On AIX, access(X_OK) may succeed when run as root even if the
execution isn't possible. This behavior is allowed by POSIX
which says:

  ... for a process with appropriate privileges, an implementation
  may indicate success for X_OK even if execute permission is not
  granted to any user.

It can lead hook programs to have their execution refused:

   git commit -m content
   fatal: cannot exec '.git/hooks/pre-commit': Permission denied

Add NEED_ACCESS_ROOT_HANDLER in order to use an access helper function.
It checks with stat if any executable flags is set when the current user
is root.
Signed-off-by: NClément Chigot <clement.chigot@atos.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Since 33533975 ("Makefile: ask "ls-files" to list source files if
available", 2011-10-18), we optionally asked "ls-files" to list the
source files that ought to exist, as a faster approximation for
"find" on working tree files.

This works reasonably well, except that it ends up listing the same
path multiple times if the index is unmerged.  Because the original
use of this construct was to name files to run etags over, and the
etags command happily takes the same filename multiple times without
causing any harm, there was no problem (other than perhaps spending
slightly more cycles, but who cares how fast the TAGS file gets
updated).

We however recently added a similar call to "ls-files" to list *.h
files, instead of using "find", in 92b88eba ("Makefile: use `git
ls-files` to list header files, if possible", 2019-03-04).  In this
new use of "ls-files", the resulting list $(LIB_H) is used for,
among other things, generating the header files to run hdr-check
target, and the duplicate unfortunately becomes a true problem.  It
causes $(MAKE) to notice that there are multiple %.hco targets and
complain.

Let the resulting list consumed by $(sort), which deduplicates,
to fix this.
Helped-by: NEric Sunshine <sunshine@sunshineco.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In order to run `rebase -i` without forking `rebase--interactive` it
will be convenient to have all the code from rebase--interactive.c in
rebase.c. This is a straight forward copy of the code from
rebase--interactive.c, it will be simplified slightly in the next
commit.
Signed-off-by: NPhillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The `git serve` built-in was introduced in ed10cb95 (serve:
introduce git-serve, 2018-03-15) as a backend to serve Git protocol v2,
probably originally intended to be spawned by `git upload-pack`.

However, in the version that the protocol v2 patches made it into core
Git, `git upload-pack` calls the `serve()` function directly instead of
spawning `git serve`; The only reason in life for `git serve` to survive
as a built-in command is to provide a way to test the protocol v2
functionality.

Meaning that it does not even have to be a built-in that is installed
with end-user facing Git installations, but it can be a test helper
instead.

Let's make it so.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In the recent years, there has been a big push to convert more and more
of Git's commands that are implemented as scripts to built-ins written
in pure, portable C, for robustness, speed and portability.

One strategy that served us well is to convert those scripts
incrementally, starting by renaming the scripts to
`git-legacy-<command>`, then introducing a built-in that does nothing
else at first than checking the config setting `<command>.useBuiltin`
(which defaults to `false` at the outset) and handing off to the legacy
script if so asked.

Obviously, those `git-legacy-<command>` commands share the documentation
with the built-in `git-<command>`, and are not intended to be called
directly anyway. So let's not try to ensure that they are documented
separately from their built-in versions.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When building with certain build options, some commands are excluded
from the build. For example, `git-credential-cache` is skipped when
building with `NO_UNIX_SOCKETS`.

Let's not build or package documentation for those excluded commands.

This issue was pointed out rightfully when running `make check-docs` on
Windows, where we do not yet have Unix sockets, and therefore the
`credential-cache` command is excluded (yet its documentation was built
and shipped).

Note: building the documentation via `make -C Documentation` leaves the
build system with no way to determine which commands have been
excluded. If called thusly, we gracefully fail to exclude their
documentation. Only when building the documentation via the top-level
Makefile will it get excluded properly, or after building
`Documentation/GIT-EXCLUDED-PROGRAMS` manually.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Among other things, the `check-docs` target ensures that
`command-list.txt` no longer contains commands that were dropped (or
that were never added in the first place).

