Previously, generator code under include/etl/generators needed to be
synchronized with the generated files under include/etl explicitly.
Leading to errors in the CI checks if forgotten.
This change separates out the generated parts to cpp03 specific
snippets. This way, most code under include/etl is now directly editable,
with only some pre-C++11 code separated out to
include/etl/generators/*_cpp03_generator.h which still needs to be
re-generated on respective changes. But in many cases, this is not necessary.
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Remove python3-cogapp from Dockerfile
Removed python3-cogapp from the Dockerfile installation.
* Update .devcontainer/Dockerfile
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Change base image and update Python package installation
* Change container to use cpp devcontainer version 2
* Point to specific version of the devcontainer
* Update devcontainer to use debian snapshot for reproducability
* Fetch sources via https
* Make devcontainer more robust for debian_snapshot
* Make fetch non-https for snapshot
* Install CMake from apt
* Update devcontainers to remove CMake version
* Change Clang version from 22 to 21
* Apply suggestion from @rolandreichweinbmw
---------
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Fix compilation with -Wsign-conversion
For tests with GCC and Clang
Fixes https://github.com/ETLCPP/etl/issues/632
* Fix support of negative Id in type_lookup.h: type_from_id
* Fix element access arithmetic in atomic_gcc_sync.h fetch_add/fetch_sub
* Fix rounded_integral_division.h: divide_round_half_odd(): direction is always 1
* Add a inplace_function constructor with a nullptr argument
* Add unit test for nullptr construction
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* add bounds and empty checks to containers
* address code rabbit review
* correct C++11 constexpr error
* rename new constexpr macro and make it global
* rename queue specializations' exceptions
* change front() implementation in locked queue specializations
* refactor usage of CONSTEXPR and NO_EXCEPT
* expand intrusive queue tests
* introduce lock guards on locked queues
* Print test names at test time (#1343)
* revert mutex and return changes on locking queues
* finish reverting the locked queues
---------
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Bug fixes in variant_variadic.h
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Fix namespace issue of floor/round in to_string_helper
---------
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Restrict etl::atomic for general types
Needs adding is_copy_assignable and is_move_assignable, and
adjustments to is_trivially_copyable and is_assignable
* Resolve mutable T value vs. volatile qualified methods
* Remove volatile method overloads
They are deprecated in C++20 because they don't work as users expect anyway.
MSVC hinted for this.
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Add ranges
* Print test names at test time (#1343)
* Fix conflit commit errors
* Cast return value of operator* to value_type
Fixed warning on VS2022
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Mark uninitialized use from std library
Similar to other cases, adds compiler pragmas against warnings
caused by std library in optimized builds of tests
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* manchester
* Added manchester code and test
* manchester
* Formatting and added missing file
* manchester
* Some functions can only be constexpr since C++14
* manchester
* Manchester decode and some refactoring
* manchester
* Added some missing typenames
* manchester
* constexpr void function not allowed in C++11
* manchester
* condition on static_assert tests
* manchester
* revert CMakeLists.txt
* Using ETL_STATIC_ASSERT
* Some cleanup
* manchester
* Added static_assert message
* manchester
* Added compile time tests
* manchester
* Added invert manchester
* Some refactoring
* manchester
* Disable test for now
* Move ETL_NODISCARD before static
* manchester
* Test for valid_span
* manchester
* Remove redundant (?) storage specifiers for template specializations. Storage specifier already given in base template
* manchester
* refactoring to get rid of specialized template functions in template class
* manchester
* cleanup
* manchester
* Added documentation comments
* Some refactoring
* manchester
* introducing namespace detail_manchester
* manchester
* Some refactoring
* Update tests
* manchester
* Some refactoring
* Removed possible undefined behavior by refactoring encode_span
* constexpr version of encode_span
* Static assertion for rare case where code doesn't work because CHAR_BIT is not the same as the number of bits in uint_least8_t
* manchester
* renamed valid to is_valid
* manchester
* renamed is_valid_span to is_valid
* Using etl exceptions in ETL_ASSERT
* manchester
* Removed _fast functions
* merged encode_in_place with encode and decode_in_place with decode
* removed _span to create normal overloads of encode and decode for span
* Some renaming and minor refactoring
* manchester
* Fix build issues
* manchester
* Conditionally compile manchester_decoded
* Update test_manchester.cpp
Removed redundant semicolon
* #1258 Manchester coding
* Formatting
* consistency: hex literals with lower case 0x
* #1258 Manchester coding
* Moved copyright to top of file
* Make constexpr encode/decode span functions equal for little and big endian platforms
* #1258 Manchester coding
* Added missing include
* Added missing 8bit/64bit guards
* Fixed is_valid for big endian platforms
* #1258 Manchester coding
* private memcpy alias
* #1258 Manchester coding
* Review comments
* #1258 Manchester coding
* Cleanup
* Fix build error
* #1258 Manchester coding
* Add manchester documentation
* #1258 Manchester coding
* Preparation for GitHub pages
* #1324 Manchester documentation
* Some small fixes
* Print test names at test time (#1343)
* IGN-280 biphasic amplitude as float
* Add big-endian devcontainer
* manchester
* fixed the configuration to work with GitHub Codespaces. The changes use cross-compilation with QEMU emulation instead of trying to use a native s390x container.
