Removed the variable name in the defaulted function to make GCC happy.

…arnings. (jbeder#1012)

* Export YAML::detail::node::m_amount

The internal header node/detail/node.h is included by public headers;
YAML::detail::node is implemented in the header itself, and thus it gets
inlined... except for its static m_amount class member, which is
instantiated in the library only. Right now all the symbols of yaml-cpp
are exported (nothing is hidden), so the linker will find node::m_amount
in the yaml-cpp library.

As solution/workaround, explicitly export YAML::detail::node::m_amount.

* CMake: use GenerateExportHeader

Make use of the GenerateExportHeader CMake module to generate the dll.h
header with export macros.

While the produced dll.h is different, the result should be the same,
i.e. nothing changes for yaml-cpp or its users.

* CMake: hide all the symbols by default

Hide all the symbols that are not explicitly exported with YAML_CPP_API.
This way the ABI will be way smaller, and only actually exposing the
public classes/functions.

…der#1038)

This reverts commit da1c8d3. and fixes jbeder#1036.

…" (jbeder#1038)" (jbeder#1045)

…ative values (jbeder#1051)

GNUInstallDirs provided may be absolute paths, in which case appending
to the install prefix is not correct. We can instead use the provided
CMAKE_INSTALL_FULL_* variables, which are precomputed absolute paths.

https://cmake.org/cmake/help/latest/module/GNUInstallDirs.html

…beder#1064)

Partial revert of "Revert "Revert "Hide most of non-public symbols by default (jbeder#984)" (jbeder#1038)" (jbeder#1045)"

This reverts commit 0733aeb.

This replaces the old Travis CI badge that didn't mean anything.

Add copying of shared library to the output directory with a test binary.

[binary] removed using of non unsigned char as argument at
'std::isspace' function that was provokes undefined behavior.

[.github/workflows/build.yml] enabled run of test at the
'windows-latest' environment.

Also refactor the build action to use a matrix.

Properly resolves jbeder#1083, jbeder#996.

…" instead of "MSVC".

Fixes issue jbeder#1053.

Complies with YAML Standard [5.4](https://yaml.org/spec/1.2.2/#54-line-break-characters) [25] instead of matching `\r` only in combination with `\n`.

After configuring the file `yaml-cpp-config.cmake.in`, the result ends up with
empty variables.  (see also the discussion in jbeder#774).

Rework this file and the call to `configure_package_config_file` according the
cmake documentation
(https://cmake.org/cmake/help/v3.22/module/CMakePackageConfigHelpers.html?highlight=configure_package_config#command:configure_package_config_file)
to overcome this issue and allow a simple `find_package` after install.

As there was some discussion about the place where to install the
`yaml-cpp-config.cmake` file, e.g. jbeder#1055, factor out the install location into
an extra variable to make it easier changing this location in the future.

Also untabify CMakeLists.txt in some places to align with the other code parts in this file.

…endencies also. (jbeder#1039)

The option `YAML_CPP_BUILD_TESTS` currently enables or disables building of tests; but unconditionally the CMake file includes CTest; this PR makes that conditional on the option.

Also, there is no option for enabling formatting, but it does check whether it can find the `clang-format` executable; this PR adds an option (default to true) that skips even looking for the executable if disabled.

Changes YAML_CPP_INSTALL from a cmake_dependent_option to an option.

Fixes jbeder#756, jbeder#847, and jbeder#1011.

Windows builds need some defines being set to use the static linking.

Also add bazel builds and test to Github-CI.

)

On Oracle Solaris the following statement is declared in file /usr/include/sys/regset.h:
#define SS 18 /* only stored on a privilege transition */

Because of this template type name SS is substituted by numeric literal, which results in a compilation error:
error: expected nested-name-specifier before numeric constant

Fixed by renaming template type names.