To do so, it compares the list of commands from that file to the
commands listed in `$(ALL_COMMANDS)`.

However, some build options exclude commands from the latter. Fix the
target to handle this situation correctly by taking the just-introduced
`$(EXCLUDED_PROGRAMS)` into account.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When built with NO_CURL or with NO_UNIX_SOCKETS, some commands are
skipped from the build. It does not make sense to list them in the
output of `git help -a`, so let's just not.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The last user was just removed; There is no longer any need to carry it
around. Should we ever run into a need for it again, it is easy enough
to revert this commit.

It is unlikely, though, that we need `NO_INSTALL` again: as we saw with
the just-removed item, `git-remote-testgit`, we have better locations
to put executables and scripts that we do not want to install, e.g.
a subdirectory in `t/`, or `contrib/`.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Teach git to read the system and global config files for
default Trace2 settings.  This allows system-wide Trace2 settings to
be installed and inherited to make it easier to manage a collection of
systems.

The original GIT_TR2* environment variables are loaded afterwards and
can be used to override the system settings.

Only the system and global config files are used.  Repo and worktree
local config files are ignored.  Likewise, the "-c" command line
arguments are also ignored.  These limits are for performance reasons.

(1) For users not using Trace2, there should be minimal overhead to
detect that Trace2 is not enabled.  In particular, Trace2 should not
allocate lots of otherwise unused data strucutres.

(2) For accurate performance measurements, Trace2 should be initialized
as early in the git process as possible, and before most of the normal
git process initialization (which involves discovering the .git directory
and reading a hierarchy of config files).
Signed-off-by: NJeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The `git-remote-testgit` script is really only used in
`t5801-remote-helpers.sh`. It does not even contain any `@@<MAGIC>@@`
placeholders that would need to be interpolated via `make
git-remote-testgit`.

Let's just move it to a new home, decluttering the top-level directory
and clarifying that this is just a test helper, not an official Git
command that we would want to ever support.

Suggested by Ævar Arnfjörð Bjarmason.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

On Windows, for example, executables (must) have the extension `.exe`.
Our `check-docs` target was not prepared for that.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When we want to see what commands are listed in `command-list.txt` but
not installed, we currently include lines that refer to guides, e.g.
`gitattributes` or `gitcli`.

Let's not include those lines, as they are not referring to commands.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

As part of the `check-docs` target, we verify that commands that are
documented are actually in the current list of commands to be built.

However, this logic broke in 5fafce0b (check-docs: get documented
command list from Makefile, 2012-08-08), when we tried to make the logic
safer by not looking at the files in the worktree, but at the list of
files to be generated in `Documentation/Makefile`. While this was the
right thing to do, it failed to accommodate for the fact that `make -C
Documentation/ print-man1`, unlike `ls Documentation/*.txt`, would *not*
print lines starting with the prefix `Documentation/`.

At long last, let's fix this.

Note: This went undetected due to a funny side effect of the
`ALL_PROGRAMS` variable starting with a space. That space, together with
the extra space we inserted before `$(ALL_PROGRAMS)` in the

	case " $(ALL_PROGRAMS)" in
	*" $$cmd ")
		[...]

construct, is responsible that this case arm is used when `cmd` is empty
(which was clearly not intended to be the case).
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Remove the rebase.useBuiltin setting, which was added as an escape
hatch to disable the builtin version of rebase first released with Git
2.20.

See [1] for the initial implementation of rebase.useBuiltin, and [2]
and [3] for the documentation and corresponding
GIT_TEST_REBASE_USE_BUILTIN option.

Carrying the legacy version is a maintenance burden as seen in
7e097e27 ("legacy-rebase: backport -C<n> and --whitespace=<option>
checks", 2018-11-20) and 9aea5e92 ("rebase: fix regression in
rebase.useBuiltin=false test mode", 2019-02-13). Since the built-in
version has been shown to be stable enough let's remove the legacy
version.

