#jsmess 2017-12-25,Mon

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)

WhoWhatWhen
bennett_eh?
no, we didn’t give permission for that - can you take it down?
[00:00]
SketchCowDFJustin: merry christmas [00:15]
.............. (idle for 1h9mn)
baihi bennett_
dunno anything about what SketchCow is talking about, but I hadn't heard of Multibowl before, looks pretty awesome
[01:24]
Vito`same, brilliant use of mame [01:29]
.... (idle for 19mn)
DFJustinmerry christmas [01:48]
..... (idle for 23mn)
bennett_thanks - just a misunderstanding about releasing it. It’s not release-ready
merry christmas also
[02:11]
baimerry christmas
how are you doing the custom win conditions, you just map the memory locations of each specific game and monitor them for specific attributes?
I've thought about doing some easter egg-like stuff like that with the web builds, like if you reach a certain score or reach a certain point of the game, you could trigger an event in the JS which is running the emulator, and do some change to the containing page
[02:11]
..... (idle for 23mn)
bennett_yeah I used the debugger to figure out where state is tracked in memory
and then we watch the memory to see when it changes
there’s a bunch of different rule definitions
[02:38]
baicool. I bet some of it is even documented in the form of game genie cheats [02:39]
SketchCowha ha
best misunderstanding ever
[02:44]
baiwell, it's the same underlying system, game genie and its ilk just worked by modifying the underlying memory locations, and people ran things through debuggers to figure out which memory locations mapped to things like whether you had a specific item, or what your score was
and MAME has abstracted that out into a cross-platform cheat system
but you're probably looking at values which aren't always useful for cheating
[02:46]
....... (idle for 30mn)
***azakai has quit IRC (Quit: Ex-Chat)
azakai has joined #jsmess
[03:17]
............ (idle for 59mn)
db48xwhat's a good system to test with wasm?
I remember there were some that were marginal
[04:17]
baijaguar's alway a good stress test [04:22]
db48xI'm only down to the C's [04:23]
baiahh, you''re doing a bulk build but want to test something that's already built, ok [04:24]
db48xI'll just build jaguar manually [04:25]
....... (idle for 30mn)
***bennett_ has quit IRC (bennett_) [04:55]
DFJustinc64 would be worth a try [05:03]
...... (idle for 29mn)
db48xI uploaded mamec64_wasm.js.gz and c64_wasm.json to go with it
I also uploaded mamejaguar_wasm.js.gz, but I don't have a json file for it yet
ok, I made a jaguar.json
they probably both need a bios image
[05:32]
baithere's a copy of Tempest somewhere on FOS
along with the machine rom
[05:36]
***Alpha64 has joined #jsmess [05:47]
..... (idle for 22mn)
DFJustinhttp://fos.textfiles.com/dfjustin/jaguar/jaguar/ [06:09]
baithat's old the asmjs build, right?
oh no, that's a wasm build
the % speed overlay isn't showing, should be f10 right?
[06:20]
DFJustinF11 [06:22]
baioh duh
oh and you have a link for it there anyway :D
[06:23]
runs around 80-100% on this system, but I was able to play it a lot further than usual before I got frustrated and gave up [06:31]
***Alpha64 has quit IRC (Quit: Alpha64) [06:39]
....... (idle for 31mn)
db48xoh, I can't upload to emularity_bios_v1
(I tried to upload the jaguar bios)
[07:10]
...... (idle for 29mn)
SketchCowdb48x: I'll get that fixed
db48x: Fixed
[07:39]
db48xawesome, thanks [07:42]
SketchCowLet me know anything else I can do.
So you're getting single-file loaders for wasm and js?
[07:42]
db48xSketchCow: another question for you: what do I need to do to simple-emulated-test-item to make it use the emularity interface? [07:44]
SketchCowSpell it emulator_ext and not emulaor_ext
Fixing it
[07:45]
db48xhah
spelling
[07:45]
SketchCowIt blows up, of course, loading WASM into javascript
Remember the &external_js=1 directive when testing
[07:48]
db48xor something [07:56]
..... (idle for 21mn)
SketchCowhuzzah [08:17]
.................................................. (idle for 4h7mn)
***JohnTalen has joined #jsmess [12:24]
JohnTalenDoes anyone familiar with IA know if there is a back door in it? By this can someone remote list directories/unmount drives without root password, stuff like this? [12:26]
db48xseems unlikely [12:38]
JohnTalendb48x: You've seen the source code for it? [12:41]
db48xonly some
why do you ask?
[12:45]
JohnTalenOnly some okay.
In the process of downloading I had my process kick out after my drive was unmounted, then remounted.
[12:46]
db48xah, well
if you're just downloading files from IA then you're not running any of the IA source code anyway
[12:53]
JohnTalenoh? how am I not? [12:55]
db48xwell, you're running wget, or curl, or a web browser
but they're just making an HTTP request and saving the response to a file
[13:03]
JohnTalenPerhaps if we could both see the source code to IA, we would know the winner of this debate. Is there a reason (besides the temporary effeciency in initial processing) that the source code for IA isn't distributed? [13:06]
SketchCow:)
This is an odd set of questions, john.