* manchester
* Made manchester work for big-endian
* Some updates to the container
* Manchester big-endian support
* Cleanup
* Manchester big-endian support
* add sourcedirectory
* Enable running with ctest
* Manchester big-endian support
* Update documentation
* Manchester big-endian support
* QA
* Manchester big-endian support
* QA
* Enable testing with ctest with cross-compiler (#5)
* Enable testing with ctest and with cross-compiler
* Clean up includes in manchester.h
---------
Co-authored-by: Timon Zijnge <timon.zijnge@imec.nl>
---------
Co-authored-by: Timon Zijnge <timon.zijnge@imec.nl>
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
* Print test names at test time (#1343)
* Adjustments in tuple.h from review
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Fix ETL_HAS_CONSTEXPR_ENDIANESS in unaligned_type.h
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Fix etl::optional for deleted copy constructors
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Fix span size type
size() returned a signed value while declared (unsigned) size_t.
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Limit clock duration to milliseconds type on narrow int type
On AVR, for 16 bit int types, nano and micro are not provided.
Now, consistently use them only conditionally.
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Add permutation functions
* Refactor permutation to use etl::less
* Add test for next_permutation where begin == end
* Update code to avoid multiple bind1st objects in loop
* Remove duplicate is_partitioned test case
Removed the duplicate is_partitioned test case from the test suite.
---------
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Print test names at test time (#1343)
* Fix comment in queue::pop
---------
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Added missing files from VS2022 project
* Added global data() overloads to complement etl::size()
* Added C++03 compatible implementation of etl::is_convertible
* Updated etl::span to more closely align with std::span
* Fix etl::rotate (#1327)
Per the C++ standard, std::rotate returns first + (last - middle):
* When first == middle, return last
* When middle == last, return first
* Fix greater_equal and less_equal (#1331)
* Align comparison operators (#1330)
In functional.h, the comparison operators for equal_to and not_equal_to
mismatch between the actual comparison execution and the type inference
for the return type. This change adjusts it by using the same operator==()
in the return type inference as used in the comparison execution.
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Add missing tests (#1321)
* Add missing tests
* Typo fixes
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Add ETL_FORMAT_NO_FLOATING_POINT control macro for etl::format (#1329)
When ETL_FORMAT_NO_FLOATING_POINT is defined, all floating-point formatting support (float, double, long double) is excluded from etl::format. This reduces code size on targets that do not require floating-point formatting.