As noted in [3] having use_builtin_rebase() shell out to get its
config doesn't make any sense anymore, that was done for the purposes
of spawning the legacy rebase without having modified any global
state. Let's instead handle this case in rebase_config().

There's still a bunch of references to git-legacy-rebase in po/*.po,
but those will be dealt with in time by the i18n effort.

Even though this configuration variable only existed two releases
let's not entirely delete the entry from the docs, but note its
absence. Individual versions of git tend to be around for a while due
to distro packaging timelines, so e.g. if we're "lucky" a given
version like 2.21 might be installed on say OSX for half a decade.

That'll mean some people probably setting this in config, and then
when they later wonder if it's needed they can Google search the
config option name or check it in git-config. It also allows us to
refer to the docs from the warning for details.

1. 55071ea2 ("rebase: start implementing it as a builtin",
   2018-08-07)
2. d8d0a546 ("rebase doc: document rebase.useBuiltin", 2018-11-14)
3. 62c23938 ("tests: add a special setup where rebase.useBuiltin is
   off", 2018-11-14)
3. https://public-inbox.org/git/nycvar.QRO.7.76.6.1903141544110.41@tvgsbejvaqbjf.bet/Acked-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The sha1dc library uses unaligned loads on platforms that support them.
This is normally what you'd want for performance, but it does cause
UBSan to complain when we compile with SANITIZE=undefined. Just like we
set -DNO_UNALIGNED_LOADS for our own code in that case, we should set
-DSHA1DC_FORCE_ALIGNED_ACCESS.

Of course that does nothing without pulling in the patches from sha1dc
to respect that define. So let's do that, too, updating both the
submodule link and our in-tree copy (from the same commit).
Signed-off-by: NJeff King <peff@peff.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

We recently converted the `git stash` command from Unix shell scripts
to builtins.

Let's end users a way out when they discover a bug in the
builtin command: `stash.useBuiltin`.

As the file name `git-stash` is already in use, let's rename the
scripted backend to `git-legacy-stash`.

To make the test suite pass with `stash.useBuiltin=false`, this commit
also backports rudimentary support for `-q` (but only *just* enough
to appease the test suite), and adds a super-ugly hack to force exit
code 129 for `git stash -h`.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NThomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The old shell script `git-stash.sh`  was removed and replaced
entirely by `builtin/stash.c`. In order to do that, `create` and
`push` were adapted to work without `stash.sh`. For example, before
this commit, `git stash create` called `git stash--helper create
--message "$*"`. If it called `git stash--helper create "$@"`, then
some of these changes wouldn't have been necessary.

This commit also removes the word `helper` since now stash is
called directly and not by a shell script.
Signed-off-by: NPaul-Sebastian Ungureanu <ungureanupaulsebastian@gmail.com>
Signed-off-by: NThomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Add a builtin helper for performing stash commands. Converting
all at once proved hard to review, so starting with just apply
lets conversion get started without the other commands being
finished.

The helper is being implemented as a drop in replacement for
stash so that when it is complete it can simply be renamed and
the shell script deleted.

Delete the contents of the apply_stash shell function and replace
it with a call to stash--helper apply until pop is also
converted.
Signed-off-by: NJoel Teichroeb <joel@teichroeb.net>
Signed-off-by: NPaul-Sebastian Ungureanu <ungureanupaulsebastian@gmail.com>
Signed-off-by: NThomas Gummerer <t.gummerer@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

If the GCRYPT_SHA256 build variable is not set, then the 'hdr-check'
target complains about the missing <gcrypt.h> header file. Add the
'sha256/gcrypt.h' header file to the exception list, if the build
variable is not defined. While here, replace the 'xdiff%' filter
pattern with 'xdiff/%' (and similarly for the compat pattern) since
the original pattern inadvertently excluded the 'xdiff-interface.h'
header.
Signed-off-by: NRamsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In d85b0dff (Makefile: use `find` to determine static header
dependencies, 2014-08-25), we switched from a static list of header
files to a dynamically-generated one, asking `find` to enumerate them.

