Hajime Hoshi
0680ca413d
internal/atlas: bug fix: the offscreen must be an independent image
...
Closes #1938
2022-01-08 23:25:06 +09:00
Hajime Hoshi
06f4142ca0
internal/driver: Optimization: Replace interface{} with driver.Uniform
...
Converting a value from/to interface{} can create a value in heap
and this is not efficient.
2021-10-30 02:58:28 +09:00
Hajime Hoshi
4b64eadfe5
internal/mipmap, internal/atlas: Integrate for-loops for color scaling
2021-09-23 00:28:20 +09:00
Hajime Hoshi
f79acf9569
internal/mipmap: Bug fix: nil should be meaningful for a mipmap map
2021-09-21 00:46:11 +09:00
Hajime Hoshi
81fd7ab8f6
internal/mipmap: Create a mipmap map lazily
2021-09-20 23:27:14 +09:00
Hajime Hoshi
f7738b812b
internal/mipmap: Skip color scaling when possible
2021-09-06 12:11:52 +09:00
Hajime Hoshi
8c9ec8fc9f
interna/affine: Avoid UnsafeElements at At
...
Updates #1796
2021-09-04 17:55:51 +09:00
Hajime Hoshi
df710a5c63
internal/affine: Change UnsafeElements to take arguments
...
This will enable to suppress unnecessary array allocations in the
future.
Updates #1796
2021-09-04 17:38:30 +09:00
Hajime Hoshi
18903db1c6
internal/atlas: Bug fix: Unexpected padding in screenshots
...
Closes #1736
2021-07-29 16:09:29 +09:00
Hajime Hoshi
21aa96f9f5
internal/affine: Refactoring: Make ColorM interface
2021-07-27 12:10:22 +09:00
Hajime Hoshi
b466a0cbd7
ebiten: Add EvenOdd to DrawTrianglesOptions and DrawShaderTrianglesOptions
...
Updates #844
Closes #1684
2021-07-05 03:35:55 +09:00
Hajime Hoshi
26b9fa20c1
internal/graphics: Bug fix: Race condition at QuadVertices
...
QuadVertices or verticesBackend.slice reused its backend slice.
This caused a race condition. QuadVertices can be accessed from
multiple goroutines, and resetting the head and copying the data
at internal/graphicscommand might not be synced.
This change fixes this issue by basically reverting
9cb631e30f
.
Closes #1546
2021-03-20 16:32:13 +09:00
Hajime Hoshi
c7330883ef
restorable: Remove Fill and make (*ebiten.Image).Fill available for sub-images
...
Now a scissor (a clipping region) can be specified, we don't have to
worry about the rendering results out of the specified region.
Replace the implmenetation of the Fill with just a DrawTriangles with
an empty white image.
As a side effect, SubImage is avilable for Fill.
Fixes #1416
2020-11-08 02:50:06 +09:00
Hajime Hoshi
ed028110cf
ebiten: Allow rendering on a sub-image by scissor test
...
Fixes #1255
2020-11-08 00:58:44 +09:00
Hajime Hoshi
fa53160e18
mipmap: Stop using negative mipmaps
...
Negative mipmaps tend to allocate extremely big images.
Instead, encourage to use images with explicit padding when enlarging
the image.
Fixes #1400
2020-10-31 02:52:40 +09:00
Hajime Hoshi
e7d080ca4a
mipmap: Bug fix: Too big scale tried to allocate too big images
...
Fixes #1399
2020-10-21 11:28:05 +09:00
Hajime Hoshi
a3cb78558b
mipmap: Bug fix: Scale could be Inf/0 and caused a forever loop
...
Fixes #1398
2020-10-21 10:55:36 +09:00
Hajime Hoshi
b2d618c2be
mipmap, shareable: Refactoring: Remove SetGraphicsDriver
2020-10-13 01:36:52 +09:00
Hajime Hoshi
bf515bb594
Update version to v2.0.0-alpha
2020-10-04 04:30:40 +09:00
Hajime Hoshi
41564533f9
ebiten: Allow SubImage at DrawRectShader
2020-09-20 04:48:10 +09:00
Hajime Hoshi
8aef1f9080
restorable: Enable to set the 'volatile' state later
...
Updates #1309
2020-08-19 00:57:23 +09:00
Hajime Hoshi
a67fd4a2b3
mipmap: Bug fix: Adopt 32 for tooBigScale
...
I found the case that 64 was not enough on Huawei ALE-L02.
2020-08-15 15:39:27 +09:00
Hajime Hoshi
62fbac2290
mipmap: Bug fix: HasHighPrecisionFloat cannot be called before the main loop
...
Instead, use the common number for the too-big scale.
Fixes #1270
2020-07-29 12:46:09 +09:00
Hajime Hoshi
da98e67345
mipmap: Bug fix: Wrong level calculation when the indices are empty
...
Fixes #1269
2020-07-28 14:02:51 +09:00
Hajime Hoshi
637afe6d67
mipmap: Skip mipmap calculation if possible
...
If we know we can skip mipmap calculation from GeoM, let's skip it.
Updates #1265
2020-07-26 12:25:03 +09:00
Hajime Hoshi
d017a1b95e
Swap the order of the dependencies 'buffered and 'mipmap'
2020-07-26 12:15:23 +09:00
Hajime Hoshi
e0d5763a60
shader: Use the fixed number of images for shaders
...
This changes uses arrays rather than slices in order to avoid heap
allocations.
