Time |
Nickname |
Message |
00:01
🔗
|
Sgeo_ |
AFK |
00:43
🔗
|
SketchCow |
1. Ha ha, IA upload queue for my admin account buffered up too high |
00:43
🔗
|
SketchCow |
2. Screenshotting... sorry, DFJustin - gotta keep it to png, hilarious reason |
00:44
🔗
|
SketchCow |
I mean jpg |
00:45
🔗
|
SketchCow |
PNGs are, and I can prove this, being screenshotted with imagemagick such that the md5 is different every single shot |
00:45
🔗
|
SketchCow |
No idea why, can't deal for now |
00:47
🔗
|
bai |
maybe png header info with a timestamp? |
00:48
🔗
|
SketchCow |
Yeah, some such |
00:48
🔗
|
SketchCow |
Regardless, not dealing for now |
00:49
🔗
|
SketchCow |
https://ia601505.us.archive.org/0/items/00101010_2003-03-02_Creators/00101010_2003-03-02_Creators_screenshot.gif |
00:59
🔗
|
bai |
trippy |
01:25
🔗
|
Sgeo_ |
Does the Ubuntu Software thingy seriously not have a search function? |
01:29
🔗
|
zino |
It does, in 18.04 it's a magnifying glass at the top. For other versions it's probably similar. |
01:31
🔗
|
Sgeo_ |
Thank you, inexplicably part of the title bar. Although didnt help in this case (looking for a graphical svn client. Found one elsewhere) |
01:31
🔗
|
Sgeo_ |
For some reason searching svn brings up nothing and searching subversion brings up a KDE thing, I'd rather use GNOME stuff or neutral stuff |
01:43
🔗
|
zino |
Back when I still used svn turtoise-svn was the go to client if you wanted a GUI. Doesn't look like it's part of the default repos any longer. Most things svn is unmaintained and on the way to be phased out. |
01:54
🔗
|
Sgeo_ |
bai, is there a config to have something yell at me if the code gets into a loop? |
02:08
🔗
|
Sgeo_ |
Going to review how much effort it would be to just fix the loop the "right" way. I mean, VICE.js person used a coroutine library to do it, I hope that that's not really necessary and the person was just being lazy... but a coroutine library may have been more effort, hmm |
02:14
🔗
|
Sgeo_ |
I thought this being not-work means I wouldn't come across painfully bad code. I was wrong.... there's an #include "6510core.c" in the middle of the mainloop function |
02:15
🔗
|
Sgeo_ |
Maybe I shouldn't rush to criticise |
02:16
🔗
|
Sgeo_ |
...my current idea involves making the code even more eye-bleedingly bad by throwing static in front of a bunch of variables |
02:21
🔗
|
bai |
Sgeo_: hmm. normally you can tell because it just hard-locks the browser, including the UI...there *might* be an option to force terminate after some period but I can't recall |
02:22
🔗
|
Sgeo_ |
bai, yeah, that's what I've been seeing. Was hoping for Emscripten warning that would magically tell me what specifically was locking up >.> |
02:23
🔗
|
Sgeo_ |
When I terminate it I get a VICE error about unable to initialize SDL sound... even when I tried to prevent that from compiling |
02:23
🔗
|
bai |
there's the dreaded SimulateInfiniteLoop exception, which can give helpful info sometimes but I'honestly not really sure when that ends up being thrown, vs just silently chewing up all the cpu cycles |
02:23
🔗
|
Sgeo_ |
Oh that reminds me of an error DOSBox has been showing for one of the items I uploaded |
02:25
🔗
|
Sgeo_ |
https://archive.org/details/stop-the-rock-live2 (big download) |
02:25
🔗
|
bai |
