Ensure that the program runs in target desktop browsers

[calebjohnston]

Test the game in Safari, Chrome, Firefox

[evndean]

Testing in Safari and Firefox shows a big performance hit when the camera starts to move. Initial research suggests that reducing the number of tilemap layers may help this. If not, I can experiment with the phaser-tiled plugin, though this seems to only work with P2 physics, so incorporating it may be fairly involved.

http://www.html5gamedevs.com/topic/11741-cannot-move-camera-with-tilemaps-on-mobile/

https://github.com/englercj/phaser-tiled

[evndean]

Results of tilemap removal tests
Branch with test code

• Safari: Performance was abysmal (20 FPS or lower) when layers had to be rendered while the camera was moving. Reducing the number of visible layers did not help. Removing the HUD update and collision handling from the Game update loop did not help.
• Firefox: Performance improved when the number of visible layers was reduced.
• Chrome: Performance was consistently at or near 60 FPS in all tests.

To get this working in all three browsers, reducing the number of tilemap layers won't cut it. I'll start experimenting with the phaser-tiled plugin.