When growing buffers, we repeatedly multiply the currently allocated
number of bytes by 1.5 until it exceeds the requested number of bytes.
This has two major problems:

    1. If the current number of bytes is tiny and one wishes to resize
       to a comparatively huge number of bytes, then we may need to loop
       thousands of times.

    2. If resizing to a value close to `SIZE_MAX` (which would fail
       anyway), then we probably hit an infinite loop as multiplying the
       current amount of bytes will repeatedly result in integer
       overflows.

When reallocating buffers, one typically chooses values close to 1.5 to
enable re-use of resulting memory holes in later reallocations. But
because of this, it really only makes sense to use a factor of 1.5
_once_, but not looping until we finally are able to fit it. Thus, we
can completely avoid the loop and just opt for the much simpler
algorithm of multiplying with 1.5 once and, if the result doesn't fit,
just use the target size. This avoids both problems of looping
extensively and hitting overflows.

This commit also adds a test that would've previously resulted in an
infinite loop.

If growing a buffer fails, we set its pointer to the static
`git_buf__oom` structure. While we correctly free the old pointer if
`git__malloc` returned an error, we do not free it if there was an
integer overflow while calculating the new allocation size. Fix this
issue by freeing the pointer to plug the memory leak.

Circular header splitting

azure: build Docker images as part of the pipeline

smart: use push_glob instead of manual filtering

The MESSAGE() function expects as first argument the message type, e.g.
STATUS or FATAL_ERROR. In some places, we were misusing this to either
not provide any type, which would then erroneously print the message to
standard error, or to use FATAL instead of FATAL_ERROR.

Fix all of these instances. Also, remove some MESSAGE invocations that
are obvious leftovers from debugging the build system.

When using mbedTLS as backend, then the user may specify the location of
where system certificates are installed. If no such location is provided
by the user, CMake will try to autodetect the location by using the
openssl executable, if installed. If no location could be detected, then
the mbedTLS is essentially worthless as it is completely unable to
verify any certificates.

To avoid use of such misconfigured mbedTLS configurations, let's error
out if we were unable to find out the location.

While we were still supporting Trusty, using Ninja as a build tool would
have required us to first setup pip and then use it to install Ninja.
As a result, the speedups from using Ninja were drowned out by the
time required to install Ninja. But as we have deprecated Trusty now,
both Xenial and Bionic have recent versions of Ninja in their
repositories and thus we can now use Ninja.

The Valgrind version shipped with Xenial has some bugs that keep our
tests from working due to bad interactions with openssl [1]. Fix this by
using the "hola-launchpad/valgrind" PPA that provides a newer version
for which the bug has been fixed.

[1]: https://bugs.launchpad.net/ubuntu/+source/valgrind/+bug/1574437

Pass the flag "--no-install-recommends" to apt-get in order to trim down
the number of packages installed, both reducing build time and image
size. As this also causes some required packages to not be installed
anymore, add these explicitly to the set of packages installed.

Reformat both Xenial and Bionic's Dockerfiles to use best practices.
Most importantly, we now run `apt-get update` and `apt-get install` in
one step followed up by removing the package lists to speed up
installation and keep down the image size.

While Xenial provides libssh2 in its repositories, it only has version
1.5.0 available. This version will unfortunately not be able to connect
to GitHub due to their removal of weak cryptographic standards [1]. To
still enable our CI to execute tests against GitHub, we thus have to
update the provided libssh2 version to a newer one.

Manually install libssh2 1.8.2 on Xenial. There's no need to do the same
for Bionic, as it already provides libssh2 1.8.0.

[1]: https://github.blog/2018-02-01-crypto-removal-notice/

We're about to phase out support for Trusty, but neither Bionic nor
Xenial images provide the mbedTLS library that's available in Trusty.
Build them for both to pull them in line with Trusty.

The Coverity build is still referencing an old "trusty-openssl"
container that is not provided by either our own now-inlined images nor
by the libgit2/libgit2-docker repository.

Convert it to build and use Xenial images instead.

Support for the LTS release Ubuntu 14.04 Trusty has been dropped in
April 2019, but Azure is still using Trusty as its primary platform to
build and test against. Let's deprecate it in favor of Xenial.

The Docker images used for our continuous integration builds currently
live in the libgit2/libgit2-docker repository. To make any changes in
them, one has to make a PR there, get it reviewed, re-build the images
and publish them to Docker Hub. This process is slow and tedious, making
it harder than necessary to perform any updates to our Docker-based
build pipeline.

To fix this, we include all Dockerfiles used by Azure from the mentioned
repository and inline them into our own repo. Instead of having to
manually push them to the CI, it will now build the required containers
on each pull request, allowing much greater flexibility.

ntlm: fix failure to find openssl headers

cmake: remove extraneous logging

open:fix memory leak when passing NULL to git_repository_open_ext

apply: Fix a patch corruption related to EOFNL handling

ignore: correct handling of nested rules overriding wild card unignore

Memory allocation fixes for diff generator

Use an HTTP scheme that supports the given credentials

Introduce an unit test to validate that git_apply__patch() properly
handles EOFNL changes in case of patches with several hunks.

problem:
filesystem_iterator loads .gitignore files in top-down order.
subsequently, ignore module evaluates them in the order they are loaded.
this creates a problem if we have unignored a rule (using a wild card)
in a sub dir and ignored it again in a level further below (see the test
included in this patch).

solution:
process ignores in reverse order.

closes #4963

apply: git_apply_to_tree fails to apply patches that add new files

Optionally read `.gitattributes` from HEAD

config: implement "onbranch" conditional

When allocating new tree iterator frames, we zero out the allocated
memory twice. Remove one of the `memset` calls.

When allocating tree iterator entries, we use GIT_ERROR_ALLOC_CHECK` to
check whether the allocation has failed. The macro will cause the
function to immediately return, though, leaving behind a partially
initialized iterator frame.

Fix the issue by manually checking for memory allocation errors and
using `goto done` in case of an error, popping the iterator frame.

When preparing options for the two iterators that are about to be
diffed, we allocate a common prefix for both iterators depending on
the options passed by the user. We do not check whether the allocation
was successful, though. In fact, this isn't much of a problem, as using
a `NULL` prefix is perfectly fine. But in the end, we probably want to
detect that the system doesn't have any memory left, as we're unlikely
to be able to continue afterwards anyway.

While the issue is being fixed in the newly created function
`diff_prepare_iterator_opts`, it has been previously existing in the
previous macro `DIFF_FROM_ITERATORS` already.

While the `DIFF_FROM_ITERATORS` does make it shorter to implement the
various `git_diff_foo_to_bar` functions, it is a complex and unreadable
beast that implicitly assumes certain local variable names. This is not
something desirable to have at all and obstructs understanding and more
importantly debugging the code by quite a bit.

The `DIFF_FROM_ITERATORS` macro basically removed the burden of having
to derive the options for both iterators from a pair of iterator flags
and the diff options. This patch introduces a new function that does the
that exact and refactors all callers to manage the iterators by
themselves.

As we potentially need to allocate a shared prefix for the
iterator, we need to tell the caller to allocate that prefix as soon as
the options aren't required anymore. Thus, the function has a `char
**prefix` out pointer that will get set to the allocated string and
subsequently be free'd by the caller.

While this patch increases the line count, I personally deem this to an
acceptable tradeoff for increased readbiblity.