diff --git a/README_8md_source.html b/README_8md_source.html
index 6ee34d080..b5217c999 100644
--- a/README_8md_source.html
+++ b/README_8md_source.html
@@ -63,7 +63,7 @@ $(function() {
3 [](https://travis-ci.org/skypjack/entt)
4 [](https://ci.appveyor.com/project/skypjack/entt)
5 [](https://coveralls.io/github/skypjack/entt?branch=master)
6 [](https://www.paypal.com/cgi-bin/webscr?cmd=_donations&business=W2HF9FESD5LJY&lc=IT&item_name=Michele%20Caini¤cy_code=EUR&bn=PP%2dDonationsBF%3abtn_donateCC_LG%2egif%3aNonHosted)
10 `EnTT` is a header-only, tiny and easy to use framework written in modern
12 It's entirely designed around an architectural pattern pattern called _ECS_ that
13 is used mostly in game development. For further details:
15 * [Entity Systems Wiki](http://entity-systems.wikidot.com/)
16 * [Evolve Your Hierarchy](http://cowboyprogramming.com/2007/01/05/evolve-your-heirachy/)
17 * [ECS on Wikipedia](https://en.wikipedia.org/wiki/Entity%E2%80%93component%E2%80%93system)
19 Originally, `EnTT` was written as a faster alternative to other well known and
20 open source entity-component systems.<br/>
21 After a while the codebase has grown and more features have become part of the
27 #include <registry.hpp>
39 void update(entt::DefaultRegistry ®istry) {
40 auto view = ecs.view<Position, Velocity>();
42 for(auto entity: view) {
43 auto &position = view.get<Position>(entity);
44 auto &velocity = view.get<Velocity>(entity);
50 entt::DefaultRegistry registry;
52 for(auto i = 0; i < 10; ++i) {
53 auto entity = registry.create();
54 registry.assign<Position>(entity, i * 1.f, i * 1.f);
55 if(i % 2 == 0) { registry.assign<Velocity>(entity, i * .1f, i * .1f); }
65 I started working on `EnTT` because of the wrong reason: my goal was to design
66 an entity-component system that beated another well known open source solution
67 in terms of performance.<br/>
68 I did it, of course, but it wasn't much satisfying. Actually it wasn't
69 satisfying at all. The fastest and nothing more, fairly little indeed. When I
70 realized it, I tried hard to keep intact the great performance of `EnTT` and to
71 add all the features I wanted to see in *my* entity-component system at the same
74 Today `EnTT` is finally what I was looking for: still faster than its _rivals_,
75 a really good API and an amazing set of features. And even more, of course.
79 As it stands right now, `EnTT` is just fast enough for my requirements if
80 compared to my first choice (that was already amazingly fast indeed).<br/>
81 Here is a comparision between the two (both of them compiled with GCC 7.2.0 on a
82 Dell XPS 13 out of the mid 2014):
84 | Benchmark | EntityX (experimental/compile_time) | EnTT |
85 |-----------|-------------|-------------|
86 | Creating 10M entities | 0.128881s | **0.0408754s** |
87 | Destroying 10M entities | **0.0531374s** | 0.0545839s |
88 | Iterating over 10M entities, unpacking one component, standard view | 0.010661s | **1.58e-07s** |
89 | Iterating over 10M entities, unpacking two components, standard view | **0.0112664s** | 0.0840068s |
90 | Iterating over 10M entities, unpacking two components, standard view, half of the entities have all the components | **0.0077951s** | 0.042168s |
91 | Iterating over 10M entities, unpacking two components, standard view, one of the entities has all the components | 0.00713398s | **8.93e-07s** |
92 | Iterating over 10M entities, unpacking two components, persistent view | 0.0112664s | **5.68e-07s** |
93 | Iterating over 10M entities, unpacking five components, standard view | **0.00905084s** | 0.137757s |
94 | Iterating over 10M entities, unpacking five components, persistent view | 0.00905084s | **2.9e-07s** |
95 | Iterating over 10M entities, unpacking ten components, standard view | **0.0104708s** | 0.388602s |
96 | Iterating over 10M entities, unpacking ten components, standard view, half of the entities have all the components | **0.00899859s** | 0.200752s |
97 | Iterating over 10M entities, unpacking ten components, standard view, one of the entities has all the components | 0.00700349s | **2.565e-06s** |
98 | Iterating over 10M entities, unpacking ten components, persistent view | 0.0104708s | **6.23e-07s** |
99 | Iterating over 50M entities, unpacking one component, standard view | 0.055194s | **2.87e-07s** |
100 | Iterating over 50M entities, unpacking two components, standard view | **0.0533921s** | 0.243197s |
101 | Iterating over 50M entities, unpacking two components, persistent view | 0.055194s | **4.47e-07s** |
102 | Sort 150k entities, one component | - | **0.0080046s** |
103 | Sort 150k entities, match two components | - | **0.00608322s** |
105 `EnTT` includes its own tests and benchmarks. See
106 [benchmark.cpp](https://github.com/skypjack/entt/blob/master/test/benchmark.cpp)
107 for further details.<br/>
108 On Github users can find also a
109 [benchmark suite](https://github.com/abeimler/ecs_benchmark) that compares a
110 bunch of different projects, one of which is `EnTT`.