for dosbox, I believe we have two different emulators, one of which has more heavy-handed use of emterpretify, which can be useful if the app's main executable spawns other processes |
02:25
🔗
|
bai |
dosbox vs dosbox-sync I think? |
02:26
🔗
|
Sgeo_ |
lol the screenshotter got the mouse cursor in it https://ia800802.us.archive.org/8/items/stop-the-rock-live2/STOPROCK_screenshot.gif |
02:30
🔗
|
Sgeo_ |
CHange I made is now faster at drawing a black screen before dying in flames |
02:32
🔗
|
bai |
archive.org still seems quite slow for large files :( |
02:33
🔗
|
bai |
like 100-200 K/s at times |
02:44
🔗
|
Sgeo_ |
Yep, dying on Uncaught SimulateInfiniteLoop |
02:45
🔗
|
Sgeo_ |
With dosbox-sync |
04:01
🔗
|
Sgeo_ |
VirtualBox is refusing to screenshot properly |
04:02
🔗
|
Sgeo_ |
https://i.imgur.com/G8HrIRq.png that log I added, "Beginning sdl_init", doesn't show up until after I stopped a long running script |
04:22
🔗
|
bai |
hmm. |
04:22
🔗
|
bai |
I'm assuming we're not dealing with threads here, right? |
04:27
🔗
|
Sgeo_ |
Ah crud, is that going to be additional difficulty? |
04:27
🔗
|
Sgeo_ |
I think it might be possible to disable threading |
04:28
🔗
|
bai |
oh, yes, you'll definitely want to do that |
04:28
🔗
|
bai |
it may be doing that already, if configure checks for -lpthreads |
04:32
🔗
|
Sgeo_ |
...There's a function called usleep |
04:32
🔗
|
Sgeo_ |
I guess I should figure out how to fix that |
04:55
🔗
|
Sgeo_ |
Trying disabling UIthreads |
05:18
🔗
|
bai |
oh yeah, usleep is one of the functions you want to either eliminate, or emterpretify everything in the call stack above it |
05:22
🔗
|
Sgeo_ |
I... tried, it's not using that one |
05:22
🔗
|
Sgeo_ |
It's like it's using a builtin one |
05:22
🔗
|
Sgeo_ |
Shouldn't the builtin one that Emscripten provides complain if it's not emterpreted properly? |
05:23
🔗
|
Sgeo_ |
RIght now trying not calling it by tearing out the SID chip completely (no sound) |
05:26
🔗
|
Sgeo_ |
Well, now it's not hard dying. I'm getting a black screen, maybe a debug statement I added a while ago is slowing things down |
05:29
🔗
|
Sgeo_ |
No error messages, just a black screen |
05:33
🔗
|
Sgeo_ |
No, it eventually froze. Took a while. Now it's the video stuff that wants to sleep |
05:33
🔗
|
Sgeo_ |
I, uh, can't really remove that |
05:33
🔗
|
Sgeo_ |
bai, why isn't Emterpretify automagicing usleep stuff? |
06:08
🔗
|
Sgeo_ |
...there's a Javascript implementation in here, that says it's doing a busy loop |
06:20
🔗
|
|
hook54321 has quit IRC (Ping timeout: 252 seconds) |
06:23
🔗
|
|
hook54321 has joined #jsmess |
06:24
🔗
|
db48x |
Sgeo_: yes, that's the only way to implement sleep/usleep in javascript |
06:24
🔗
|
Sgeo_ |
I was vaguely under the impression using emterpreter would do some sort of setTimeout shenanigans |
06:24
🔗
|
db48x |
a C program can rely on the OS to set a timer interrupt and resume execution of the program so that sleep/usleep can return |
06:25
🔗
|
Sgeo_ |
I feel like maybe Emterpreter isn't being activated properly |
06:25
🔗
|
Sgeo_ |
Does it need to be enabled for all c compilation? |
06:25
🔗
|
db48x |
to be honest I've no idea |
06:25
🔗
|
Sgeo_ |
(during emmake?) |
06:26
🔗
|
db48x |
https://github.com/kripken/emscripten/wiki/Emterpreter says yes |
06:26
🔗
|
db48x |
"To use the Emterpreter, build with -s EMTERPRETIFY=1." |
06:27
🔗
|
Sgeo_ |
Was doing that with emc |
06:27
🔗
|
Sgeo_ |
I'll try adding to the makefile soon |