Guarded sections:
- #include <cmath>
- float/double/long double in supported_format_types variant
- float/double/long double constructors in basic_format_arg
- format_floating_* functions and format_aligned_floating
- formatter<float>, formatter<double>, formatter<long double>
- Floating-point test cases in test_format.cpp
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
# Conflicts:
# include/etl/platform.h
* Manchester documentation (#1325)
* manchester
* Added manchester code and test
* manchester
* Formatting and added missing file
* manchester
* Some functions can only be constexpr since C++14
* manchester
* Manchester decode and some refactoring
* manchester
* Added some missing typenames
* manchester
* constexpr void function not allowed in C++11
* manchester
* condition on static_assert tests
* manchester
* revert CMakeLists.txt
* Using ETL_STATIC_ASSERT
* Some cleanup
* manchester
* Added static_assert message
* manchester
* Added compile time tests
* manchester
* Added invert manchester
* Some refactoring
* manchester
* Disable test for now
* Move ETL_NODISCARD before static
* manchester
* Test for valid_span
* manchester
* Remove redundant (?) storage specifiers for template specializations. Storage specifier already given in base template
* manchester
* refactoring to get rid of specialized template functions in template class
* manchester
* cleanup
* manchester
* Added documentation comments
* Some refactoring
* manchester
* introducing namespace detail_manchester
* manchester
* Some refactoring
* Update tests
* manchester
* Some refactoring
* Removed possible undefined behavior by refactoring encode_span
* constexpr version of encode_span
* Static assertion for rare case where code doesn't work because CHAR_BIT is not the same as the number of bits in uint_least8_t
* manchester
* renamed valid to is_valid
* manchester
* renamed is_valid_span to is_valid
* Using etl exceptions in ETL_ASSERT
* manchester
* Removed _fast functions
* merged encode_in_place with encode and decode_in_place with decode
* removed _span to create normal overloads of encode and decode for span
* Some renaming and minor refactoring
* manchester
* Fix build issues
* manchester
* Conditionally compile manchester_decoded
* Update test_manchester.cpp
Removed redundant semicolon
* #1258 Manchester coding
* Formatting
* consistency: hex literals with lower case 0x
* #1258 Manchester coding
* Moved copyright to top of file
* Make constexpr encode/decode span functions equal for little and big endian platforms
* #1258 Manchester coding
* Added missing include
* Added missing 8bit/64bit guards
* Fixed is_valid for big endian platforms
* #1258 Manchester coding
* private memcpy alias
* #1258 Manchester coding
* Review comments
* #1258 Manchester coding
* Cleanup
* Fix build error
* #1258 Manchester coding
* Add manchester documentation
* #1258 Manchester coding
* Preparation for GitHub pages
* #1324 Manchester documentation
* Some small fixes
---------
Co-authored-by: Timon Zijnge <timon.zijnge@imec.nl>
* Moved and split has_size_and_data in span.h to has_size & has_data in type_traits.h
* Removed has_size_and_data traits, and move to type_traits.h
Added ETL_ASSERT for for fixed extent constructors from iterator range and begin/size
* Added macro ETL_NOEXCEPT_IF that takes a compile time boolean expression
* Changed two fixed span constructors to ETL_CONSTEXPR14 due to ETL_ASSERT in the constructor bodies
Added ETL_NOEXCEPT_IF for simpler boolean conditions
Added tests for construction from mismatched sizes
* Added definition for ETL_NOEXCEPT_IF in no C++11 path
* Changes to disable construction from rvalue temporaries
---------
Co-authored-by: John Wellbelove <john.wellbelove@etlcpp.com>
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: Niu Zhihong <zhihong@nzhnb.com>
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Co-authored-by: Timon Zijnge <47081647+tzijnge@users.noreply.github.com>
Co-authored-by: Timon Zijnge <timon.zijnge@imec.nl>
* Updated message handling to be worst case O(logN)
* Copied optimised message handling from etl::fsm
* Updated fsm generator
* Updated message_router generator
* Added optimised accepts() member function
* Modified comment, as the FSM doesn't support a successor
* Updated version and release notes
* Hotfix/etl multiset iterator invalidation during erase leads to incorrect sorted order in depth first traversal (#1317)
* Fixed issue for both multiset and multimap
* Added std::is_sorted checks to all map/set tests
* Updated with coderabbit suggestions
---------
Co-authored-by: John Wellbelove <john.wellbelove@etlcpp.com>
* Updated release notes and version
* Changed std::is_same to etl::is_same in struct type_list_is_unique (#1320)
Co-authored-by: John Wellbelove <john.wellbelove@etlcpp.com>
* Updated release notes and version
* Fix etl::rotate (#1327)
Per the C++ standard, std::rotate returns first + (last - middle):
* When first == middle, return last
* When middle == last, return first
* Added missing files from VS2022 project
* Fix greater_equal and less_equal (#1331)
* Align comparison operators (#1330)
In functional.h, the comparison operators for equal_to and not_equal_to
mismatch between the actual comparison execution and the type inference
for the return type. This change adjusts it by using the same operator==()
in the return type inference as used in the comparison execution.
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Add missing tests (#1321)
* Add missing tests
* Typo fixes
---------
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Add ETL_FORMAT_NO_FLOATING_POINT control macro for etl::format (#1329)
When ETL_FORMAT_NO_FLOATING_POINT is defined, all floating-point formatting support (float, double, long double) is excluded from etl::format. This reduces code size on targets that do not require floating-point formatting.