Updates #1193
2020-07-18 18:27:47 +09:00
Hajime Hoshi
52773e1d12
mipmap: Bug fix: Crash at DrawTrianglesWithShader
2020-07-16 23:02:33 +09:00
Hajime Hoshi
c4c9771b78
mipmap: Bug fix: Infinite loop at (*Mipmap).mipmapLevelFromDistance
2020-07-16 03:00:10 +09:00
Hajime Hoshi
bf56492d2d
mipmap: Bug fix: mipmapLevelFromDistance must be called from the source
2020-07-16 02:43:08 +09:00
Hajime Hoshi
f927e09f56
mipmap: Unify DrawImage and DrawTriangles
...
Fixes #909
2020-07-16 02:31:17 +09:00
Hajime Hoshi
7f2be42410
mipmap: Use mipmap for DrawTriangles
...
Updates #909
2020-07-16 02:31:11 +09:00
Hajime Hoshi
275dd438fa
mipmap: Refactoring: Do not use mipmap when the image is volatile
2020-07-16 01:19:57 +09:00
Hajime Hoshi
04b26adf3d
mipmap: Refactoring
2020-07-16 00:52:41 +09:00
Hajime Hoshi
38d3811f13
Revert "mipmap: Bug fix: Wrong maximum size of the negative-level mipmap image"
...
This reverts commit 48b192dbe9
.
2020-07-16 00:51:38 +09:00
Hajime Hoshi
48b192dbe9
mipmap: Bug fix: Wrong maximum size of the negative-level mipmap image
2020-07-15 22:59:27 +09:00
Hajime Hoshi
7f70797a6d
ebiten: Rename DrawTriaglesWithShaderOptions.Textures to Images
2020-07-15 03:49:05 +09:00
Hajime Hoshi
f8956941b7
mipmap: Do not allocate independent mipmaps for each sub-image
...
Fixes #1247
2020-07-15 02:23:45 +09:00
Hajime Hoshi
2a9fd5ef13
mipmap: Avoid using the vertex backend when GopherJS is not used
...
The backend logic is especially for GopherJS. As the backend logic
tends to consume a lot of memory, let's avoid this when possible.
Updates #797
2020-07-14 12:16:37 +09:00
Hajime Hoshi
d217bc6033
ebiten: Sparate textures from uniforms at DrawTrianglesWithShader
...
Updates #1193
Updates #1239
2020-07-09 01:59:21 +09:00
Hajime Hoshi
b83f0acc4d
Remove source-region information from vertices
...
Fixes #1210
2020-07-02 23:38:39 +09:00
Hajime Hoshi
71c9e7ac40
driver: Add Region and sourceRegion parameter at Draw
...
This is a preparation to remove source-region information from
vertices.
Updates #1210
2020-07-02 03:26:05 +09:00
Hajime Hoshi
2a63512c6e
driver: Add AddressUnsafe
...
This skips the source-region check and reduces 'if' branches from
shader programs.
AddressUnsafe is internal only so far. We might expose this value
later.
Updates #1210
2020-06-25 01:00:38 +09:00
Hajime Hoshi
2f843c49a6
shader: Fix the case when the source image is nil for shaders
2020-06-04 01:35:35 +09:00
Hajime Hoshi
f92253487f
buffered: Add Shader
2020-05-30 03:36:28 +09:00
Hajime Hoshi
218b6fc172
mipmap: Add Shader
2020-05-29 04:12:41 +09:00
Hajime Hoshi
9c637c65be
shareable: Add Shader
2020-05-29 00:15:42 +09:00
Hajime Hoshi
3550abef7a
shareable: Replace At with Pixels
...
This change replaces the API At with Pixels to reduce mutex locks.
Updates #1137
2020-04-18 21:05:51 +09:00
Hajime Hoshi
c36d4678a3
mipmap: Bug fix: DrawTriangles with a scale-only color matrix should be optimized
...
Fixes #1101
2020-03-17 02:48:29 +09:00
corfe83
b3c567de89
Fix another ~300 allocations per frame in my test project ( #1089 )
...
This simple change brings my simple test project from 752 allocations per frame to 474 allocations per frame. It seems a shame that go's escape analysis is not smart enough to leave this variable on the stack.
GeoM is a 24-byte struct and there is a slight perf difference that we are storing it in stack, but also copying it around with this change (instead of an 8-byte pointer). This could make things faster (due to stack / CPU cache) or slower (due to copying more memory) - when I try a stress test (drawing 100K images per frame), I can't see any actual performance difference (but I do see 100K fewer allocations, and GC is no longer running almost all the time).
2020-02-22 13:40:17 +09:00
Hajime Hoshi
46601bb516
graphics: Fill the screenshot in black when the screen is not transparent
...
Fixes #997
2020-02-06 03:08:16 +09:00
Hajime Hoshi
b3bdf51905
graphicscommand: Return the error immediately ( #1060 )
...
Now grpahicscommand saves the error and shows the error after a
while. This was good to simplify the API but was the cause to hide
some issues.
This change fixes all the errors to be returned immediately, and
buffer this in the ebiten package instead.
Fixes #971
2020-01-19 01:18:56 +09:00
Hajime Hoshi
2cbd5aa16e
mipmap: Remove unused members
2020-01-08 02:32:28 +09:00
Hajime Hoshi
4c8137ccf5
mipmap: Bug fix: Use more negative mipmaps when float precision is low
...
Fixes #1044
2020-01-08 02:06:36 +09:00
Hajime Hoshi
c99fd1a742
mipmap: Create mipmap package and bufferd.Image uses it
...
Mipmap calculation must be executed after the main loop starts
because the graphics driver's HasHighPrecisionFloat is needed.
Then, operations on mipmap images must be called from images in
buffered package.
Updates #1044
2020-01-08 02:06:01 +09:00