Thomas Harte
c173777d12
Extends TextureTarget so that targets can be created with a one-bit stencil.
2018-11-17 15:51:12 -05:00
Thomas Harte
16dfeb3fc8
Discards empty lines, yet makes some attempt at restoring transparency.
...
The two things conflict more than a little, so work to do.
2018-11-15 21:51:27 -05:00
Thomas Harte
5a31891048
Returns Amstrad CPC output.
...
Which is probably it until I get some more composite processing back in.
2018-11-15 21:32:22 -05:00
Thomas Harte
8b37496447
Restores video output to the Master System.
2018-11-15 21:21:54 -05:00
Thomas Harte
8f6664f0d7
Starts towards picking an input shader based on data type and pipeline.
2018-11-15 21:02:46 -05:00
Thomas Harte
8a699b6072
Kills setup_output definitively, saving some indirection. set_scan_target takes its place.
2018-11-14 21:52:57 -05:00
Thomas Harte
87df8b9e85
Makes an attempt at pre-emptive line buffer clearing.
2018-11-14 21:19:14 -05:00
Thomas Harte
91b19c5c70
Adds bookending, and finally kills the TextureBuilder. Farewell.
2018-11-14 20:49:06 -05:00
Thomas Harte
0487580a1a
Corrects initial state of is_drawing_ and expands lines to full display.
2018-11-14 20:10:38 -05:00
Thomas Harte
3dca836571
Ensures no overflow, and adds a couple of consts.
2018-11-14 20:09:57 -05:00
Thomas Harte
6ba02c44d0
Better binds buffer sizes.
2018-11-13 23:08:51 -05:00
Thomas Harte
bf3ab4e260
Proceeds as drawing to the unprocessed line buffer and drawing from it.
...
Very, very slowly, and without yet clearing.
2018-11-13 21:15:33 -05:00
Thomas Harte
02f9cada43
Communicates the colour subcarrier frequency, and uses it to pick a buffer width.
2018-11-13 18:33:44 -05:00
Thomas Harte
654a19ea15
Switches back to working on the scan shaders.
...
Pixels from the emulated machine are now starting to appear.
2018-11-12 22:52:26 -05:00
Thomas Harte
ecb5504bd1
Switches enable_vertex_attribute_with_pointer to silent failure (versus glGetError).
2018-11-12 22:51:44 -05:00
Thomas Harte
2adf3d353e
Subtracts retrace periods from output scale.
2018-11-12 20:20:09 -05:00
Thomas Harte
3045e85004
Ensures redraws when resizing; declines to busy wait otherwise.
2018-11-12 20:15:38 -05:00
Thomas Harte
e9d1afd515
Appears to demonstrates that the line buffer is approximately working.
2018-11-12 19:10:48 -05:00
Thomas Harte
833ab7945b
Slow steps towards switching to line output.
2018-11-12 18:56:54 -05:00
Thomas Harte
0af1d668a6
Takes a first step towards generality, and thereby starts submitting lines.
2018-11-12 18:47:55 -05:00
Thomas Harte
0ac62e3805
Flips and properly sizes output scans.
2018-11-12 18:28:09 -05:00
Thomas Harte
938d09f34a
Corrects scan outline generation.
2018-11-12 18:23:45 -05:00
Thomas Harte
dce52d740d
Finally gets some pixels back on screen.
...
For now, just the raw scans, direct to the framebuffer, with no intermediate processing. But it seems to prove that at least some of the proper data is reaching the GPU.
2018-11-11 23:23:42 -05:00
Thomas Harte
3ae333fa84
Edges further towards reviving the shaders.
2018-11-11 21:41:13 -05:00
Thomas Harte
d5af1f3948
Removes some migrated work.
2018-11-11 16:22:14 -05:00
Thomas Harte
0ba3ae53ab
Connects up the necessary recording to use intermediate composite buffers.
2018-11-11 15:20:18 -05:00
Thomas Harte
be12d78c83
Corrects vertical event announcement, and adjusts namespaces for OpenGL primitives.
2018-11-11 15:11:32 -05:00
Thomas Harte
b70227ac1b
Ensures proper write area locations end up in the scans.
2018-11-10 21:10:33 -05:00
Thomas Harte
6d277fecd5
Makes ScanTarget a little more communicative and orthogonal.
2018-11-10 19:52:57 -05:00
Thomas Harte
491817d85c
Corrects allocation error and begins submitting raw textures.
2018-11-08 23:02:36 -05:00
Thomas Harte
20faf4e477
Adds submission of scans to the GPU.
2018-11-08 22:21:11 -05:00
Thomas Harte
4fe5c7c24e
Conspires to handle multithreading side of things in a lockless fashion.
...
At least on x86-64.
2018-11-08 21:57:28 -05:00
Thomas Harte
36bf640c6f
Acts as if it is going to submit scans, at least.
2018-11-07 22:53:46 -05:00
Thomas Harte
7881e40e0b
Shuffles the OpenGL primitives into their own collection.
2018-11-07 19:11:01 -05:00
Thomas Harte
55da1e9c0f
Simplifies semantics a little and starts accepting a single buffer of pixel data.
2018-11-06 22:23:38 -05:00
Thomas Harte
9799aa0975
Completes documentation and rounds out implementation.
2018-11-04 22:17:33 -05:00
Thomas Harte
1effb97b74
Reintroduces colour phase acquisition from the colour burst.
2018-11-04 21:57:46 -05:00
Thomas Harte
eb28095041
Ensures proper accumulation and reporting of colour phase across lines.
2018-11-04 21:44:22 -05:00
Thomas Harte
014da41471
Ensures scan positions are communicated with a specified range, and switches manner of pixel clock communication.
2018-11-04 21:06:25 -05:00
Thomas Harte
0446e350d3
Resolves sizing of texture coordinates, and improves constness slightly.
2018-11-03 23:51:26 -04:00
Thomas Harte
05fb7db147
Reduces CRT chattiness.
2018-11-03 23:47:41 -04:00
Thomas Harte
f6562de325
Possibly adds enough for the Electron and ZX80 to start outputting dummy lines.
...
Let's see!
2018-11-03 23:40:39 -04:00
Thomas Harte
b40211d2c0
Starts to bend 'CRTMachine' to a world farther from owning the GPU relationship.
2018-11-03 21:54:25 -04:00
Thomas Harte
da4d883321
Adds first, incomplete attempts to talk to a ScanTarget from the CRT.
...
Does away with the hassle of `unsigned` while I'm here; that was a schoolboy error.
2018-11-03 19:58:44 -04:00
Thomas Harte
373820f080
Attempts to establish interface to decouple scan output from generation.
...
Restores some functionality that had dropped out in the interim: diagonal scans, decoupling of scan scaling from timing of the composite subcarrier.
2018-10-30 21:50:35 -04:00
Thomas Harte
4701aa149a
Adds first draft of an interface to separate CRT logic from the GPU-side stuff.
2018-10-29 22:08:17 -04:00
Thomas Harte
7eeefd2602
Ensures LOGs look like statements even in release builds.
2018-10-22 22:37:11 -04:00
Thomas Harte
7c65cfd932
Adds default values for WriteArea.
2018-10-21 21:18:54 -04:00
Thomas Harte
e582b4c8ca
Eliminates some dangling references to iCoordinate.
2018-09-13 19:35:15 -04:00
Thomas Harte
a7508bc2ae
Switching explicitly to one pixel per sample eliminates the need for a bookender.
2018-09-12 20:11:17 -04:00