Back in those days, we did not use `$(LIB_H)` by default, and many a
`make` implementation seems smart enough not to run that `find` command
in that case, so it was deemed okay to run `find` for special targets
requiring this macro.

However, as of ebb7baf0 (Makefile: add a hdr-check target,
2018-09-19), $(LIB_H) is part of a global rule and therefore must be
expanded. Meaning: this `find` command has to be run upon every
`make` invocation. In the presence of many a worktree, this can tax the
developers' patience quite a bit.

Even in the absence of worktrees or other untracked files and
directories, the cost of I/O to generate that list of header files is
simply a lot larger than a simple `git ls-files` call.

Therefore, just like in 33533975 (Makefile: ask "ls-files" to list
source files if available, 2011-10-18), we now prefer to use `git
ls-files` to enumerate the header files to enumerating them via `find`,
falling back to the latter if the former failed (which would be the case
e.g. in a worktree that was extracted from a source .tar file rather
than from a clone of Git's sources).

This has one notable consequence: we no longer include `command-list.h`
in `LIB_H`, as it is a generated file, not a tracked one, but that is
easily worked around. Of the three sites that use `LIB_H`, two
(`LOCALIZED_C` and `CHK_HDRS`) already handle generated headers
separately. In the third, the computed-dependency fallback, we can just
add in a reference to $(GENERATED_H).

Likewise, we no longer include not-yet-tracked header files in `LIB_H`.

Given the speed improvements, these consequences seem a comparably small
price to pay.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Acked-by: NRamsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Ever since the DEVELOPER=1 facility introduced there's been no way to
have custom CFLAGS (e.g. CFLAGS="-O0 -g -ggdb3") while still
benefiting from the set of warnings and assertions DEVELOPER=1
enables.

This is because the semantics of variables in the Makefile are such
that the user setting CFLAGS overrides anything we set, including what
we're doing in config.mak.dev[1].

So let's introduce a "DEVELOPER_CFLAGS" variable in config.mak.dev and
add it to ALL_CFLAGS. Before this the ALL_CFLAGS variable
would (basically, there's some nuance we won't go into) be set to:

    $(CPPFLAGS) [$(CFLAGS) *or* $(CFLAGS) in config.mak.dev] $(BASIC_CFLAGS) $(EXTRA_CPPFLAGS)

But will now be:

    $(DEVELOPER_CFLAGS) $(CPPFLAGS) $(CFLAGS) $(BASIC_CFLAGS) $(EXTRA_CPPFLAGS)

The reason for putting DEVELOPER_CFLAGS first is to allow for
selectively overriding something DEVELOPER=1 brings in. On both GCC
and Clang later settings override earlier ones. E.g. "-Wextra
-Wno-extra" will enable no "extra" warnings, but not if those two
arguments are reversed.

Examples of things that weren't possible before, but are now:

    # Use -O0 instead of -O2 for less painful debuggng
    DEVELOPER=1 CFLAGS="-O0 -g"
    # DEVELOPER=1 plus -Wextra, but disable some of the warnings
    DEVELOPER=1 DEVOPTS="no-error extra-all" CFLAGS="-O0 -g -Wno-unused-parameter"

The reason for the patches leading up to this one re-arranged the
various *FLAGS assignments and includes is just for readability. The
Makefile supports assignments out of order, e.g.:

    $ cat Makefile
    X = $(A) $(B) $(C)
    A = A
    B = B
    include c.mak
    all:
            @echo $(X)
    $ cat c.mak
    C=C
    $ make
    A B C

So we could have gotten away with the much smaller change of changing
"CFLAGS" in config.mak.dev to "DEVELOPER_CFLAGS" and adding that to
ALL_CFLAGS earlier in the Makefile "before" the config.mak.*
includes. But I think it's more readable to use variables "after"
they're included.

1. https://www.gnu.org/software/make/manual/html_node/Overriding.htmlSigned-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Move the setting of variables like CFLAGS down past settings like
"prefix" and default programs like "TAR" to just before we do the
include from "config.mak.*".

