#jsmess 2017-07-15,Sat

↑back Search

Time Nickname Message
00:24 🔗 db48x bai: hah, those are good
00:25 🔗 db48x bai: it's easy enough to add -confirm_quit 1 to the list of arguments in Emularity
00:30 🔗 bai that's what I hoped but when I try it causes it to abort, heh
00:30 🔗 bai I'm probably doing it wrong
00:30 🔗 db48x heh
00:31 🔗 bai https://pastebin.com/gvxJN81u
00:31 🔗 bai does that look right?
00:31 🔗 db48x yep
00:34 🔗 bai weird
00:37 🔗 bai seems like any args will do it
00:38 🔗 db48x we're already passing in a dozen or more things; what's two more?
00:39 🔗 bai oh I see what the problem is, its not "-confirm_quit 1" it's just "-confirm_quit" and then the "1" is being treated as the rom name, heh
00:39 🔗 db48x lol
00:39 🔗 bai and the other arg I tried wasn't working because I misremembered it
00:40 🔗 db48x details
00:40 🔗 db48x bbiab
00:40 🔗 bai computers should just automatically know what I want and do it!!
00:41 🔗 db48x agreed!
01:17 🔗 DFJustin oh right derp
01:22 🔗 bai works well, might be nice that it pauses when you hit esc to drop out of fullscreen or release the pointer lock
01:22 🔗 bai only weird thing is that the last frame stays on screen after you exit, so it just feels like it hung
01:50 🔗 JohnTalen bai: that's precisely what the code enforces.
01:50 🔗 JohnTalen (I was looking at your links last night)
01:51 🔗 JohnTalen it's that whole pausing functionality.
01:52 🔗 bai yeah. we could probably detect when the emulator exists in Emularity and blank the canvas or do whatever
01:53 🔗 JohnTalen bai: It's pretty wild your working inside the actual MAME project.
01:54 🔗 JohnTalen The Vice code that I'm fixing is all new unfinished code.
01:54 🔗 JohnTalen Thing is once this works I have to do it all over again for 3.1.
01:56 🔗 bai we try to keep our footprint in the mame code to a minimum, but I think by now the work of convincing people it's a valid platform has paid off, and now they just want to see the code get a little less hacky
01:56 🔗 JohnTalen But this is worth it because this project *worked* in the browser. It worked but no where near the state as the release on github. This repeats a 2d cad program port I consulted on. They didn't give me the latest source because they obviously didn't want me to steal out.
01:56 🔗 JohnTalen s/out/it
01:56 🔗 JohnTalen bai: interesting!
01:58 🔗 JohnTalen The magnetic/gravitional properties of this network/channel make me perform at least one typo in every output.
02:04 🔗 bai also, interesting fact - mame running in my brower is significantly faster than native mame running in the virtualbox vm I use to compile it
02:05 🔗 bai I dunno whether to be surprised, angry, or nonplussed about that
02:09 🔗 JohnTalen bai: bizarre. webassembly or asm.js?
02:13 🔗 JohnTalen bai: That speed difference will most likely be dependening on the repeition of the most used inner loops of main switch case. Perhaps time retrieval to determine screen syncs is vastly different between the platforms. I'd put some cout/printfs of how time computation is handing at those points.
02:21 🔗 bai wasm build. I think it's most likely just that virtualbox's video drivers are particularly crappy
02:21 🔗 bai sad that in 2017 we still can't use virtual machines for anything involving graphics
02:31 🔗 JohnTalen bai: It's been like that from the begining of emulation. Even Desqview in 1991 provided multitasking. But it was dog slow. But that was mostly because of RAM and a little bit of an overburndened CPU.
02:32 🔗 JohnTalen bai: The cause could be a compbination of drivers and how the host OS handles updates.
02:32 🔗 JohnTalen At this rate... I know WebGL will require at least another 10 years before anything serious under the browser.
02:34 🔗 JohnTalen It all comes down to how state is managed. If state were completely functional we'd be able to offload processing to other cores.
02:34 🔗 JohnTalen Redux in Javascript tries to do this. But in affect it's just a link list of states.