06:28
🔗
|
|
hook54321 has quit IRC (se.hub irc.underworld.no) |
06:28
🔗
|
|
dboard420 has quit IRC (se.hub irc.underworld.no) |
06:28
🔗
|
|
mx has quit IRC (se.hub irc.underworld.no) |
06:28
🔗
|
|
i0npulse has quit IRC (se.hub irc.underworld.no) |
06:37
🔗
|
|
db421 has joined #jsmess |
06:44
🔗
|
|
db421 is now known as dboard420 |
07:01
🔗
|
bai |
I can't recall if emterpretify is a compile-time or a link-time option |
07:05
🔗
|
Sgeo_ |
I'm going to try something silly, even though I told myself this is the time I would go to sleep |
07:10
🔗
|
Sgeo_ |
I replaced usleep in the js with a do nothing function |
07:10
🔗
|
Sgeo_ |
Still getting errors: |
07:10
🔗
|
Sgeo_ |
Error: Script terminated by timeout at: |
07:10
🔗
|
Sgeo_ |
emterpret@http://localhost:8000/x64.js:26912:1 |
07:10
🔗
|
Sgeo_ |
When I press Stop when it freezes up |
07:10
🔗
|
Sgeo_ |
Bedtime |
07:13
🔗
|
|
arkiver has quit IRC (Ping timeout: 607 seconds) |
07:17
🔗
|
|
arkiver has joined #jsmess |
07:19
🔗
|
|
datajerk has quit IRC (Ping timeout: 606 seconds) |
07:26
🔗
|
Sgeo_ |
Huh. The coroutine stuff in vice.js is basically a preprocessor trick |
07:27
🔗
|
Sgeo_ |
A bunch of stuff you can throw into a function to turn it into a state machine, I think |
07:39
🔗
|
|
datajerk has joined #jsmess |
07:41
🔗
|
Sgeo_ |
Tomorrow will manually look at the .patch and start applying relevant looking stuff. Didn't have luck attempting to automatically apply |
07:57
🔗
|
|
mx has joined #jsmess |
07:58
🔗
|
|
i0npulse has joined #jsmess |
08:42
🔗
|
SketchCow |
Screenshots |
08:42
🔗
|
SketchCow |
Going well |
08:42
🔗
|
SketchCow |
Sgeo_: I think that's a good plan |
12:41
🔗
|
|
zino has quit IRC (Quit: Leaving) |
12:42
🔗
|
|
zino has joined #jsmess |
14:13
🔗
|
* |
db48x sighs |
14:13
🔗
|
db48x |
https://github.com/db48x/emularity/issues/39 |
15:47
🔗
|
bai |
is that....bad? |
17:05
🔗
|
db48x |
dunno |
18:43
🔗
|
SketchCow |
I saw that, because I watch over everything, like Jesus |
18:43
🔗
|
SketchCow |
And I thought it was odd because it's not an issue, and I'm not sure how netplay works with it, if it works with us, then it works with pastebin |
18:45
🔗
|
SketchCow |
http://teamarchive1.fnf.archive.org/DOSSCREENS/ now works |
18:45
🔗
|
SketchCow |
It'll reset often, as I stop by to use it for QA and clear out ones that died or went bad |
21:20
🔗
|
Sgeo_ |
Why would SDL_EventState be a problem |
21:20
🔗
|
Sgeo_ |
Ugh I feel unfocused |
21:25
🔗
|
Sgeo_ |
The VICE.js patch has a wrappre for a necessary reason but the comment explaining why is wrong |
21:25
🔗
|
Sgeo_ |
(I think) |
22:11
🔗
|
Sgeo_ |
....Is everything burning because the VICE error message box is waiting for input???? |
22:26
🔗
|
Sgeo_ |
https://i.imgur.com/f0rLGle.png getting here (eventually) with no error message about the script slowing the page down |
23:02
🔗
|
bai |
could well be, that looks like exactly the type of simplistic blocking modal dialog which would lead to a a locked browser |
23:05
🔗
|
Sgeo_ |
(I removed the loop that checks for input in the UI. Probably not ideal) |
23:57
🔗
|
Sgeo_ |
Doing a checkout on my Windows machine where maybe I won't get frustrated at my tools |
23:57
🔗
|
Sgeo_ |
Going to apply the patch (or at least most of it, going to skip the stuff that messes with the keyboard if I can), copy over to Linux, try to build |