112 Of course, probably I'll try to get out of `EnTT` more features and better
113 performance in the future, mainly for fun.<br/>
114 If you want to contribute and/or have any suggestion, feel free to make a PR or
115 open an issue to discuss your idea.
121 To be able to use `EnTT`, users must provide a full-featured compiler that
122 supports at least C++14.<br/>
123 The requirements below are mandatory to compile the tests and to extract the
126 * CMake version 3.2 or later.
127 * Doxygen version 1.8 or later.
131 `EnTT` is a header-only library. This means that including the `entt.hpp`
132 header is enough to include the whole framework and use it. For those who are
133 interested only in the entity-component system, consider to include the sole
134 `entity/registry.hpp` header instead.<br/>
135 It's a matter of adding the following line to the top of a file:
138 #include <entt/entt.hpp>
141 Use the line below to include only the entity-component system instead:
144 #include <entt/entity/registry.hpp>
147 Then pass the proper `-I` argument to the compiler to add the `src` directory to
152 The documentation is based on [doxygen](http://www.stack.nl/~dimitri/doxygen/).
159 The API reference will be created in HTML format within the directory
160 `build/docs/html`. To navigate it with your favorite browser:
163 $ your_favorite_browser docs/html/index.html
165 The API reference is also available [online](https://skypjack.github.io/entt/)
166 for the latest version.
170 To compile and run the tests, `EnTT` requires *googletest*.<br/>
171 `cmake` will download and compile the library before to compile anything else.
180 To build the benchmarks, use the following line instead:
182 * `$ cmake -DCMAKE_BUILD_TYPE=Release ..`
184 Benchmarks are compiled only in release mode currently.
190 The `Registry` to store, the `View`s to iterate. That's all.
192 An entity (the _E_ of an _ECS_) is an opaque identifier that users should just
193 use as-is and store around if needed. Do not try to inspect an entity
194 identifier, its type can change in future and a registry offers all the
195 functionalities to query them out-of-the-box. The underlying type of an entity
196 (either `std::uint16_t`, `std::uint32_t` or `std::uint64_t`) can be specified
197 when defining a registry (actually the DefaultRegistry is nothing more than a
198 Registry where the type of the entities is `std::uint32_t`).<br/>
199 Components (the _C_ of an _ECS_) should be plain old data structures or more
200 complex and moveable data structures with a proper constructor. They are list
201 initialized by using the parameters provided to construct the component. No need
202 to register components or their types neither with the registry nor with the
203 entity-component system at all.<br/>
204 Systems (the _S_ of an _ECS_) are just plain functions, functors, lambdas or
205 whatever the users want. They can accept a Registry, a View or a PersistentView
206 and use them the way they prefer. No need to register systems or their types
207 neither with the registry nor with the entity-component system at all.
209 The following sections will explain in short how to use the entity-component
210 system, the core part of the `EnTT` framework.<br/>
211 In fact, the framework is composed of many other classes in addition to those
212 describe below. For more details, please refer to the
213 [online documentation](https://skypjack.github.io/entt/).
215 ## The Registry, the Entity and the Component
217 A registry is used to store and manage entities as well as to create views to
218 iterate the underlying data structures.<br/>
219 Registry is a class template that lets the users decide what's the preferred
220 type to represent an entity. Because `std::uint32_t` is large enough for almost
221 all the cases, there exists also an alias named DefaultRegistry for
222 `Registry<std::uint32_t>`.
224 Entities are represented by _entitiy identifiers_. An entity identifier is an
225 opaque type that users should not inspect or modify in any way. It carries
226 information about the entity itself and its version.