Guarded sections:
- #include <cmath>
- float/double/long double in supported_format_types variant
- float/double/long double constructors in basic_format_arg
- format_floating_* functions and format_aligned_floating
- formatter<float>, formatter<double>, formatter<long double>
- Floating-point test cases in test_format.cpp
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Co-authored-by: John Wellbelove <jwellbelove@users.noreply.github.com>
* Manchester documentation (#1325)
* manchester
* Added manchester code and test
* manchester
* Formatting and added missing file
* manchester
* Some functions can only be constexpr since C++14
* manchester
* Manchester decode and some refactoring
* manchester
* Added some missing typenames
* manchester
* constexpr void function not allowed in C++11
* manchester
* condition on static_assert tests
* manchester
* revert CMakeLists.txt
* Using ETL_STATIC_ASSERT
* Some cleanup
* manchester
* Added static_assert message
* manchester
* Added compile time tests
* manchester
* Added invert manchester
* Some refactoring
* manchester
* Disable test for now
* Move ETL_NODISCARD before static
* manchester
* Test for valid_span
* manchester
* Remove redundant (?) storage specifiers for template specializations. Storage specifier already given in base template
* manchester
* refactoring to get rid of specialized template functions in template class
* manchester
* cleanup
* manchester
* Added documentation comments
* Some refactoring
* manchester
* introducing namespace detail_manchester
* manchester
* Some refactoring
* Update tests
* manchester
* Some refactoring
* Removed possible undefined behavior by refactoring encode_span
* constexpr version of encode_span
* Static assertion for rare case where code doesn't work because CHAR_BIT is not the same as the number of bits in uint_least8_t
* manchester
* renamed valid to is_valid
* manchester
* renamed is_valid_span to is_valid
* Using etl exceptions in ETL_ASSERT
* manchester
* Removed _fast functions
* merged encode_in_place with encode and decode_in_place with decode
* removed _span to create normal overloads of encode and decode for span
* Some renaming and minor refactoring
* manchester
* Fix build issues
* manchester
* Conditionally compile manchester_decoded
* Update test_manchester.cpp
Removed redundant semicolon
* #1258 Manchester coding
* Formatting
* consistency: hex literals with lower case 0x
* #1258 Manchester coding
* Moved copyright to top of file
* Make constexpr encode/decode span functions equal for little and big endian platforms
* #1258 Manchester coding
* Added missing include
* Added missing 8bit/64bit guards
* Fixed is_valid for big endian platforms
* #1258 Manchester coding
* private memcpy alias
* #1258 Manchester coding
* Review comments
* #1258 Manchester coding
* Cleanup
* Fix build error
* #1258 Manchester coding
* Add manchester documentation
* #1258 Manchester coding
* Preparation for GitHub pages
* #1324 Manchester documentation
* Some small fixes
---------
Co-authored-by: Timon Zijnge <timon.zijnge@imec.nl>
* Changes from review of algorithm.h on development branch (#1340)
* Add missing constexpr in algorithm.h
* Fix call of nth_element
2nd argument (nth) was missing
* Replace partition point with O(log(N)) algorithm
The C++ standard defines O(log(N)) calls of predicate as the
complexity of partition_point(). The old algorithm was linear.
* Use predicate in calculation of is_permutation consistently
In case of predicate not equal_to, the calculation previously
returned wron results
* Omit swap in selection_sort if iterators are equal
* Use difference_type in rotate_general() instead of int
* Typo fix in algorithm.h
* Simplifications in algorithm.h
Application of plain refactoring by keeping semantics
* Guard against past-end iterator in etl::rotate()
And fix scope of rotate_right_by_one for etl::rotate()
* Support empty ranges in selection_sort
* Add tests for swap_ranges
* Add tests for binary_search
* Add tests for find_end
* Add tests for accumulate
* Add tests for move_s
* Added tests for is_heap and sort_heap
* Remove early exit for empty input
* Add adjacent_find
* Add unique
* Add unique_copy
* Add merge
* Add inplace_merge
* Add partial_sort
* Add partial_sort_copy
* copilot review change
---------
Co-authored-by: John Wellbelove <john.wellbelove@etlcpp.com>
Co-authored-by: Roland Reichwein <Roland.Reichwein@bmw.de>
Co-authored-by: Niu Zhihong <zhihong@nzhnb.com>
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Co-authored-by: Timon Zijnge <47081647+tzijnge@users.noreply.github.com>
Co-authored-by: Timon Zijnge <timon.zijnge@imec.nl>
