Containers library

Thread safety

1. All container functions can be called concurrently by different threads on different containers. More generally, the C++ standard library functions do not read objects accessible by other threads unless those objects are directly or indirectly accessible via the function arguments, including the this pointer.
2. All const member functions can be called concurrently by different threads on the same container. In addition, the member functions begin(), end(), rbegin(), rend(), front(), back(), data(), find(), lower_bound(), upper_bound(), equal_range(), at(), and, except in associative containers, operator[], behave as const for the purposes of thread safety (that is, they can also be called concurrently by different threads on the same container). More generally, the C++ standard library functions do not modify objects unless those objects are accessible, directly or indirectly, via the function's non-const arguments, including the this pointer.
3. Different elements in the same container can be modified concurrently by different threads, except for the elements of std::vector<bool> (for example, a vector of std::future objects can be receiving values from multiple threads).
4. Iterator operations (e.g. incrementing an iterator) read, but do not modify the underlying container, and may be executed concurrently with operations on other iterators on the same container, with the const member functions, or reads from the elements. Container operations that invalidate any iterators modify the container and cannot be executed concurrently with any operations on existing iterators even if those iterators are not invalidated.
5. Elements of the same container can be modified concurrently with those member functions that are not specified to access these elements. More generally, the C++ standard library functions do not read objects indirectly accessible through their arguments (including other elements of a container) except when required by its specification.
6. In any case, container operations (as well as algorithms, or any other C++ standard library functions) may be parallelized internally as long as this does not change the user-visible results (e.g. std::transform may be parallelized, but not std::for_each which is specified to visit each element of a sequence in order)