228 A registry can be used both to construct and to destroy entities:
231 // constructs a naked entity with no components ad returns its identifier
232 auto entity = registry.create();
234 // constructs an entity and assigns it default-initialized components
235 auto another = registry.create<Position, Velocity>();
237 // destroys an entity and all its components
238 registry.destroy(entity);
241 Once an entity is deleted, the registry can freely reuse it internally with a
242 slightly different identifier. In particular, the version of an entity is
243 increased each and every time it's destroyed.<br/>
244 In case entity identifiers are stored around, the registry offers all the
245 functionalities required to test them and get out of the them all the
246 information they carry:
249 // returns true if the entity is still valid, false otherwise
250 bool b = registry.valid(entity);
252 // gets the version contained in the entity identifier
253 auto version = registry.version(entity);
255 // gets the actual version for the given entity
256 auto curr = registry.current(entity);
259 Components can be assigned to or removed from entities at any time with a few
260 calls to member functions of the registry. As for the entities, the registry
261 offers also a set of functionalities users can use to work with the components.
263 The `assign` member function template creates, initializes and assigns to an
264 entity the given component. It accepts a variable number of arguments that are
265 used to construct the component itself if present:
268 registry.assign<Position>(entity, 0., 0.);
272 auto &velocity = registry.assign<Velocity>(entity);
277 If the entity already has the given component, the `replace` member function
278 template can be used to replace it:
281 registry.replace<Position>(entity, 0., 0.);
285 auto &velocity = registry.replace<Velocity>(entity);
290 In case users want to assign a component to an entity, but it's unknown whether
291 the entity already has it or not, `accomodate` does the work in a single call
292 (of course, there is a performance penalty to pay for that mainly due to the
293 fact that it must check if `entity` already has the given component or not):
296 registry.accomodate<Position>(entity, 0., 0.);
300 auto &velocity = registry.accomodate<Velocity>(entity);
305 Note that `accomodate` is a sliglhty faster alternative for the following
306 `if`/`else` statement and nothing more:
309 if(registry.has<Comp>(entity)) {
310 registry.replace<Comp>(entity, arg1, argN);
312 registry.assign<Comp>(entity, arg1, argN);
316 As already shown, if in doubt about whether or not an entity has one or more
317 components, the `has` member function template may be useful:
320 bool b = registry.has<Position, Velocity>(entity);
323 On the other side, if the goal is to delete a single component, the `remove`
324 member function template is the way to go when it's certain that the entity owns
325 a copy of the component:
328 registry.remove<Position>(entity);
331 Otherwise consider to use the `reset` member function. It behaves similarly to
332 `remove` but with a strictly defined behaviour (and a performance penalty is the
333 price to pay for that). In particular it removes the component if and only if it
334 exists, otherwise it returns safely to the caller:
337 registry.reset<Position>(entity);
340 There exist also two other _versions_ of the `reset` member function:
342 * If no entity is passed to it, `reset` will remove the given component from
343 each entity that has it:
346 registry.reset<Position>();
349 * If neither the entity nor the component are specified, all the entities and
350 their components are destroyed:
356 Finally, references to components can be retrieved by just doing this:
359 // either a non-const reference ...
360 DefaultRegistry registry;
361 auto &position = registry.get<Position>(entity);
363 // ... or a const one
364 const auto &cregistry = registry;
365 const auto &position = cregistry.get<Position>(entity);
368 The `get` member function template gives direct access to the component of an
369 entity stored in the underlying data structures of the registry.
371 ### Sorting: is it possible?
373 Of course, sorting entities and components is possible with `EnTT`.<br/>
374 In fact, there are two functions that respond to slightly different needs:
376 * Components can be sorted directly:
379 registry.sort<Renderable>([](const auto &lhs, const auto &rhs) {
380 return lhs.z < rhs.z;
384 * Components can be sorted according to the order imposed by another component:
387 registry.sort<Movement, Physics>();
390 In this case, instances of `Movement` are arranged in memory so that cache
391 misses are minimized when the two components are iterated together.
393 ## View: to persist or not to persist?
395 There are mainly two kinds of views: standard (also known as View) and
396 persistent (alsa known as PersistentView).<br/>
397 Both of them have pros and cons to take in consideration. In particular:
402 * They work out-of-the-box and don't require any dedicated data
404 * Creating and destroying them isn't expensive at all because they don't
405 have any type of initialization.
406 * They are the best tool to iterate single components.
