Update.
This commit is contained in:
@@ -1,4 +1,5 @@
|
|||||||
------------------------CVS after 3.00 beta 6rc2----------------
|
------------------------CVS after 3.00 beta 6rc2----------------
|
||||||
|
BUG FIX: Crash handler restarts correctly.
|
||||||
BUG FIX: Fixed choosing song filenames when saving.
|
BUG FIX: Fixed choosing song filenames when saving.
|
||||||
BUG FIX: Fix Endless crash when a song folder has over 99 minutes of songs.
|
BUG FIX: Fix Endless crash when a song folder has over 99 minutes of songs.
|
||||||
BUG FIX: Can no longer keep a hold lit if it's missed completely.
|
BUG FIX: Can no longer keep a hold lit if it's missed completely.
|
||||||
|
|||||||
@@ -6,6 +6,33 @@ Some of this is stuff I'm not quite sure how to deal with, but too
|
|||||||
low-pri to bother bringing up on the list.
|
low-pri to bother bringing up on the list.
|
||||||
********
|
********
|
||||||
|
|
||||||
|
Resizing tiled textures is a pain: if we scale down a tiled 1024x1024
|
||||||
|
texture to 256x256, then we need *four pixels* of buffer between each
|
||||||
|
tile to ensure they don't bleed into each other. Splitting all tiled
|
||||||
|
bitmaps into lots of smaller textures sounds good (for other reasons,
|
||||||
|
too). But for fonts, this would create hundreds of tiny textures; that
|
||||||
|
would be a performance hit (four fonts could mean over 1000 textures,
|
||||||
|
and we'd have to flush the render queue *every character*--ack.) Maybe
|
||||||
|
BitmapTexts could pre-render their text; then they'd be nearly free,
|
||||||
|
but changing text would be more expensive. Maybe too expensive; eg.
|
||||||
|
the Oni timer changes every frame. But we might have to do that anyway
|
||||||
|
with fallback fonts; how is Jared implementing that?
|
||||||
|
|
||||||
|
If we split tiled sprites apart, we can also tile BGAs with only one
|
||||||
|
render, by using the texture matrix (I think--maybe we'd just have to
|
||||||
|
stretch.) Then we might be able to use multitexturing (if available)
|
||||||
|
for BGAs, to render multiple BGALayers in one pass; they get slow. Tricky
|
||||||
|
to interface, though.
|
||||||
|
|
||||||
|
Even with texture wrapping off, I'm still seeing borders. With my new
|
||||||
|
texture loader, the bottom gameplay frame has a black border instead of
|
||||||
|
a gray one. Is that the border color or is it still wrapping? Either
|
||||||
|
is a problem; even if it *is* the border color, it shouldn't render anything
|
||||||
|
there. (Can the border be set transparent? But that would cause transparent
|
||||||
|
borders, which might cause seaming ...)
|
||||||
|
|
||||||
|
********
|
||||||
|
|
||||||
Abstract NoteData completely, and then make the arrays vectors, so they
|
Abstract NoteData completely, and then make the arrays vectors, so they
|
||||||
don't have arbitrary limits and don't take lots of extra memory.
|
don't have arbitrary limits and don't take lots of extra memory.
|
||||||
|
|
||||||
@@ -237,6 +264,8 @@ obvious, minor ones here that nonetheless need to be done:
|
|||||||
* no #pragma once
|
* no #pragma once
|
||||||
* MFC->STL is hard, because most STL implementations aren't very STL.
|
* MFC->STL is hard, because most STL implementations aren't very STL.
|
||||||
g++ 3.1 is much better, but VC6 isn't. (How is VC7?)
|
g++ 3.1 is much better, but VC6 isn't. (How is VC7?)
|
||||||
|
Hmm. Things aren't as bad as they seemed a while back; we should be
|
||||||
|
careful, but maybe it's not hopeless.
|
||||||
|
|
||||||
(This is long-term.)
|
(This is long-term.)
|
||||||
|
|
||||||
@@ -303,3 +332,7 @@ This is probably difficult to implement reliably without having
|
|||||||
access to all of the distinct adapter types, since this is very
|
access to all of the distinct adapter types, since this is very
|
||||||
special-cased.
|
special-cased.
|
||||||
|
|
||||||
|
********
|
||||||
|
|
||||||
|
Option to skip the music select screen and always pick all music.
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user