When thees is an unreadable folder, we should still be able
to enumerate status.

Pass unconverted Unicode path data when iconv doesn't like it

Be more careful with user-supplied buffers

Some improvements to the cert checking

Configuration snapshotting

Add filter options and ALLOW_UNSAFE

It seems that with the various recent changes to reference updating
and reflog writing, that the thread safety of refdb updates has
been reduced (either that or it was never thread safe and the
window for error has increased).  Either way, this test is now
sometimes segfaulting which is no good, so let's disable the test
for now.  We don't really make any public promises about thread
safety for this type of operation, so I think this is acceptable,
at least in the short term.

Only on a filesystem that is composed/decomposed insensitive,
should be testing that a branch can be looked up by the opposite
form and still work correctly.

One of the test helpers provides a quick way for looking up a
boolean key.  But if the key way missing completely, the check
would actually raise an error.  Given the way we use this helper,
if the key is missing, this should just return false, I think.

indexer: avoid memory moves

When using Iconv to convert unicode data and iconv doesn't like
the source data (because it thinks that it's not actual UTF-8),
instead of stopping the operation, just use the unconverted data.
This will generally do the right thing on the filesystem, since
that is the source of the non-UTF-8 path data anyhow.

This adds some tests for creating and looking up branches with
messy Unicode names.  Also, this takes the helper function that
was previously internal to `git_repository_init` and makes it
into `git_path_does_fs_decompose_unicode` which is a useful in
tests to understand what the expected results should be.

Our vector does a move of the rest of the array when we remove an
item. Doing this repeatedly can be expensive, and we do this a lot in
the indexer. Instead, set the value to NULL and skip those entries.

perf reported around 30% of `index-pack` time was going into
memmove. With this change, that goes away and we spent most of the time
hashing and inflating data.

iconv debugging aids

This is quite close to running "git for-each-ref" except:

  1. It does not take any formatting or selection options at
     all.

  2. The output is not sorted.

I wrote it to look at debugging some issues with ref
iteration, but there's no reason it can't live on as an
example command.

The cmake module we provide is in the file FindIconv.cmake,
so we must match the case correctly. It happens to work in
practice because we only turn on ICONV on Darwin, and people
generally have case-insensitive filesystems there.

Note that we only need to update the package name here. The
package itself still sets the all-uppercase ICONV_FOUND
flag, so we continue to use uppercase in the rest of cmake.

This adds in missing calls to `git_buf_sanitize` and fixes a
number of places where `git_buf` APIs could inadvertently write
NUL terminator bytes into invalid buffers.  This also changes the
behavior of `git_buf_sanitize` to NUL terminate a buffer if it can
and of `git_buf_shorten` to do nothing if it can.

Adds tests of filtering code with zeroed (i.e. unsanitized) buffer
which was previously triggering a segfault.

Accessing the repository's config and immediately taking a snapshot of
it is a common operation, so let's provide a convenience function for
it.

Fix the issues in git_shutdown

Diff and status do not want core.safecrlf to actually raise an
error regardless of the setting, so this extends the filter API
with an additional options flags parameter and adds a flag so that
filters can be applied with GIT_FILTER_OPT_ALLOW_UNSAFE, indicating
that unsafe filter application should be downgraded from a failure
to a warning.

odb: ignore files in the objects dir

1) Call to git_shutdown results in setting git__n_shutdown_callbacks
to -1. Next call to git__on_shutdown results in ABW (Array Bound Write)
for array git__shutdown_callbacks. In the current Implementation,
git_atomic_dec is called git__n_shutdown_callbacks + 1 times. I have
modified it to a for loop so that it is more readable. It would not
set git__n_shutdown_callbacks to a negative number and reset the
elements of git__shutdown_callbacks to NULL.

2) In function git_sysdir_get, shutdown function is registered only if
git_sysdir__dirs_shutdown_set is set to 0. However, after this variable
is set to 1, it is never reset to 0. If git_sysdir_global_init is
called again from synchronized_threads_init it does not register
shutdown function for this subsystem.

The brace in the check for peel's return was surrounding the wrong
thing, which made 'error' be set to 1 when there was an error instead of
the error code.

We assume that everything under GIT_DIR/objects/ is a directory. This is
not necessarily the case if some process left a stray file in there.

Check beforehand if we do have a directory and ignore the entry
otherwise.

Reduce excessive OID calculation for diff and stat

commit: safer commit creation with reference update

Trying to find other issues where tests may not clean up quite
properly when they are through...

There are a few tests that set up a fake home directory and a
fake GLOBAL search path so that we can test things in global
ignore or attribute or config files.  This cleans up that code to
work more robustly even if there is a test failure.  This also
fixes some valgrind warnings where scanning search paths for
separators could end up doing a little bit of sketchy data access
when coming to the end of search list.

There were a couple of "init_opts()" functions a few more cases
of structure initialization that I somehow missed.

Use an unsigned int for the version and add a helper macro so the
code is simplified (and so the error message is a common string).