Co-authored-by: Roman Degtyar <[email protected]>

…er#1055)

Minimize warnings when not the top-level project

Should fix jbeder#970 and jbeder#764 when trying to add yaml-cpp to other project

Add definitions to static yaml-cpp target (fixes jbeder#1116)

Also updates CONTRIBUTING instructions for tests

This pull request adds two new exported variables to the CMake config:

YAML_CPP_LIBRARY_DIR - points to the directory containing the built library files
YAML_CPP_SHARED_LIBS_BUILT - boolean value, lets users know whether shared libraries were built or not

It secures the repo against erroneous or malicious actions from external jobs you call from your workflow. It's specially important for the case they get compromised, for example.

* Test(CMake) set NAME and COMMAND in add_test

* (CMake) add enable_testing()

* (CMake) move cmake_dependent_option up, before using them

* (CMake) use YAML_CPP_MAIN_PROJECT in cmake_dependent_option

* (CMake) log values regarding tests

* (CMake) always find CTest, but don't enable tests

* (CMAKE)(temp) fix logging

* (actions) set YAML_CPP_BUILD_TESTS for tests

* (actions) provide YAML_CPP_BUILD_TESTS to ctest

* (actions) set -DYAML_CPP_BUILD_TESTS at build

* (actions) don't fail false

* (actions) build tests in Test step

* (actions) run tests verbose

* (CMake) remove temp logging

* (actions) split building from running tests

* (actions) ctest Debug

* (actions) ctest Debug

* Remove enable_testing

This reverts commit 74f63c1.

This matches the name for the ALIAS targets, and fixes jbeder#1025.

cd /<<PKGBUILDDIR>>/build-static/test && /usr/bin/c++ -DYAML_CPP_STATIC_DEFINE -I/<<PKGBUILDDIR>>/test/integration -I/<<PKGBUILDDIR>>/test -I/<<PKGBUILDDIR>>/src -I/<<PKGBUILDDIR>>/include -isystem /usr/src/googletest/googlemock/include -isystem /usr/src/googletest/googlemock -isystem /usr/src/googletest/googletest/include -isystem /usr/src/googletest/googletest -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=. -flto=auto -ffat-lto-objects -fstack-protector-strong -Wformat -Werror=format-security -fdebug-prefix-map=/<<PKGBUILDDIR>>=/usr/src/yaml-cpp-0.8.0+dfsg-1~build1 -Wdate-time -D_FORTIFY_SOURCE=2 -std=gnu++14 -Wno-variadic-macros -Wno-sign-compare -DGTEST_HAS_PTHREAD=1 -MD -MT test/CMakeFiles/yaml-cpp-tests.dir/binary_test.cpp.o -MF CMakeFiles/yaml-cpp-tests.dir/binary_test.cpp.o.d -o CMakeFiles/yaml-cpp-tests.dir/binary_test.cpp.o -c /<<PKGBUILDDIR>>/test/binary_test.cpp
/<<PKGBUILDDIR>>/test/binary_test.cpp: In member function ‘virtual void BinaryTest_DecodingNoCrashOnNegative_Test::TestBody()’:
/<<PKGBUILDDIR>>/test/binary_test.cpp:11:38: error: narrowing conversion of ‘-58’ from ‘int’ to ‘char’ [-Wnarrowing]
   11 |   std::string input{-58, -1, -99, 109};
      |                                      ^

…beder#1225)

This reverts commit 6262201.

in 6262201, we wanted address the needs to use the `string_view`
converter in C++98, but that requirement was based on wrong
preconditions. `std::string_view` was introduced in C++17, and
popular standard libraries like libstdc++ and libc++ both provide
`std::string_view` when the source is built with C++17.

furthermore 6262201 is buggy. because it uses `<version>` to tell
the feature set provided by the standard library. but `<version>`
is a part of C++20. so this defeats the purpose of the change of
6262201.

Fixes jbeder#1223

* CI: test with supported C++ standards

let's test with the supported standards for better coverage.
C++23 standard was just out, but the support on the toolchain is
not quite ready yet. so let's leave it for a future change.

Signed-off-by: Kefu Chai <[email protected]>

* CI: drop redundant Build step

"Build Tests" is strictly a superset of "Build". in addition to
the library, the former builds the tests also. both these
steps share the same set of command line arguments. by removing
"Build" step, we don't lose anything regarding the test coverage
and information for further investigation if the build fails.

