sprintf(header->mtime, "%011lo", (unsigned long) args->time);
+
+	sprintf(header->uid, "%07o", 0);
+	sprintf(header->gid, "%07o", 0);
+	strlcpy(header->uname, "root", sizeof(header->uname));
+	strlcpy(header->gname, "root", sizeof(header->gname));
+	sprintf(header->devmajor, "%07o", 0);
+	sprintf(header->devminor, "%07o", 0);
+
+	memcpy(header->magic, "ustar", 6);
+	memcpy(header->version, "00", 2);
+
+	sprintf(header->chksum, "%07o", ustar_header_chksum(header));
+}
+
+static int write_extended_header(struct archiver_args *args,
+				 const unsigned char *sha1,
+				 const void *buffer, unsigned long size)
+{
+	struct ustar_header header;
+	unsigned int mode;
+	memset(&header, 0, sizeof(header));
+	*header.typeflag = TYPEFLAG_EXT_HEADER;
+	mode = 0100666;
+	sprintf(header.name, "%s.paxheader", sha1_to_hex(sha1));
+	prepare_header(args, &header, mode, size);
+	write_blocked(&header, sizeof(header));
+	write_blocked(buffer, size);
+	return 0;
+}
+
 static int write_tar_entry(struct archiver_args *args,
-		const unsigned char *sha1, const char *path, size_t pathlen,
-		unsigned int mode, void *buffer, unsigned long size)
+			   const unsigned char *sha1,
+			   const char *path, size_t pathlen,
+			   unsigned int mode)
 {
 	struct ustar_header header;
 	struct strbuf ext_header = STRBUF_INIT;
+	unsigned int old_mode = mode;
+	unsigned long size;
+	void *buffer;
 	int err = 0;
 
 	memset(&header, 0, sizeof(header));
 
-	if (!sha1) {
-		*header.typeflag = TYPEFLAG_GLOBAL_HEADER;
-		mode = 0100666;
-		strcpy(header.name, "pax_global_header");
-	} else if (!path) {
-		*header.typeflag = TYPEFLAG_EXT_HEADER;
-		mode = 0100666;
-		sprintf(header.name, "%s.paxheader", sha1_to_hex(sha1));
+	if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
+		*header.typeflag = TYPEFLAG_DIR;
+		mode = (mode | 0777) & ~tar_umask;
+	} else if (S_ISLNK(mode)) {
+		*header.typeflag = TYPEFLAG_LNK;
+		mode |= 0777;
+	} else if (S_ISREG(mode)) {
+		*header.typeflag = TYPEFLAG_REG;
+		mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
 	} else {
-		if (S_ISDIR(mode) || S_ISGITLINK(mode)) {
-			*header.typeflag = TYPEFLAG_DIR;
-			mode = (mode | 0777) & ~tar_umask;
-		} else if (S_ISLNK(mode)) {
-			*header.typeflag = TYPEFLAG_LNK;
-			mode |= 0777;
-		} else if (S_ISREG(mode)) {
-			*header.typeflag = TYPEFLAG_REG;
-			mode = (mode | ((mode & 0100) ? 0777 : 0666)) & ~tar_umask;
+		return error("unsupported file mode: 0%o (SHA1: %s)",
+			     mode, sha1_to_hex(sha1));
+	}
+	if (pathlen > sizeof(header.name)) {
+		size_t plen = get_path_prefix(path, pathlen,
+					      sizeof(header.prefix));
+		size_t rest = pathlen - plen - 1;
+		if (plen > 0 && rest <= sizeof(header.name)) {
+			memcpy(header.prefix, path, plen);
+				memcpy(header.name, path + plen + 1, rest);
 		} else {
-			return error("unsupported file mode: 0%o (SHA1: %s)",
-					mode, sha1_to_hex(sha1));
+			sprintf(header.name, "%s.data",
+				sha1_to_hex(sha1));
+			strbuf_append_ext_header(&ext_header, "path",
+						 path, pathlen);
 		}
-		if (pathlen > sizeof(header.name)) {
-			size_t plen = get_path_prefix(path, pathlen,
-					sizeof(header.prefix));
-			size_t rest = pathlen - plen - 1;
-			if (plen > 0 && rest <= sizeof(header.name)) {
-				memcpy(header.prefix, path, plen);
-				memcpy(header.name, path + plen + 1, rest);
-			} else {
-				sprintf(header.name, "%s.data",
-				        sha1_to_hex(sha1));
-				strbuf_append_ext_header(&ext_header, "path",
-						path, pathlen);
-			}
-		} else
-			memcpy(header.name, path, pathlen);
+	} else
+		memcpy(header.name, path, pathlen);
+
+	if (S_ISREG(mode) && !args->convert &&
+	    sha1_object_info(sha1, &size) == OBJ_BLOB &&
+	    size > big_file_threshold)
+		buffer = NULL;
+	else if (S_ISLNK(mode) || S_ISREG(mode)) {
+		enum object_type type;
+		buffer = sha1_file_to_archive(args, path, sha1, old_mode, &type, &size);
+		if (!buffer)
+			return error("cannot read %s", sha1_to_hex(sha1));
+	} else {
+		buffer = NULL;
+		size = 0;
 	}
 
