| Time |
Nickname |
Message |
|
00:04
🔗
|
|
pengan has quit IRC (Ping timeout: 633 seconds) |
|
00:46
🔗
|
Sgeo |
I tried throwing Emterpreter at VICE.... the menu works but the screen does not |
|
00:46
🔗
|
Sgeo |
My mind boggles |
|
00:48
🔗
|
Sgeo |
Oh actually it's just unbearably slow |
|
00:50
🔗
|
Sgeo |
If I can separate the menu stuff into its own main loop (which I would have to do for the coroutine.h approach anyway), I think I can get Emterpreter to only deal with that part |
|
01:53
🔗
|
|
pengan has joined #jsmess |
|
02:15
🔗
|
|
azakai has joined #jsmess |
|
02:15
🔗
|
Sgeo |
bai, Setting EMTERPRETIFY=1 forces js-opts=1, because we convert to the emterpreter’s binary format using a js-optimizer pass. |
|
02:15
🔗
|
Sgeo |
What js-optimizer is this? Because I'm still mad at it for breaking my perfectly valid JS syntax |
|
02:15
🔗
|
bai |
no idea |
|
02:50
🔗
|
Sgeo |
bai, for Emterpretify async, do all functions called by the main loop on down need to be in the whitelist? |
|
02:52
🔗
|
Sgeo |
Or maybe as a blocker? (e.g. emscripten_push_main_loop_blocker) |
|
02:59
🔗
|
bai |
anything that's up the call stack from the blocking function, yes |
|
03:06
🔗
|
Sgeo |
I'm having trouble configuring EMTERPRETIFY_WHITELIST |
|
03:06
🔗
|
Sgeo |
-s EMTERPRETIFY_WHITELIST='["_sdl_ui_trap"]' |
|
03:06
🔗
|
Sgeo |
Is that not the correct syntax? |
|
03:07
🔗
|
Sgeo |
.38.11/src/asm2wasm.h:2247: wasm::Asm2WasmBuilder::processFunction(cashew::Ref)::<lambda(cashew::Ref)>: Assertion `target[0] == SUB && target[1]->isString() && target[2][0] == BINARY && target[2][1] == AND && target[2][3]->isNumber()' failed. |
|
03:07
🔗
|
Sgeo |
ERROR:root:'/home/sgeo/emsdk/clang/e1.38.11_64bit/binaryen/bin/asm2wasm ../js/x64.temp.asm.js --total-memory=67108864 --trap-mode=allow -O3 --mem-init=../js/x64.js.mem --mem-base=1024 -g -o ../js/x64.wast -S' failed |
|
03:12
🔗
|
SketchCow |
Sgeo: I am going to bed to work all day tomorrow clearing out my storage once and for all. I'll be active tomorrow night. |
|
03:12
🔗
|
Sgeo |
SketchCow, ok, night |
|
03:22
🔗
|
Sgeo |
bai, this is profoundly unhelpful advise: |
|
03:22
🔗
|
Sgeo |
Suggested list of functions to run in the emterpreter: |
|
03:22
🔗
|
Sgeo |
-s EMTERPRETIFY_WHITELIST='[]' |
|
03:22
🔗
|
Sgeo |
Unless it's trying to tell me everything should be in the whitelist. Which is also unhelpful in its own way |
|
03:42
🔗
|
bai |
heh |
|
03:55
🔗
|
* |
Sgeo starts experimenting with the non-recommended asyncify >.> |
|
03:59
🔗
|
Sgeo |
LLVM ERROR: Stores must be a multiple of 8 bits |
|
03:59
🔗
|
Sgeo |
I, uh, have no idea what that means |
|
04:28
🔗
|
Sgeo |
bai, halp what does this mean (trying to use emterpretify_whitelist) |
|
04:28
🔗
|
Sgeo |
asm2wasm: /home/ubuntu/emslave/buildslave/ubuntu-tagbuilder/emsdk/binaryen/tag-1.38.11/src/asm2wasm.h:2247: wasm::Asm2WasmBuilder::processFunction(cashew::Ref)::<lambda(cashew::Ref)>: Assertion `target[0] == SUB && target[1]->isString() && target[2][0] == BINARY && target[2][1] == AND && target[2][3]->isNumber()' failed. |
|
04:32
🔗
|
Sgeo |
Ugh going to try disabling WASM for now |
|
04:56
🔗
|
Sgeo |
Successfully excluding maincpu_mainloop from Emterpretify... I think |
|
04:56
🔗
|
Sgeo |
So why is it still slow? |
|
05:11
🔗
|
Sgeo |
7% -> 27% in statusbar by disabling assertions and "-g"... I forgot what the latter was for, but I was using that to help determine the whitelist |
|
05:29
🔗
|
Sgeo |