Signed-off-by: Kefu Chai <[email protected]>

---------

Signed-off-by: Kefu Chai <[email protected]>

…w` properly when the library is compiled by MSVC on Windows. (jbeder#1227)

…#1231)

… be disabled by YAML_CPP_DISABLE_UNINSTALL

add_custom_target uninstall causes a clash with the same target in rtmidi (https://github.com/thestk/rtmidi).
"add_custom_target cannot create target "uninstall" because another target with the same name already exists.  The existing target is a custom target created in source directory"

Adds new option YAML_CPP_DISABLE_UNINSTALL 

fixes jbeder#1151

---------

Co-authored-by: Megamouse <[email protected]>
Co-authored-by: Jesse Beder <[email protected]>

* CMake: Add option to set the package install dir.

* CMake: Fix generated config.

- `YAML_CPP_SHARED_LIBS_BUILT` should not be set with a `PATH_VAR` as it
would always evaluate to true.
- `YAML_CPP_LIBRARIES` should used the exported target name including
the namespace, but `check_required_components` shouldn't.
- Use `CMAKE_CURRENT_LIST_DIR` to find the target file, instead of a
`PATH_VAR`. Package managers such as vcpkg move CMake configs after
installing.

* CI: Test the generated CMake package.

* CMake: Create add a deprecated yaml-cpp target.

This target is meant to provide compatibility with versions prior to
0.8.0.

* CMake: mark the yaml-cpp target as IMPORTED.

---------

Co-authored-by: Jesse Beder <[email protected]>

…dlers (jbeder#1193)

…ailable. (jbeder#1035)

There is no need to use the embedded gtest code copy in Linux systems, if they already provide the googletest framework system-wide.
Search for it, and fallback to the embedded one if the system one is not detected.

This patch has been also contributed by Simon Quigley <[email protected]>

Support for YAML Merge keys ( <<: [*dict1, *dict2] ) is added. The merge
key is a specific scalar with value << (and tag !!merge) that implies
that during node construction, the map (or sequence of maps) are merged
into the current map. The priority rules are that each key from maps
within the value associated with  << are added iff the key is not yet
present in the current map (and first map gets higher priority). Test
cases have been added accordingly.

Protect from regressions due to use of undefined or
implementation-specific behavior when using `std::` containers and smart
pointers.

This only has effect on platforms with the GNU standard C++ library.

Refer to https://gcc.gnu.org/onlinedocs/libstdc++/manual/debug_mode.html.

Signed-off-by: Diogo Teles Sant'Anna <[email protected]>

Signed-off-by: Diogo Teles Sant'Anna <[email protected]>

Bumps the github-actions group with 1 update: [actions/checkout](https://github.com/actions/checkout).

- [Release notes](https://github.com/actions/checkout/releases)
- [Commits](actions/checkout@v2...v4)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-type: direct:production
  update-type: version-update:semver-major
  dependency-group: github-actions
...

Signed-off-by: dependabot[bot] <[email protected]>

…rge-operator

…the node)

The problem is that const_map_to.get(*j->first) compares only the
shared_ptr's, while we need to compare the key itself.

v2: remove const iterator
v3: fix use-after-free due to const reference

Consider the following YAML:

    trait1: &t1
      foo: 1

    trait2: &t2
      foo: 2

    merged:
      <<: *t1
      <<: *t2

yq reports:

    $ yq .merged.foo < /tmp/yaml
    2

while the order that yaml-cpp returns is different, since it will
firstly handle 1, and will not replace it with 2:

    $ util/parse < /tmp/yaml
    trait1:
      ? &1 foo
      : &2 1
    trait2:
      foo: 2
    merged:
      *1 : *2

(Don't mix up "*2" with "2", it is trait1)

* merge-operator:
  Fix order for merging iterator
  Fix merge operator support (that can be visible by iterating through the node)
  Fix merge-key handling in case the dictionary contains a sub-dictionary
  Adding support for handling YAML Merge Key (jbeder#41)