My general-purpose libraries and reusable code written and stored over the years. Except where specifically mentioned, these are avilable under a dual licence: public domain where such a thing exists, and MIT otherwise. See the LICENSE file for the MIT-license.
The various single-file libraries. The following table describes them, and how "fit for production" they are, as well as which other single-file libraries they rely on. Most support custom allocators if relevant.
For more complete descriptions and instructions, see the comments at the top of each file.
| File | Description | FFP* | Dependencies | Notes |
|---|---|---|---|---|
| vul_astar.h | Generic A* implementation | ☆ | vul_priority_heap, vul_queue, vul_stack, vul_stable_array | |
| vul_audio.h | OS-agnostic audio wrapper around PulseAudio, ALSA, OSS, WaveOut and CoreAudio | ⚐ | WIP: OSS is untested, >2 channels is untested. ALSA uses a hack instead of proper polling | |
| vul_benchmark.h | Simple function benchmarking | ☆ | vul_timer, vul_sort | |
| vul_cl.h | OpenCL wrappers and helper functions | ⚐ | vul_stable_array | |
| vul_cmath.h | C vector and matrix math library | ☆ | Needs tests, but used a lot | |
| vul_csp.h | Constraint satisfaction problem solver using GAC | ☆ | vul_astar | |
| vul_distributions.h | Contains a halton series generator | ⚐ | Intended to contain more distributions as the need arises | |
| vul_exploding_comment.h | Timed comments-as-macros. Compile-time error at set date (in debug). | ⚐ | Requires C++11 (constexpr) | |
| vul_file.h | System-agnosting mmap and file change monitoring + some stb.h file-related functions | ⚐ | vul_string | Has seen some use, but has no tests. OS X file monitoring missing. |
| vul_gl.h | Some minor GL helper functions (error printing and glu-matrix-likes) | ☆ | ||
| vul_hash_map.h | Generic C hash-map using Robin Hood hashing. | ⚐ | Has decent test coverage, but has seen limited use. | |
| vul_hash_map_linked_list.h | Generic C hash-map. Buckets are linked lists (which is bad) | ⚐ | vul_linked_list | Has seen plenty of use and is stable, but slow. Left in for legacy, use vul_hash_map.h instead! |
| vul_linalg.h | Linear system and gen. linear least squares solvers, singular value decomposition | ☆ | ||
| vul_linked_list.h | Non-intrusive linked list | ☆ | ||
| vul_noise.h | Various noise functions | ⚐ | Currently generates gaussian and worley noise only | |
| vul_priority_heap.h | Generic fibonacci heap | ☆ | Needs tests | |
| vul_queue.h | Generic queue (linked list of fixed-size arrays) | ☆ | vul_linked_list | |
| vul_raycast.h | Triangle-soup ray-caster using SSE | ⚠ | WIP (BVH version is incomplete, both untested) | |
| vul_resizable_array.h | Stretchy buffer. Re-allocates on resize | ☆ | ||
| vul_rngs.h | A number of PRNG implementations | ☆ | The PCG32 function implementation is Apache 2.0 licenced. See comment in source | |
| vul_shapes.h | Arbitraty precision 3D shape construction functions | ⚠ | WIP, only supports spheres (by tetrahedron subdivision) | |
| vul_skip_list.h | Skip-list implementation. Not indexable | ⚠ | WIP/Broken | |
| vul_socket.h | OS-agnostic TCP sockets | ⚐ | Needs more testing on windows and OS X, but has been used for game jams | |
| vul_sort.h | Sorting for vul_resizable_array. Insertion-, shell-, quick- and a variation of Timsort (using shellsort for runs) | ☆ | vul_resizable_array | |
| vul_stable_array.h | Generic resizable array that does not reallocate on resize | ☆ | ||
| vul_stack.h | Generic stack | ☆ | vul_stable_array | Pointers to elements are stable |
| vul_string.h | Some useful string functions (UTF8 & search/pattern matching (KMP)) | ⚐ | UTF-8 functions are due stb.h | |
| vul_thread.h | OS-agnostic threads and mutexes | ⚠ | Useful when you want to avoid the C++ stdlib, but required dynamic linking for pthreads on non-windows platforms | |
| vul_timer.h | OS-agnostic timer & sleep function | ☆ | ||
| vul_types.h | Defines standard types in my favourable form | ☆ | Useful if you like the form, horrible otherwise. |
* FFP (Fit-for-production) is a 3-tier system:
- ☆ means this either has extensive tests or has been used to a degree that it should be low on bugs.
- ⚐ means it either has tests but has not been used, or has been used some, but not extensively, and should probably be extensively tested before it is used in production.
- ⚠ indicates it has neither been tested nor used, and is probably a WIP.
A scalar linear math library containing Vectors, Matrices, Points, Quaternions, AABBs general affine transformations as well as fixed point and half precision floating point types.
The library contains both a C++11 version and a C++98-version, the primary difference being the way we construct our types. The C++11 style uses normal constructors and initializer lists:
Vector< f32, 3 > foo( a ), bar{ b, c, d };
foo = Vector< f32, 3 >( a );
while the C++98-version uses a function style:
Vector< f32, 3 > foo = makeVector< f32 >( a, b, c );
foo = makeVector< f32, 3 >( a );
For SIMD work on our vectors we use an AOSOA architecture; we pack 2-8 vectors into
vectors of simd types, f.e. 4 Vector< f32, 3 >s into 1 Vector< __m128, 3 >, then operate
on those.
Include vul_math.hpp only to use the normal features. For the bare-bones bezier tracing functionality or linear solvers, these headers must be included separately after vul_math.hpp. Note that vul_linear.hpp is is the beginning of a port of vul_blas.h, but is neither feature complete nor bug-free; avoid for now!