Patch (#104)

* flat_shape convert dim_t constructor

* basic_shape convert dim type constructor

* ttl/functional

* view(const std::vector<R> &x)

merge from dev branch (#101)

* remove tensor.h

* rm src/tensor.cpp

* new api :: consistent_variable

* update basic_flat_shape

* check bound

* fix windows build

fix clang build (#98)

* fix build

* install libgtest-dev

* sudo

compatibility fix (to be released as v0.9.1) (#92)

* add an option USE_STRICT to allow disable strict build

* BUILD_LIB=OFF in cmake

* windows CI (#93)

dev-0.9.0 (#87)

* remove deprecated APIs

* tensor type name

* #include <climits>

* move ttl::experimental::zip to upstream

* fix namespace

* fix namespace

* use size_t in basic_allocator<R, cuda_memory>

* size_t

* show cuda error string (#88)

dev-0.8.0 (#85)

* more examples

* flat_tensor::ranked

* raw_tensor::typed<R, r>

* rename as_ranked -> ranked

* rename raw_shape -> flat_shape

* [[gnu::unused]]

* more detailed exception

* invalid_type_reification

* make make_shape constexpr

* flatten

* batch and vectorize

* raw_tensor::{rank,size,dims}

* new header <ttl/flat_shape>

* ref,view for basic_raw_tensor

* <ttl/type>

* move type_encoder to ttl/

* std_except.hpp

* fix cuda raw tensor (#86)

dev-0.7.0 (#81)

* new API :: device_type

* more constructors

* chebyshev_distenace

* stdc++fs

* COC

* deprecate <stdtensor>

* fix test

* support .at() for rank 0 tensor

* raw to flat

* flat_tensor::{rank,size,dims}

* simplify opencv example

* fix test for gcc

cherrypicks from dev-0.7.0 (#83)

* new API :: device_type

* more constructors

dev-0.6.0 (#78)

* INSTALL

* make ttl::range less generic

* ttl::copy

* --with-cuda

* add .size() and .dims() to tensor types (#79)

* add size method to tensor

* dims()

* deprecate from_host, to_host (#80)

* support customize install prefix

fix flatten for rank 0 (#75)