The block of code can be removed in the future if CMake 3.15 ever
becomes the minimum

Move pugixml-config.cmake.in to scripts/ to prevent polluting the root

New Features:

pugixml-shared and pugixml-static are "always" available, but not always
built. This allows downstream projects to still use them, and to make
sure our configurations are correct regardless of if they are being
built. They are not always installed however.

pugixml-shared and pugixml-static have ALIAS libraries of
pugixml::shared and pugixml::static respectively. These names are also
respected in the `find_package(pugixml CONFIG)` generated files, so its
safe to use pugixml in a CMake project regardless of whether it is
installed locally, or if its a subproject via `add_subdirectory`.

pugixml will automatically select the correct library type based on
BUILD_SHARED_LIBS. A pugixml::pugixml ALIAS is also available.

CMAKE_MSVC_RUNTIME_LIBRARY from CMake 3.15 has been backported.

CMake will now rely on generator expressions (a powerful abstraction in
large projects with many subdirectories) for most work. This offloads
work from the single-threaded configure stage to the multithreaded
generation stage.

pugixml now uses CTest as the runner.

Some settings are automatically disabled if pugixml is used as a
subdirectory. These are still able to be manually set, but are hidden
from folks who choose to use CMake GUI.

 Bump CMake minimum to 3.4
 pugixml no longer requires a C compiler to be found or set to
compile correctly. This speeds up configuration and building on windows.
 Begin laying groundwork to backport MSVC_RUNTIME_LIBRARY
property

Fixes #285 

This keeps src/ folder clean of auxiliary files only required for
special builds; note that CMakeLists.txt already depends on scripts/
(specifically for pkgconfig setup).

Closes #294 

The newly added tests make sure that during node/attribute destruction
we deallocate a few memory pages; this makes sure that we don't read
node data after it's being destroyed.

Also clean up formatting/style in the remove_* implementation a bit.

These functions remove all attributes / child nodes in bulk which is faster than removing them one at a time.

Migrate from licenseUrl to license since licenseUrl is deprecated and
fix grammar for VS version list.

Update changelog and regenerate HTML documentation.

According to XML spec, > sometimes needs to be escaped in PCDATA (when
it occurs as a ]]> pattern), but it doesn't need to be escaped in
attribute values.

Contributes to #272.

When using double quotes for attributes, we don't need to escape '; when
using single quotes, we don't need to escape ".

This changes behavior to match 1.9 by default (where we don't escape ').

Contributes to #272.

Note: this chang also updates PUGIXML_VERSION macro to allow for
double-digit minor versions; this preserves the continuity of versions
so PUGIXML_VERSION >= 190 will still work.

Create visual studio projects that are vs2019 compliant.
* nuget_build.ps1 :
Introduce a new argument that will define how we implement the nuget
build. For now we accept 201{9,7.5.3} as possible argument values.

* pugixml_vs2019{,_static}.vcxproj :
Add two visual studio projects that build pugi with the latest SDK and
build tools

* appveyor.yml
- Add Visual Studio 2019 to build targets
- Add Visual Studio 201{9,3,5} to build_scripts. And call
  nuget_build.ps1 with a new argument.
- Add Visual Studio 2019 to the test_scripts.

- Typos were found by codespell v1.15.0

Self-assignment generates a Wself-assign-overloaded warning.

appending the suffix to the build product need not be a function of
whether CMAKE_CONFIGURATION_TYPES is set.  for example, having two ninja
build trees---one for debug and another for release---is a fine use-case
for USE_POSTFIX.

This change adds format_attribute_single_quote flag that uses single quotes (`'`) instead of double quotes (`"`) for formatting attribute values.

Internal quotation marks are escaped using `"` and `'`.

std::random_shuffle is removed in current standard

use <random> 's generator and std::shuffle instead

Also split one-liner if/for loops for better readability and coverage.

Remove contrib folder

This change removes contrib folder since it doesn't seem very useful;
Natvis scripts can and should go to scripts/ along with project files.

foreach.hpp is supporting a severely outdated BOOST_FOREACH construct;
in C++11 ranged for loop can be used with xml_node::children/attributes,
and BOOST_FOREACH can work with these as well.

foreach.hpp was also accidentally licensed as public domain which isn't
very helpful when the actual library is MIT. We could fix the license
but it seems better to just remove it.

Fixes #264.

Also fix code style.

Includes wording fixes and format_skip_control_chars

format_skip_control_chars

We now have two tests: one tests behavior when we run out of space when
appending the node set (in which case the append fails), another one
tests behavior when we run out of space when filtering the node set (in
which case the set still contains redundant data).

Given an unsorted sequence, remove_duplicates would sort it using the
pointer value of attributes/nodes and then remove consecutive
duplicates.

This was problematic because it meant that the result of XPath queries
was dependent on the memory allocation pattern. While it's technically
incorrect to rely on the order, this results in easy to miss bugs.

This is particularly common when XPath queries use union operators -
although we also will call remove_duplicates in other cases.

This change reworks the code to use a hash set instead, using the same
hash function we use for compact storage. To make sure it performs well,
we allocate enough buckets for count * 1.5 (assuming all elements are
unique); since each bucket is a single pointer unlike xpath_node which
is two pointers, we need somewhere between size * 0.75 and size * 1.5
temporary storage.

The resulting filtering is stable - we remove elements that we have seen
before but we don't change the order - and is actually significantly
faster than sorting was.

With a large union operation, before this change it took ~56 ms per 100
query invocations to remove duplicates, and after this change it takes
~20ms.

Fixes #254.

This test is very sensitive to the particular implementation of union
aggregation; for now lets disable this.

We need a more robust way to test union allocation failures.