There's no functional changes here yet, but move note that
"ALL_CFLAGS" and "ALL_LDFLAGS" are moved below the include. A
follow-up change will tweak those depending on a variable set in
config.mak.dev.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Now the only other non-program assignment in the previous list is
PTHREAD_CFLAGS, which'll be moved elsewhere in a follow-up change.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The top of the Makfile is mostly separated into logical steps like set
default configuration, set programs etc., but there's some deviation
from that.

Let's add mostly comments where they're missing, remove those that
don't add anything. The whitespace tweaking makes subsequent patches
smaller.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Move the assignment of the "STRIP" variable down to where we're
setting variables with the names of other programs.

For consistency with those use "=" for the assignment instead of
"?=". I can't imagine why this would need to be different than the
rest, and 4dc00021 ("Makefile: add 'strip' target", 2006-01-12)
which added it doesn't provide an explanation.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Remove a comment referring to a caveat that hasn't been applicable
since 18b0fc1c ("Git.pm: Kill Git.xs for now", 2006-09-23).

At the time of 8d7f586f ("Git.pm: Support for perl/ being built by a
different compiler", 2006-06-25) some of the code in perl would be
built by a C compiler, but support for that went away a few months
later in 18b0fc1c discussed above.

Since my 20d2a30f ("Makefile: replace perl/Makefile.PL with simple
make rules", 2017-12-10) the perl/ directory doesn't even have its own
build process.
Signed-off-by: NÆvar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Create unit tests for Trace2.
Signed-off-by: NJeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Create a new unified tracing facility for git.  The eventual intent is to
replace the current trace_printf* and trace_performance* routines with a
unified set of git_trace2* routines.

In addition to the usual printf-style API, trace2 provides higer-level
event verbs with fixed-fields allowing structured data to be written.
This makes post-processing and analysis easier for external tools.

Trace2 defines 3 output targets.  These are set using the environment
variables "GIT_TR2", "GIT_TR2_PERF", and "GIT_TR2_EVENT".  These may be
set to "1" or to an absolute pathname (just like the current GIT_TRACE).

* GIT_TR2 is intended to be a replacement for GIT_TRACE and logs command
  summary data.

* GIT_TR2_PERF is intended as a replacement for GIT_TRACE_PERFORMANCE.
  It extends the output with columns for the command process, thread,
  repo, absolute and relative elapsed times.  It reports events for
  child process start/stop, thread start/stop, and per-thread function
  nesting.

* GIT_TR2_EVENT is a new structured format. It writes event data as a
  series of JSON records.

Calls to trace2 functions log to any of the 3 output targets enabled
without the need to call different trace_printf* or trace_performance*
routines.
Signed-off-by: NJeff Hostetler <jeffhost@microsoft.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In cc95bc20 (t5562: replace /dev/zero with a pipe from
generate_zero_bytes, 2019-02-09), we replaced usage of /dev/zero (which
is not available on NonStop, apparently) by a Perl script snippet to
generate NUL bytes.

Sadly, it does not seem to work on NonStop, as t5562 reportedly hangs.

Worse, this also hangs in the Ubuntu 16.04 agents of the CI builds on
Azure Pipelines: for some reason, the Perl script snippet that is run
via `generate_zero_bytes` in t5562's 'CONTENT_LENGTH overflow ssite_t'
test case tries to write out an infinite amount of NUL bytes unless a
broken pipe is encountered, that snippet never encounters the broken
pipe, and keeps going until the build times out.

Oddly enough, this does not reproduce on the Windows and macOS agents,
nor in a local Ubuntu 18.04.

This developer tried for a day to figure out the exact circumstances
under which this hang happens, to no avail, the details remain a
mystery.

In the end, though, what counts is that this here change incidentally
fixes that hang (maybe also on NonStop?). Even more positively, it gets
rid of yet another unnecessary Perl invocation.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

On various BSD's, fileno(fp) is implemented as a macro that directly
accesses the fields in the FILE * object, which breaks a function that
accepts a "void *fp" parameter and calls fileno(fp) and expect it to
work.

