30 Mar 2023 |
Guldoman | https://freetype.org/freetype2/docs/reference/ft2-error_enumerations.html | 14:31:43 |
Takase actual | freetype more like free me from eternal suffering | 14:31:53 |
Guldoman | ah fuck, found a nice optimization for canvas | 14:47:05 |
Guldoman | aka match the goddamn window pixel format | 14:47:18 |
Guldoman | now the full window rendering of doom is smooth | 14:47:50 |
Guldoman | and uses 50% of the CPU | 14:47:57 |
Guldoman | I wonder if we have some optimizations like this around our renderer | 14:48:14 |
Guldoman | * I wonder if we have some optimizations like this to do around our renderer | 14:48:21 |
aethy#7194 | You're begging for it | 15:36:51 |
aethy#7194 | https://www.youtube.com/watch?v=E2SVkZu-yNY&ab_channel=Yronno%27sClips | 15:37:05 |
aethy#7194 | I really do need to write a bot | 16:51:09 |
aethy#7194 | Trigger on \b[aA]{3,}\b | 16:52:06 |
aethy#7194 | * Trigger on \b[aA]{3,}\b | 16:55:15 |
Guldoman | I'm adding a parameter to canvas creation to specify if the canvas should be transparent | 17:03:54 |
Guldoman | if it's not, we just use the window pixel format | 17:04:03 |
Guldoman | this improves performance by a lot | 17:04:14 |
Guldoman | what should be the default? | 17:04:45 |
Guldoman | transparent? | 17:04:49 |
Guldoman | also what name should I use for the parameter? opaque? transparent? | 17:05:05 |
Guldoman | this raises a problem tho | 17:13:45 |
Guldoman | what if the purpose of the canvas wasn't to draw on the window? | 17:14:07 |
Guldoman | could using the window pixel format cause problems for people that want to use canvas as an intermediary? | 17:14:37 |
Guldoman | * this raises a question tho | 17:14:58 |
Guldoman | at this point I might just allow specifying a pixel format for the canvas itself | 18:10:04 |
Guldoman | so it would be like either boolean to indicate optimized for transparency or not, or a pixelformat table | 18:10:44 |
Guldoman | WTF SDL, why does creating an RGBA8888 take double the time to create an ARGB8888 | 21:29:04 |
Guldoman | do I need to dive in SDL again, or does anyone already know why? | 21:29:40 |
Guldoman | ok no maybe the issue is the scaled blitting??? why is blitting from an RGBA8888 to a surface with the same format slower than blitting onto an ARGB8888 | 22:17:22 |
Guldoman | AGH | 22:17:35 |
Guldoman | alright I'll think more about it tomorrow | 22:17:43 |