closes https://github.com/nlohmann/json/issues/670

Fix typo in README.md

Fixes https://github.com/nlohmann/json/issues/1486.

Implement contains() member function

Fix documentation

Remove two instance of duplicate paragraphs.

Disable Clang -Wmismatched-tags warning on tuple_size / tuple_element

Disable installation when used as meson subproject. #1463

remove extra semicolon from readme

docs: README typo

Just another small typo I found in the README.

in example of NLOHMANN_JSON_SERIALIZE_ENUM

Closes #1447.

attempt to fix #1445, flush buffer in serializer::dump_escaped in UTF8_REJECT case.

allow push_back() and pop_back() calls on json_pointer

serializer use fixed buffer. Whenever it is nearly full, it is flushed to `output_adapter_t<char> o`
But the code forgets to flush when there is a invalid utf8 code point
So there will be buffer overflow.

🚨 fixed unused variable warning

Add instructions about using nlohmann/json with the conda package manager

Since 6503e83e "Improve dump_integer performance by implementing a more efficient int2ascii"
an unused variable and thus a warning existed.

Putting pop_back() to public and creating a trivial push_back()
method allows users of nlohmann::json_pointer to manipulate an
existing json-pointer by adding or removing keys at the end.

This is useful for traversing a JSON-instance and keeping track
of its "absolute path" at any moment.

In my case for a schema-validator error-handler.

Closes #1400.

Closes #1405

- Catch 1.12.0 -> 1.12.2
- Google Benchmark 1.3.0 -> 1.4.1

Improve dump_integer performance

Fix x64 target platform for appveyor

Add some unit tests for formatting integers
to keep code coverage as before.

This benchmark is a sample of 1 million "small" integers
in range [-1000000 1000000) sampled from uniform distribution.

This commits implements a faster int2ascii inspired by "Fastware" talk given
by Andrei Alexandrescu.
See: https://www.youtube.com/watch?v=o4-CwDo2zpg