Work it around by adding a compile-time knob FILENO_IS_A_MACRO that
inserts a real helper function in the middle of the callchain.
Signed-off-by: NNguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When serializing UTF-16 (and UTF-32), there are three possible ways to
write the stream. One can write the data with a BOM in either big-endian
or little-endian format, or one can write the data without a BOM in
big-endian format.

Most systems' iconv implementations choose to write it with a BOM in
some endianness, since this is the most foolproof, and it is resistant
to misinterpretation on Windows, where UTF-16 and the little-endian
serialization are very common. For compatibility with Windows and to
avoid accidental misuse there, Git always wants to write UTF-16 with a
BOM, and will refuse to read UTF-16 without it.

However, musl's iconv implementation writes UTF-16 without a BOM,
relying on the user to interpret it as big-endian. This causes t0028 and
the related functionality to fail, since Git won't read the file without
a BOM.

Add a Makefile and #define knob, ICONV_OMITS_BOM, that can be set if the
iconv implementation has this behavior. When set, Git will write a BOM
manually for UTF-16 and UTF-32 and then force the data to be written in
UTF-16BE or UTF-32BE. We choose big-endian behavior here because the
tests use the raw "UTF-16" encoding, which will be big-endian when the
implementation requires this knob to be set.

Update the tests to detect this case and write test data with an added
BOM if necessary. Always write the BOM in the tests in big-endian
format, since all iconv implementations that omit a BOM must use
big-endian serialization according to the Unicode standard.

Preserve the existing behavior for systems which do not have this knob
enabled, since they may use optimized implementations, including
defaulting to the native endianness, which may improve performance.
Signed-off-by: Nbrian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In order to enable greater user customisation of the SPARSE_FLAGS
variable, we introduce a new SP_EXTRA_FLAGS variable to use for
target specific settings. Without using the new variable, setting
the SPARSE_FLAGS on the 'make' command-line would also override the
value set by the target-specific rules in the Makefile (effectively
making them useless). Also, this enables the SP_EXTRA_FLAGS to be
used in the future for any other internal customisations, such as
for some platform specific values.

In addition, we initialise the SPARSE_FLAGS to the default (empty)
value using a conditional assignment (?=). This allows SPARSE_FLAGS
to be set from the environment as well as from the command-line.
Signed-off-by: NRamsay Jones <ramsay@ramsayjones.plus.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Sometimes there are test failures in the 'pu' branch. This
is somewhat expected for a branch that takes the very latest
topics under development, and those sometimes have semantic
conflicts that only show up during test runs. This also can
happen when running the test suite with different GIT_TEST_*
environment variables that interact in unexpected ways

This causes a problem for the test coverage reports, as
the typical 'make coverage-test coverage-report' run halts
at the first failed test. If that test is early in the
suite, then many valuable tests are not exercising the code
and the coverage report becomes noisy with false positives.

Add a new 'coverage-prove' target to the Makefile,
modeled after the 'coverage-test' target. This compiles
the source using the coverage flags, then runs the test
suite using the 'prove' tool. Since the coverage
machinery is not thread-safe, enforce that the tests
are run in sequence by appending '-j1' to GIT_PROVE_OPTS.
Signed-off-by: NDerrick Stolee <dstolee@microsoft.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The fact that Git's test suite is implemented in Unix shell script that
is as portable as we can muster, combined with the fact that Unix shell
scripting is foreign to Windows (and therefore has to be emulated),
results in pretty abysmal speed of the test suite on that platform, for
pretty much no other reason than that language choice.

For comparison: while the Linux build & test is typically done within
about 8 minutes, the Windows build & test typically lasts about 80
minutes in Azure Pipelines.

To help with that, let's use the Azure Pipeline feature where you can
parallelize jobs, make jobs depend on each other, and pass artifacts
between them.

The tests are distributed using the following heuristic: listing all
test scripts ordered by size in descending order (as a cheap way to
estimate the overall run time), every Nth script is run (where N is the
total number of parallel jobs), starting at the index corresponding to
the parallel job. This slicing is performed by a new function that is
added to the `test-tool`.

