[00:11] build machine builds [00:11] but the resulting js files are 25MB [00:17] hey bai [00:18] yo DFJustin [02:28] hey SketchCow [02:37] hey SketchCow [02:47] bai: send me your ssh key [02:47] bai: ed25519 preferred [02:48] db48x2: I'll get you one on monday, all I have is my cell phone right now [02:48] ok [02:48] bai: are you still camping [02:49] yeah, until sunday [02:49] when do you leave? [02:49] thursday [02:50] we habe made advances in tbe build box [02:50] why are you still on irc [03:22] haha, we're napping in the tent right now, we just walked up and down the river [03:22] this is calif9rnia camping, the campsite has wifi [03:22] so of course I have to upload my vacation snaps in realtime [03:26] So, the box is built [03:26] We have a machine to do builds [03:26] We spent some time creating a list of every emulator the emularity needs, by number of titles [03:27] (So some get done sooner because they have 10,000 titles, over ones that have 5, etc.) [03:27] 20:01:19.833 TypeError: Module is null[Learn More] mamea2600.js:1:56 [03:27] 1. We are having issues with the compilation arguments [03:27] 2. We find it crashes, and we're trying to figure out why [03:32] Our compile arguments may be wrong or old. maybe you or DFJustin jabe ideas [03:34] what are you building [03:36] DFJustin: mamea2600.js, in this case [03:37] emmake make -j4 SUBTARGET=a2600 NOWERROR=1 REGENIE=1 SOURCES=src/mame/drivers/a2600.cpp [03:37] that seems reasonable [03:38] I have no idea what REGENIE does; SketchCow made me do it [03:39] it forces the build system to regenerate the makefiles even if it thinks it doesn't need to [03:40] it's necessary when you switch up the BLAHBLAH=1 parameters but doesn't hurt to just do it all the time [03:40] it forces mame to rebuild the lua build scripts but it should only be needed if you changed shit, I think [03:40] oh I guess it generates the makefiles from the lua script, DFJustin is mire correct :D [03:40] more* [03:42] this error is even better: [03:42] 20:35:16.574 uncaught exception: undefined (unknown) [03:42] you can't define me, maaaan [03:55] the arguments look ok, and I'm able to get a working 2600 http://fos.textfiles.com/dfjustin/a2600/a2600/ [03:57] and when uploaded to IA, it apparently works just fine [03:57] so it may just be my local environment that is broken [04:00] ah [04:17] OK, so, DFJustin [04:17] We put in all the new stuff! We jumped from 0.177 for all the MAMEs to 0.187 [04:17] Do we want to try going to the Apple II floppy that's not the faster one? Is it better? [04:34] Little help, DFJustin [04:34] Go look at https://archive.org/details/apple2_demo_ibiza_2014 and the error it now gives [04:34] I tried to fix it up, it's failing [04:39] Any ideas? I need to snap it back soon [04:45] OK, snapping back [04:46] MAME exception: Unknown slot option '""' in slot 'sl4' loader.js:1129:42 [04:46] Attempting to fall into debugger loader.js:1129:42 [04:46] exception thrown: 9282696 mameapple2e.js.gz:1:3016211 [04:46] Successfully compiled asm.js code (loaded from cache in 247ms) mameapple2e.js.gz [04:46] TypeError: asm.js link error: bad typed array constructor mameapple2e.js.gz:2:10 [04:46] TypeError: asm.js type error: Disabled by 'asmjs' runtime option mameapple2e.js.gz [04:46] uncaught exception: 9282696 (unknown) [04:48] oh sorry was watching super metroid [04:52] not sure what's going on with that exception, -sl4 "" is to disable the mockingboard and that appears to still work with 0.187 native [04:53] WOw, 0.187 is slow [04:53] unacceptably slow [04:53] the slower floppy is better for copy protections and whatnot, in my testing it was fast with webassembly but I would assume it's as slow as ever with asm.js [04:54] It's super slow. [04:55] Are the floppy sounds somewhere? [04:55] afaik they're not hooked up on apple [04:55] Sorry I'm hitting you from all sides. [04:56] 1. The whole Apple 2e is slower. Regardless of disk. [04:56] is anything specific slow on 0.187 or just everything [04:56] 2. Obviously the floppy thing is notably slower so that's noticable on a demo I have that pulls from the apple disk [04:56] So we get lots of crackly [04:57] benchmarking native, 0.187 is somewhat slower than 0.177 but not by enough that I would expect much impact [04:57] on apple2e [04:57] if the parameters are fucking up somehow that could be causing it though [04:58] { [04:58] "name": "Apple IIe", [04:58] "js_filename": "mameapple2e.js.gz", [04:58] "bios_filenames": ["apple2e.zip"], [04:58] "peripherals": ["flop1"], [04:58] "native_resolution": [560,384], [04:58] "extra_args": [""], [04:58] "driver": "apple2e" [04:58] } [04:58] So I kept it simple now [04:59] yeah that's gonna get you mockingboard + slow floppy [04:59] I was trying to give slow floppy a chance. [05:00] I'm going to revert, this is all too slow. [05:00] I'm of the opinion we should be switching over to wasm which fixes that [05:00] That's soon [05:08] are you able to get 0.187 apple2e going with the fast floppy [05:10] either mame is slower, or we didn't compile it properly [05:10] or the JS just got bigger which slowed down execution [05:11] bbl [05:11] same [05:12] *** db48x2 has quit IRC (Read error: Operation timed out) [05:26] hey dfjustin [05:27] in odyssey2.json, resolution should be 340x243, not 320x243 [05:28] can you tell us the options to compole webassembly? [05:29] add WEBASSEMBLY=1 to the make command [05:29] emularity should just habdle it now [05:30] you may need to update emscripten if it's an old version [05:45] we ate spankong ne2 [05:45] new [05:46] db is on bart [07:28] db48x: if you have the energy, we could try a webassembly of a single platform, like a800 [15:54] i also came up with a crazy plan for thr apple 2 [23:02] SketchCow: seems to work to me [23:03] it loads the wasm file [23:05] SketchCow: grab glowing-computing-machine.db48x.net:~db48x/mamea2600.js and glowing-computing-machine.db48x.net:~db48x/mamea2600.wasm [23:48] hmm, we may need a minor change to a2600.json to go along with that [23:50] yea, add this line: file_locations: { "mamea2600.wasm": "mamea2600.wasm.gz" },