In short:
- No more grid-for-alpha; everything uses generated textures now
(With proper bilinear filtering so it looks like the old method)
- All tiles are now shown partially uncovered at the same time;
selecting one is no longer needed
- Gems can now be local (associated to a tile) or global.
Local games move with their tile, global ones stay where they were placed
Background:
Originally there were two possible implementations of how to render the world map:
- One used write-alpha-to-texture to implement graual uncovering.
- The other (permanently enabled) used the DrawGrid to render the map tiles
as a fine grid, each little square having its own alpha value
The downside of the first method was that it didn't look as good as the
second, so i guess that's why it was never fully finished.
The main downside of the second method was that it burned a lot of vertices
just to do alpha, so only one tile at a time could show the detailed grid.
I also never liked how an entire tile was effectively fully uncovered once
the map was first entered, taking away a lot of the exploration feeling
that could have been there if everything that hasn't been explored would be
completely invisible.
I've added this worldmap uncovering method as an optional config param,
<WorldMap revealMethod="1"/> but i've decided to fully switch over now.
Things left to be done:
- create a WorldMapRender instance only once and keep the tiles across map loads
- add debug option to reload/recreate worldmap at runtime
- cleanup gem storage and carry over the player gem properly (ged rid of std::list)
- remove "worldmap" grid render type
- Add more user "pyramid" gems as world map markers. More colors!
- check that gems and beacons still work as they should
Eliminates stutter on slow machines or in debug mode when operating doors.
Also the game shouldn't freeze anymore when marking very large tiles
as obstructing.
One such case is the editor; if Ctrl was mapped to RMB then rotating
tiles in 45° increment mode is impossible
(since the editor checks for Ctrl to enable increment mode,
and it did the wrong thing when releasing "both" at the same time)
First i thought this would fix the bug fixed in the prev. commit,
but it was unrelated.
Now that i've changed it and made it more readable, may as well keep it.
The original use case of the update flag was to prevent those layers that
only had map Element tiles on them to update each tile, which would be
very slow and most tiles never had any reason to be updated.
(Those that needed to were tracked in a separate list.)
Since the map Element class does no longer exist and tiles are handled
much more efficiently now, it makes sense to not arbitarily disable
object updates for layers that used to have only map tiles on them.
Note to self: This fixes the skeleton entities properly reacting to
shots now when placed on tile layers.
more accurate check than previously; because the normal isOnScreen()
uses a bounding circle check... which gets pretty large for something
that spans the entire screen width.
- remove top & bottom black bars
- fix minimap position to take virtual y offset into account
- same for editor bar, and version label
- fix grid render errorneously assuming that core->cameraPos.y was always 0
- fix typo in gradient height adjustment
- don't slide minimap when there's enough virtual space so that it's not in the way
This commit changes a bunch of internal rendering logic to
use FBOs in a way that doesn't violate the GL spec.
The water surface FBO's output texture was bound for reading
while at the same time rendering the water surface back into the
same FBO! Depending on the card/driver/load/zoom factor/moon phase,
this could lead to water surface flickering, chessboard effects,
and other visual glitches.
In order to fix this an extra FBO is needed.
In theory this is a simple fix but in practice this is the Aquaria
codebase and everything is more complicated than it has any right to be.
Couple other things:
- FBOs no longer have a depth renderbuffer. Aquaria never uses the depth
buffer for anything, so this can go to save some memory.
Also remove renderbuffer GL function pointers.
- Make FBOs multi-"paged". This is supposedly more efficient on desktop GL,
if glDrawBuffer() supports GL_COLOR_ATTACHMENTn. This is currently checked
via the presence of glDrawBuffersARB().
- Main core FBO now has 2 pages becaus it's needed for the water surface.
The same 2 pages are later used by the after effect manager to ping-pong
postprocessing shaders. Remove private after effect FBO.
TODO:
- There's still a bug in the one-fbo-multiple-binding-points code path.
-> for now glDrawBuffersARB must be NULL to work properly.
- no longer uses GL matrix stack manipulation, all matrix math is done locally
- use fixed rotation matrixes for fh/fv
- don't touch GL buffers (fixes crash with prev. commit)