02:35 🔗 JohnTalen Of course graphics at various times are sequential mostly but the common tools for older systems were rectangular invalidation and off screen rendering before swapping to visual video memory.
02:37 🔗 JohnTalen Vice 3.1 does an awsome job for the c64. The sprite blur doesn't flicker as it did in early versions. And the sprite 'blur' is damned near perfect rendition of a stock c64.
02:46 🔗 JohnTalen <Gets off soapbox before falling off and tries to remember to make a typo> ;D
03:01 🔗 JohnTalen bai: My firefox 52.2 has javascript.options.wasm = false. Time to upgrade.
03:11 🔗 bai yeah sounds like it. both chrome and ff toggled wasm on by default a couple months back
03:14 🔗 JohnTalen yeah, i was just doing a little webasembly. nice stuff.
03:15 🔗 JohnTalen you can still use it under 47+. just flip the switch.
04:52 🔗 bai pull request accepted!
04:53 🔗 bai I was hoping to slip in some additional changes but they were too quick, so I've opened another pull request to export that new functionality - https://github.com/mamedev/mame/pull/2479
05:26 🔗 DFJustin congrats
05:28 🔗 bai second one was accepted too
05:29 🔗 bai so we can now save and load states or reset/exit via the JS API
05:30 🔗 bai if there's any other functionality which would be useful to trigger from the js just let me know and I'll look into that too
08:36 🔗 JohnTalen has quit IRC (Quit: leaving)
15:12 🔗 SketchCow hey!!
15:14 🔗 SketchCow can you compile jaguar and the other wasm tests in the new flow/revision?
19:39 🔗 azakai has quit IRC (Remote host closed the connection)
20:21 🔗 db48x has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 devesine has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 balrog has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 zino has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 gsathya has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 arkiver has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 SketchCow has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 bai has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 bwn has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 datajerk has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 Lord_Nigh has quit IRC (ny.us.hub hub.efnet.us)
20:21 🔗 bwn has joined #jsmess
20:21 🔗 SketchCow has joined #jsmess
20:21 🔗 balrog has joined #jsmess
20:21 🔗 db48x has joined #jsmess
20:21 🔗 zino has joined #jsmess
20:21 🔗 datajerk has joined #jsmess
20:21 🔗 gsathya has joined #jsmess
20:21 🔗 devesine has joined #jsmess
20:21 🔗 bai has joined #jsmess
20:21 🔗 Lord_Nigh has joined #jsmess
20:21 🔗 arkiver has joined #jsmess
20:21 🔗 hub.efnet.us sets mode: +oo SketchCow arkiver
20:33 🔗 db48x has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 devesine has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 balrog has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 zino has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 gsathya has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 arkiver has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 SketchCow has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 bai has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 bwn has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 datajerk has quit IRC (ny.us.hub hub.efnet.us)
20:33 🔗 Lord_Nigh has quit IRC (ny.us.hub hub.efnet.us)
20:41 🔗 bwn has joined #jsmess
20:41 🔗 SketchCow has joined #jsmess
20:41 🔗 balrog has joined #jsmess
20:41 🔗 db48x has joined #jsmess
20:41 🔗 zino has joined #jsmess
20:41 🔗 datajerk has joined #jsmess
20:41 🔗 gsathya has joined #jsmess
20:41 🔗 devesine has joined #jsmess
20:41 🔗 bai has joined #jsmess
20:41 🔗 Lord_Nigh has joined #jsmess
20:41 🔗 arkiver has joined #jsmess
20:41 🔗 ircd.choopa.net sets mode: +oo SketchCow arkiver
20:41 🔗 DFJustin has quit IRC (Remote host closed the connection)
20:48 🔗 DFJustin has joined #jsmess

irclogger-viewer