Uploaded, tried on main computer. 33% |
|
05:29
🔗
|
Sgeo |
WHY |
|
05:30
🔗
|
Sgeo |
Firefox is coping much better |
|
05:40
🔗
|
Sgeo |
Yeah, I think the problem is it really needs WebAssembly to run well. I'm getting a warning in asm.js mode about an optimization that's disabled because I'm allowing memory to grow, and I'm only allowing memory to grow because resid needed more memory |
|
05:40
🔗
|
Sgeo |
Maybe I shouldn't allow memory to grow, and just double the memory amount |
|
05:44
🔗
|
DFJustin |
yeah that's a big one I think, you have to just give it a bigger fixed amount |
|
05:46
🔗
|
Sgeo |
Didn't help much. Think I really need WASM for this |
|
05:47
🔗
|
Sgeo |
Although the old vice.js was fine on ASM.JS somehow |
|
07:07
🔗
|
Sgeo |
I think Emscripten might be emitting technically invalid asm.js which asm2wasm chokes on |
|
09:38
🔗
|
|
pengan has quit IRC (Read error: Operation timed out) |
|
14:37
🔗
|
|
i0npulse has quit IRC (Ping timeout: 252 seconds) |
|
14:52
🔗
|
|
hook54321 has joined #jsmess |
|
14:54
🔗
|
|
i0npulse has joined #jsmess |
|
15:59
🔗
|
|
pengan has joined #jsmess |
|
16:12
🔗
|
|
pengan has quit IRC (Ping timeout: 633 seconds) |
|
18:15
🔗
|
|
i0npulse has quit IRC (Ping timeout: 252 seconds) |
|
18:16
🔗
|
|
hook54321 has quit IRC (Ping timeout: 252 seconds) |
|
18:43
🔗
|
|
i0npulse has joined #jsmess |
|
18:46
🔗
|
|
hook54321 has joined #jsmess |
|
20:39
🔗
|
|
Lord_Nigh has quit IRC (Quit: ZNC - http://znc.in) |
|
20:40
🔗
|
SketchCow |
OKAY WOW WHAT A DAY |
|
20:40
🔗
|
SketchCow |
Anyway, thing is uploading disks, thing is screenshotting vice, thing is uploading screenshots, thing is checking screenshots for "boring" |
|
20:41
🔗
|
SketchCow |
So, did we just become PC98 ready? |
|
20:41
🔗
|
|
Lord_Nigh has joined #jsmess |
|
20:42
🔗
|
SketchCow |
If so... we got a party waiting |
|
20:42
🔗
|
SketchCow |
https://archive.org/details/Pc-98SoftwareDump |
|
20:55
🔗
|
DFJustin |
would be best to use the neo kobe collection |
|
20:56
🔗
|
DFJustin |
https://archive.org/details/NeoKobe-NecPc-98012017-11-17 |
|
22:30
🔗
|
|
pengan has joined #jsmess |
|
22:51
🔗
|
|
pengan has quit IRC (Read error: Operation timed out) |
|
23:32
🔗
|
Sgeo |
I mentioned the VICE 3.2 JS stuff to Neohabitat Slack |
|
23:33
🔗
|
Sgeo |
Most useful for them would probably be hotkeys, Alt-N (flip disk) is essential |
|
23:46
🔗
|
SketchCow |
Agreed |
|
23:46
🔗
|
SketchCow |
It's always been so hard to do though |
|
23:46
🔗
|
SketchCow |
db48x: What's left to make dosbox-x work |
|
23:46
🔗
|
SketchCow |
And does the loader.js modifications the kid submitted make sense or give us something we need |
|
23:46
🔗
|
db48x |
yes, they're great |
|
23:47
🔗
|
db48x |
if you want it to work on IA, then we just need a json file, and a copy of the compiled dosbox-x |
|
23:47
🔗
|
SketchCow |
I did that |
|
23:47
🔗
|
db48x |
oh, excellent |
|
23:47
🔗
|
SketchCow |
There is now a dosbox-x.js.gz now |
|
23:47
🔗
|
SketchCow |
(Ripped it from the demo page) |
|
23:47
🔗
|
db48x |
:) |
|
23:47
🔗
|
SketchCow |
Touched a dosbox.cfg and put in emularity_config_v1 |
|
23:48
🔗
|
SketchCow |
Made my best attempt at a dosbox-x.json and flung it up on emularity_engine_v1 |
|
23:48
🔗
|
db48x |
awesome |
|
23:49
🔗
|
SketchCow |
If there's more we need I don't know it |
|
23:49
🔗
|
SketchCow |
If there's a wasm I couldn't find it |
|
23:56
🔗
|
SketchCow |
What did I get wrong, what else do we need |