407 * They are the best tool to iterate multiple components at once when
408 tags are involved or one of the component is assigned to a
409 significantly low number of entities.
410 * They don't affect any other operations of the registry.
413 * Their performance tend to degenerate when the number of components
414 to iterate grows up and the most of the entities have all of them.
419 * Once prepared, creating and destroying them isn't expensive at all
420 because they don't have any type of initialization.
421 * They are the best tool to iterate multiple components at once when
422 the most of the entities have all of them.
425 * They have dedicated data structures and thus affect the memory
426 pressure to a minimal extent.
427 * If not previously prepared, the first time they are used they go
428 through an initialization step that could take a while.
429 * They affect to a minimum the creation and destruction of entities and
430 components. In other terms: the more persistent views there will be,
431 the less performing will be creating and destroying entities and
434 To sum up and as a rule of thumb, use a standard view:
435 * To iterate entities for a single component.
436 * To iterate entities for multiple components when a significantly low
437 number of entities have one of the components.
438 * In all those cases where a persistent view would give a boost to
439 performance but the iteration isn't performed frequently.
441 Use a persistent view in all the other cases.
443 To easily iterate entities, all the views offer _C++-ish_ `begin` and `end`
444 member functions that allow users to use them in a typical range-for loop.<br/>
445 Continue reading for more details or refer to the
446 [official documentation](https://skypjack.github.io/entt/).
450 A standard view behaves differently if it's constructed for a single component
451 or if it has been requested to iterate multiple components. Even the API is
452 different in the two cases.<br/>
453 All that they share is the way they are created by means of a registry:
456 // single component standard view
457 auto single = registry.view<Position>();
459 // multi component standard view
460 auto multi = registry.view<Position, Velocity>();
463 For all that remains, it's worth discussing them separately.<br/>
465 #### Single component standard view
467 Single component standard views are specialized in order to give a boost in
468 terms of performance in all the situation. This kind of views can access the
469 underlying data structures directly and avoid superflous checks.<br/>
470 They offer a bunch of functionalities to get the number of entities they are
471 going to return and a raw access to the entity list as well as to the component
473 Refer to the [official documentation](https://skypjack.github.io/entt/) for all
476 There is no need to store views around for they are extremely cheap to
477 construct, even though they can be copied without problems and reused
478 freely. In fact, they return newly created and correctly initialized iterators
479 whenever `begin` or `end` are invoked.<br/>
480 To iterate a single component standard view, just use it in range-for:
483 auto view = registry.view<Renderable>();
485 for(auto entity: view) {
486 auto &renderable = view.get(entity);
492 **Note**: prefer the `get` member function of the view instead of the `get`
493 member function template of the registry during iterations.
495 #### Multi component standard view
497 Multi component standard views iterate entities that have at least all the given
498 components in their bags. During construction, these views look at the number
499 of entities available for each component and pick up a reference to the smallest
500 set of candidates in order to speed up iterations.<br/>
501 They offer fewer functionalities than their companion views for single
502 component, the most important of which can be used to reset the view and refresh
503 the reference to the set of candidate entities to iterate.<br/>
504 Refer to the [official documentation](https://skypjack.github.io/entt/) for all
507 There is no need to store views around for they are extremely cheap to
508 construct, even though they can be copied without problems and reused
509 freely. In fact, they return newly created and correctly initialized iterators
510 whenever `begin` or `end` are invoked.<br/>
511 To iterate a multi component standard view, just use it in range-for:
514 auto view = registry.view<Position, Velocity>();
516 for(auto entity: view) {
517 auto &position = view.get<Position>(entity);
518 auto &velocity = view.get<Velocity>(entity);
524 **Note**: prefer the `get` member function template of the view instead of the
525 `get` member function template of the registry during iterations.
529 A persistent view returns all the entities and only the entities that have at
530 least the given components. Moreover, it's guaranteed that the entity list is
531 thightly packed in memory for fast iterations.<br/>
532 In general, persistent views don't stay true to the order of any set of
533 components unless users explicitly sort them.
535 Persistent views can be used only to iterate multiple components. Create them
539 auto view = registry.persistent<Position, Velocity>();
542 There is no need to store views around for they are extremely cheap to
543 construct, even though they can be copied without problems and reused
544 freely. In fact, they return newly created and correctly initialized iterators
545 whenever `begin` or `end` are invoked.<br/>
546 That being said, persistent views perform an initialization step the very first
547 time they are constructed and this could be quite costly. To avoid it, consider
548 asking to the registry to _prepare_ them when no entities have been created yet:
551 registry.prepare<Position, Velocity>();
554 If the registry is empty, preparation is extremely fast. Moreover the `prepare`
555 member function template is idempotent. Feel free to invoke it even more than
556 once: if the view has been alreadt prepared before, the function returns
557 immediately and does nothing.