To optimize the overall runtime of the entire Pipeline, we need to move
the Windows jobs to the beginning (otherwise there would be a very
decent chance for the Pipeline to be run only the Windows build, while
all the parallel Windows test jobs wait for this single one).

We use Azure Pipelines Artifacts for both the minimal Git for Windows
SDK as well as the built executables, as deduplication and caching close
to the agents makes that really fast. For comparison: while downloading
and unpacking the minimal Git for Windows SDK via PowerShell takes only
one minute (down from anywhere between 2.5 to 7 when using a shallow
clone), uploading it as Pipeline Artifact takes less than 30s and
downloading and unpacking less than 20s (sometimes even as little as
only twelve seconds).
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

This will come in handy when publishing the results of Git's test suite
during an automated Azure DevOps run.

Note: we need to make extra sure that invalid UTF-8 encoding is turned
into valid UTF-8 (using the Replacement Character, \uFFFD) because
t9902's trace contains such invalid byte sequences, and the task in the
Azure Pipeline that uploads the test results would refuse to do anything
if it was asked to parse an .xml file with invalid UTF-8 in it.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The comment explaining how to build the fuzzers was broken in
927c77e7 ("Makefile: use FUZZ_CXXFLAGS for linking fuzzers",
2018-11-14).

When building fuzzers, all .c files must be compiled with coverage
tracing enabled. This is not possible when using only FUZZ_CXXFLAGS, as
that flag is only applied to the fuzzers themselves. Switching back to
CFLAGS fixes the issue.
Signed-off-by: NJosh Steadmon <steadmon@google.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Break load_commit_graph_one() into a new function, parse_commit_graph().
The latter function operates on arbitrary buffers, which makes it
suitable as a fuzzing target. Since parse_commit_graph() is only called
by load_commit_graph_one() (and the fuzzer described below), we omit
error messages that would be duplicated by the caller.

Adds fuzz-commit-graph.c, which provides a fuzzing entry point
compatible with libFuzzer (and possibly other fuzzing engines).
Signed-off-by: NJosh Steadmon <steadmon@google.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

As part of an ongoing effort to make rebase have more uniform behavior,
modify the merge backend to behave like the interactive one, by
re-implementing it on top of the latter.

Interactive rebases are implemented in terms of cherry-pick rather than
the merge-recursive builtin, but cherry-pick also calls into the
recursive merge machinery by default and can accept special merge
strategies and/or special strategy options.  As such, there really is
not any need for having both git-rebase--merge and
git-rebase--interactive anymore.  Delete git-rebase--merge.sh and
instead implement it in builtin/rebase.c.

This results in a few deliberate but small user-visible changes:
  * The progress output is modified (see t3406 and t3420 for examples)
  * A few known test failures are now fixed (see t3421)
  * bash-prompt during a rebase --merge is now REBASE-i instead of
    REBASE-m.  Reason: The prompt is a reflection of the backend in use;
    this allows users to report an issue to the git mailing list with
    the appropriate backend information, and allows advanced users to
    know where to search for relevant control files.  (see t9903)

testcase modification notes:
  t3406: --interactive and --merge had slightly different progress output
         while running; adjust a test to match the new expectation
  t3420: these test precise output while running, but rebase--am,
         rebase--merge, and rebase--interactive all were built on very
         different commands (am, merge-recursive, cherry-pick), so the
         tests expected different output for each type.  Now we expect
         --merge and --interactive to have the same output.
  t3421: --interactive fixes some bugs in --merge!  Wahoo!
  t9903: --merge uses the interactive backend so the prompt expected is
         now REBASE-i.
Signed-off-by: NElijah Newren <newren@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

OSS-Fuzz requires C++-specific flags to link fuzzers. Passing these in
CFLAGS causes lots of build warnings. Using separate FUZZ_CXXFLAGS
avoids this.
Signed-off-by: NJosh Steadmon <steadmon@google.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>