Cache the exposed area used for pathfinding and rendering FoW #3745
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…kenView() In the future, this could enable a complete separation of whether tokens are available vs whether we actually want a global view. For now, it makes the calling code more self-descriptive as the logic is now clearly about whether we want a global view and not so much whether there happens to be tokens available.
After removing `tempArea` and recognizing that the `combinedView.isEmpty()` case didn't add anything, it was clear that three of the four cases were really the same thing and could be combined. The fourth case really is distinct from the others as it does not render according to the current player view. In practice, it only renders any previously exposed area for owned tokens that no longer have sight.
Since the rendering logic is identical for any exposed areas, there is no need to include it in each case. And now that it is separated, we can time each portion separately rather than sometimes including the extra work in one timer that really belongs in another.
The `ZoneRenderer` and `AbstractAStarWalker` can make use of this cache to avoid repeating the expensive calculations. The exposed area cache is invalidated exactly whenever the visible area cache is invalidated as the two are related.
Preserving the fog buffer no longer carries much benefit as the expensive part (calculated the exposed area) is now cached on its own, and the rendering itself is typically quite cheap. And since the buffer contents have to be invalidated during the key points of user interaction (e.g., moving the map), caching the contents does not offer a perceivable benefit to the user.
Phergus
approved these changes
Nov 23, 2022
There was a problem hiding this comment.
Reviewed 4 of 4 files at r1, all commit messages.
Reviewable status:complete! all files reviewed, all discussions resolved (waiting on @kwvanderlinde)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Identify the Bug or Feature request
Addresses #3744
Description of the Change
This PR adds caching of the view-specific exposed areas. There are a few separate changes that come together to make this happen, which can be seen more easily in the commit history rather than the full diff.
The first major change is to simplify the algorithm for calculating the exposed area. The behaviour is the same, but three of the four special cases have been merged as they actually do the same thing (the fourth case is fundamentally different). Also the calculation of the exposed area is now separated from the rendering logic as the rendering logic was the same for all four cases.
The second major change is to move the exposed area calculation into
ZoneViewwhere it can be cached. BothZoneRendererandAbstractAStarCellWalkernow make use that cache to improve their responsiveness.The final major change is to remove the existing cache of the fog buffer contents. Now that the expensive portion of
renderFog()is cached in its own right, there is much less benefit to cache the buffer contents. And importantly, the buffer contents would be invalidated anyways in some of the most user-noticeable cases, namely panning/zooming the map. Removing this cache means one less thing forZoneRendererto keep track of.Sample times
Some (highly variable) timings for typical
renderFogruns using the linked campaign in #3744:Of course, when the cache is invalidated, times still aren't especially good, but are typically about 2x better than before the change.
Possible Drawbacks
This change is meant to be a pure refactoring, so shouldn't be any issues unless I changed some subtle behaviour or didn't get the cache invalidation correct.
Documentation Notes
N/A
Release Notes
This change is