-	if (S_ISLNK(mode) && buffer) {
+	if (S_ISLNK(mode)) {
 		if (size > sizeof(header.linkname)) {
 			sprintf(header.linkname, "see %s.paxheader",
 			        sha1_to_hex(sha1));
@@ -182,32 +265,25 @@ static int write_tar_entry(struct archiver_args *args,
 			memcpy(header.linkname, buffer, size);
 	}
 
-	sprintf(header.mode, "%07o", mode & 07777);
-	sprintf(header.size, "%011lo", S_ISREG(mode) ? size : 0);
-	sprintf(header.mtime, "%011lo", (unsigned long) args->time);
-
-	sprintf(header.uid, "%07o", 0);
-	sprintf(header.gid, "%07o", 0);
-	strlcpy(header.uname, "root", sizeof(header.uname));
-	strlcpy(header.gname, "root", sizeof(header.gname));
-	sprintf(header.devmajor, "%07o", 0);
-	sprintf(header.devminor, "%07o", 0);
-
-	memcpy(header.magic, "ustar", 6);
-	memcpy(header.version, "00", 2);
-
-	sprintf(header.chksum, "%07o", ustar_header_chksum(&header));
+	prepare_header(args, &header, mode, size);
 
 	if (ext_header.len > 0) {
-		err = write_tar_entry(args, sha1, NULL, 0, 0, ext_header.buf,
-				ext_header.len);
-		if (err)
+		err = write_extended_header(args, sha1, ext_header.buf,
+					    ext_header.len);
+		if (err) {
+			free(buffer);
 			return err;
+		}
 	}
 	strbuf_release(&ext_header);
 	write_blocked(&header, sizeof(header));
-	if (S_ISREG(mode) && buffer && size > 0)
-		write_blocked(buffer, size);
+	if (S_ISREG(mode) && size > 0) {
+		if (buffer)
+			write_blocked(buffer, size);
+		else
+			err = stream_blocked(sha1);
+	}
+	free(buffer);
 	return err;
 }
 
@@ -215,11 +291,18 @@ static int write_global_extended_header(struct archiver_args *args)
 {
 	const unsigned char *sha1 = args->commit_sha1;
 	struct strbuf ext_header = STRBUF_INIT;
-	int err;
+	struct ustar_header header;
+	unsigned int mode;
+	int err = 0;
 
 	strbuf_append_ext_header(&ext_header, "comment", sha1_to_hex(sha1), 40);
-	err = write_tar_entry(args, NULL, NULL, 0, 0, ext_header.buf,
-			ext_header.len);
+	memset(&header, 0, sizeof(header));
+	*header.typeflag = TYPEFLAG_GLOBAL_HEADER;
+	mode = 0100666;
+	strcpy(header.name, "pax_global_header");
+	prepare_header(args, &header, mode, ext_header.len);
+	write_blocked(&header, sizeof(header));
+	write_blocked(ext_header.buf, ext_header.len);
 	strbuf_release(&ext_header);
 	return err;
 }
diff --git a/archive-zip.c b/archive-zip.c
index 02d1f3787a..f5af81f904 100644
--- a/archive-zip.c
+++ b/archive-zip.c
@@ -3,6 +3,7 @@
  */
 #include "cache.h"
 #include "archive.h"
+#include "streaming.h"
 
 static int zip_date;
 static int zip_time;
@@ -15,6 +16,7 @@ static unsigned int zip_dir_offset;
 static unsigned int zip_dir_entries;
 
 #define ZIP_DIRECTORY_MIN_SIZE	(1024 * 1024)
+#define ZIP_STREAM (8)
 
 struct zip_local_header {
 	unsigned char magic[4];
@@ -31,6 +33,14 @@ struct zip_local_header {
 	unsigned char _end[1];
 };
 
+struct zip_data_desc {
+	unsigned char magic[4];
+	unsigned char crc32[4];
+	unsigned char compressed_size[4];
+	unsigned char size[4];
+	unsigned char _end[1];
+};
+
 struct zip_dir_header {
 	unsigned char magic[4];
 	unsigned char creator_version[2];
@@ -70,6 +80,7 @@ struct zip_dir_trailer {
  * we're interested in.
  */
 #define ZIP_LOCAL_HEADER_SIZE	offsetof(struct zip_local_header, _end)
+#define ZIP_DATA_DESC_SIZE	offsetof(struct zip_data_desc, _end)
 #define ZIP_DIR_HEADER_SIZE	offsetof(struct zip_dir_header, _end)
 #define ZIP_DIR_TRAILER_SIZE	offsetof(struct zip_dir_trailer, _end)
 
@@ -120,20 +131,59 @@ static void *zlib_deflate(void *data, unsigned long size,
 	return buffer;
 }
 
+static void write_zip_data_desc(unsigned long size,
+				unsigned long compressed_size,
+				unsigned long crc)
+{
+	struct zip_data_desc trailer;
+
+	copy_le32(trailer.magic, 0x08074b50);
+	copy_le32(trailer.crc32, crc);
+	copy_le32(trailer.compressed_size, compressed_size);
+	copy_le32(trailer.size, size);
+	write_or_die(1, &trailer, ZIP_DATA_DESC_SIZE);
+}
+
+static void set_zip_dir_data_desc(struct zip_dir_header *header,
+				  unsigned long size,