559 A persistent view offers a bunch of functionalities to get the number of
560 entities it's going to return, a raw access to the entity list and the
561 possibility to sort the underlying data structures according to the order of one
562 of the components for which it has been constructed.<br/>
563 Refer to the [official documentation](https://skypjack.github.io/entt/) for all
566 To iterate a persistent view, just use it in range-for:
569 auto view = registry.persistent<Position, Velocity>();
571 for(auto entity: view) {
572 auto &position = view.get<Position>(entity);
573 auto &velocity = view.get<Velocity>(entity);
579 **Note**: prefer the `get` member function template of the view instead of the
580 `get` member function template of the registry during iterations.
584 * Entity identifiers are numbers and nothing more. They are not classes and they
585 have no member functions at all. As already mentioned, do no try to inspect or
586 modify an entity descriptor in any way.
588 * As shown in the examples above, the preferred way to get references to the
589 components while iterating a view is by using the view itself. It's a faster
590 alternative to the `get` member function template that is part of the API of
591 the Registry. That's because the registry must ensure that a pool for the
592 given component exists before to use it; on the other side, views force the
593 construction of the pools for all their components and access them directly,
594 thus avoiding all the checks.
596 * Most of the _ECS_ available out there have an annoying limitation (at least
597 from my point of view): entities and components cannot be created and/or
598 deleted during iterations.<br/>
599 `EnTT` partially solves the problem with a few limitations:
601 * Creating entities and components is allowed during iterations.
602 * Deleting an entity or removing its components is allowed during
603 iterations if it's the one currently returned by a view. For all the
604 other entities, destroying them or removing their components isn't
605 allowed and it can result in undefined behavior.
607 Iterators are invalidated and the behaviour is undefined if an entity is
608 modified or destroyed and it's not the one currently returned by the
610 To work around it, possible approaches are:
612 * Store aside the entities and the components to be removed and perform the
613 operations at the end of the iteration.
614 * Mark entities and components with a proper tag component that indicates
615 they must be purged, then perform a second iteration to clean them up one
618 * Views and thus their iterators aren't thread safe. Do no try to iterate a set
619 of components and modify the same set concurrently.<br/>
620 That being said, as long as a thread iterates the entities that have the
621 component `X` or assign and removes that component from a set of entities,
622 another thread can safely do the same with components `Y` and `Z` and
623 everything will work like a charm.<br/>
624 As an example, users can freely execute the rendering system and iterate the
625 renderable entities while updating a physic component concurrently on a
626 separate thread if needed.
630 The `EnTT` framework is moving its first steps. More and more will come in the
631 future and hopefully I'm going to work on it for a long time.<br/>
632 Here is a brief list of what it offers today:
634 * Statically generated integer identifiers for types.
635 * An entity-component system based on sparse sets.
636 * Signal handlers and event emitters of any type.
638 * Any other business.
640 Consider it a work in progress. For more details and an updated list, please
641 refer to the [online documentation](https://skypjack.github.io/entt/).
645 If you want to contribute, please send patches as pull requests against the
646 branch `master`.<br/>
648 [contributors list](https://github.com/skypjack/entt/blob/master/AUTHORS) to see
649 who has partecipated so far.
653 Code and documentation Copyright (c) 2017 Michele Caini.<br/>
655 [the MIT license](https://github.com/skypjack/entt/blob/master/LICENSE).
657 [Creative Commons](https://github.com/skypjack/entt/blob/master/docs/LICENSE).
661 Developing and maintaining `EnTT` takes some time and lots of coffee. I'd like
662 to add more and more functionalities in future and turn it in a full-featured
664 If you want to support this project, you can offer me an espresso. I'm from
665 Italy, we're used to turning the best coffee ever in code. If you find that
666 it's not enough, feel free to support me the way you prefer.<br/>
667 Take a look at the donation button at the top of the page for more details or
668 just click [here](https://www.paypal.com/cgi-bin/webscr?cmd=_donations&business=W2HF9FESD5LJY&lc=IT&item_name=Michele%20Caini¤cy_code=EUR&bn=PP%2dDonationsBF%3abtn_donateCC_LG%2egif%3aNonHosted).