Brett 1759b5f42f | ||
---|---|---|
.idea | ||
.vscode | ||
cmake | ||
include/blt | ||
libraries | ||
src/blt | ||
tests | ||
.gitignore | ||
.gitmodules | ||
CMakeLists.txt | ||
CMakeSettings.json | ||
LICENSE | ||
README.md | ||
arg_parse_variant_map.drawio | ||
blt_argparse_variant_map.png | ||
build_and_run_debug.sh | ||
cloc.sh | ||
commit.py | ||
commit.py.save | ||
commit.sh | ||
design.txt | ||
exclude.txt | ||
icon.png | ||
icon_large.png | ||
icon_small.png | ||
py_commit_helper.sh |
README.md
BLT v0.20
A C++17 common utilities library to make thing easy!
Features
-
blt/fs
-
loader.h
- std::string blt::fs::getFile(std::string_view path)
- Gets the entire file as a string.
- std::vector<std::string> blt::fs::getLinesFromFile(std::string_view path)
- Gets the entire file as a string, then splits on the new line character. Then returns a vector of those lines
- std::vector<std::string> blt::fs::recursiveInclude(std::string_view path, std::string include_header, std::vector<include_guard> guards);
- Recursively include in order based on the include string (include_header) marked by the include guards
- Defaults to C/C++/GLSL preprocessor style (Was designed for GLSL)
- std::string blt::fs::loadBrainFuckFile(const std::string& path)
- Load a brainfuck file recursively, uses ~ to mark the include path
- std::string blt::fs::getFile(std::string_view path)
-
nbt.h
- probably needs to be remade (TODO)
-
-
blt/math
-
averages.h
- blt::averagizer_o_matic
- Computes a rolling average in a dynamically allocated array.
- Useful for average FPS over a period of time
- blt::averagizer_o_matic
-
fixed_point.h
- Provides a 64bit fixed point number with 2^32 bits of precision on each side of the decimal
- Mostly optimal assembly, avoids sign extension by using 64 bit numbers
- Requires your compiler to support 128 bit numbers
- Docs: TODO
-
fixed_point_vector.h
- Not finished due to complications in the vector type, this is a TODO
-
log_util.h
- provides overloads for logging math types
-
math.h
- blt::toRadians
- to radians function (from deg)
- blt::f_randi
- fast random integer from seed
- blt::fsqrt
- fast inverse square root algorithm, floats only
- blt::pow
- power function
- blt::round_up
- rounds up to a variable number of decimal places
- blt::toRadians
-
Specialties
blt::logging.v2
Found in the header file include/std/logging.h
, blt::logging is a fast and simple
single header logging library which can be used independently of BLT. Just include
the header file anywhere in your project and in ONE cpp file define #define BLT_LOGGING_IMPLEMENTATION
.
If you are using BLT as a CMake library (as you should!) this is done for you.
-
Compile Time Config
- Using CMake
- Several options are provided which disable various logging contexts, as such global logging can be disabled by passing
-DBLT_DISABLE_LOGGING:BOOL=ON
- Options follow the pattern of
BLT_DISABLE_LEVEL
where level is one ofTRACE
,DEBUG
,INFO
,WARN
,ERROR
, orFATAL
.- This allows for individual_t logging levels to be disabled while leaving the others functional. These options can be combined.
- See CMakeLists.txt for a complete list of options.
- Several options are provided which disable various logging contexts, as such global logging can be disabled by passing
- Standalone
- The CMake options define global variables in a config file. If you are using logging standalone you will need to remove the config include.
- Macros can be used in the same pattern of
BLT_DISABLE_LEVEL
to disable the various logging levels
- Using CMake
blt::arg_parse
blt::arg_parse found in blt/parse/argparse.h
is an attempt at a near 1:1 replication of Python's argparse in C++.
Python's argparse is one of the best command line argument parsing solutions with some of the best documentation I've seen.
The goal was to create an API which would function exactly as the Python docs described, simplifying how much I needed to remember and document for this library.
blt::profile_t (v2)
The newest version of my profiler solution now features a CPU cycle counter, CPU thread execution time, and wall time. It has more options for
sorting and general printing while featuring a nicer codebase. It is an improvement over blt::profiling while maintaining (almost) complete backwards
compatability. Due to changes in the API the BLT_PRINT_PROFILE
macro deletes the internal profiles and intervals AND takes arguments which match
the new formatting options for profiling v2. However, BLT_START_INTERVAL
, BLT_WRITE_PROFILE
, and BLT_END_INTERVAL
are still the same.
Documentation for this is coming soon, along with more profiling features. the BLT_*
macros can be disabled by the standard BLT_DISABLE_PROFILING
from the v1 profiler. It is encouraged to use the new blt::* profile functions over the macros however these currently cannot be disabled. (TODO)
Features / Examples
-
Data Structures
- Queue / Stack
- faster than std::queue / std::stack
- backed by a contiguous array
- Queue / Stack
-
Utility
- Simple Random Wrapper Interface
- Simple random functions based on the PCG Hash
-
String Functions
- starts_with
- ends_with
- contains
- toLowerCase
- toUpperCase
- split
- trim
- Logging
- See blt::logging section above
- Time
- Current time in nanoseconds (without all the c++ gobbledygook)
- Java's currentTimeMilliseconds
- nanoTime as well
std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::high_resolution_clock::now().time_since_epoch()).count()
becomesblt::system::nanoTime()
- Formatted time string with year/month/date + current time
- Current time in nanoseconds (without all the c++ gobbledygook)
-
Profiling
- Basic profiler with history and formatted output