[13:07]
JohnTalenSketchCow: You can see from the IA logs why I am asking them. [13:08]
SketchCowWhen you say "IA Logs", what do you mean [13:09]
JohnTalenDon't get me wrong, what services you're providing is wonderful.
SketchCow: The profile information of my account. I was downloading Run Magazine and the drive I was backing up to unmounted, then remounted. This is Linux. That doesn't happen (tm).
[13:09]
db48xactually, it can happen
errors on the SATA bus can cause the drive to reset
[13:11]
JohnTalenPerhaps, but your assuming I'm using the SATA bus. Of course, errors can happen at any level on anything. Is it okay to get the souce code to IA?
source code.
[13:13]
db48xdb48x shrugs. SCSI and IDE drives can have similar problems [13:15]
SketchCowI guess there's about a dozen ways I could answer here. [13:15]
JohnTalendb48x: You're now assuming I'm using SCSCI or IDE. What doesn't have problems? [13:16]
SketchCowFunctionally, no, there is not a IA "source code" for you to get in this realm, so I can't help you there. [13:16]
JohnTalenah okay! [13:16]
SketchCowI can answer questions on the infrastructure.
But realize there is an entire set of, I believe, 8 "web-heads" connected to dozens and dozens of machines playing the role of processors, ingestors, and storage.
In three building locations.
Also, we have between 5-10 drive failures a week, like actual drives dying a sad death, quickly replaced from elsewhere in the infrastructure and data restored and backed up.
And that just I, me, Jason, I'm causing thousands of transactions of uploads and downloads a single day, and I'm just one user (albeit a heavy one) and we get 3 million users a day.
[13:16]
JohnTalenHere is my log: https://pastebin.com/LYm4SJgq
As you can see 207.241.228.14 is an IA server and the read occurred at my firewall breach attempt.
(with a 12 second difference)
jesus 3 million?!
Insane.
[13:21]
SketchCowI guess I don't quite know what I'm looking at and I'm confused by the term "firewall breach attempt"
Anyway. db48x - Good to see you still hackiing away
I had a drive nearly fill and I've got 5 windows cleaning up said drive, as well as mirroring multiple floppies
I mean CDs
[13:24]
JohnTalenSketchCow: thank you for your time. [13:28]
You might want to check your servers for sending out ICMP bursts to it clients. [13:37]
SketchCowHow's the vice work going.
db48x: Anything you need from me?
[13:38]
JohnTalenSketchCow: Excellent. I haven't worked on it for about 1 1/2 weeks. I'm at a point where it's drawing black to the canvas, but it is erroring out on allocation. My next step is to add more debugging information to it. It got mildly far into the code before this happened. [13:40]
baio_O
if your local system's log contains disk errors I suggest you look ito that separately of any questions about IA's infrastructure, heh
nothing IA's code could possibly do could conceivably cause your local disk to throw errors
[13:41]
JohnTalenbai: No problems with this drive since I've had it until a few bits ago.. and I have logs of a large variety of Internet Archive.org's service blasting my client with ICMP packets just before the I/O error -5 write on said drive.
bai: Is there a reason for more than 3 ICMP a minute? Is *should* just use my ESTABLISHED/RELATED connection until the download is down. ICMP is normally just used to keep the connection open.
[13:47]
baiwell, ICMP packets are a normal part of network traffic, the particular icmp packet referenced in that log is sent by a router somewhere, it's saying the destination is unreachable [13:49]
JohnTalenbai: While that is true, it's not normal for dozens of services from the IA domain to blast one little old client with them. [13:50]
baiit.....kind of is..... [13:51]
JohnTalenI mean I have the log. [13:51]
baiicmp happens, all the time [13:51]
JohnTalenYes, they do happen all the time, but my client was under attack. [13:51]
baithe fact that it happened 12 seconds prior to your local filesystem throwing a directory read error is almost certainly a coincidence, not IA hacking you :) [13:51]
JohnTalenEvery 4 - 12 seconds of ICMP is not normal from one domain copying one file at a time. [13:52]
bai*shrug* [13:52]
JohnTalenping is ICMP. [13:53]
baiping is one type of ICMP packet, yes [13:53]
JohnTalenDoes IA use an ESTABLISHED/RELATED connection for it's downloading process?
It's problably just an inefficiency process by either your servers or my client.
[13:53]
baithis is the specific ICMP packet you're seeing, and every possible reason a router might send you that message, according to various TCP/IP specs http://www.networksorcery.com/enp/protocol/icmp/msg3.htm
it's just a normal message that your system receives if there's a network error
[13:59]
JohnTalenICMP is used not by file transfer but by connection information retrieval. [14:00]
baiit's used by IP, file transfers happen over HTTP, which is a TCP/IP protocol [14:00]
JohnTalenbai: Dozens of IA servers, not just a few. It's called ineffciency.
I have it logged if you have interest in it.
[14:00]
baithis is internet :) [14:01]
JohnTalenbai: My question remains open then.
Here is the evidence of dozens of IA servers banging my machine before my hard drive write fails: https://pastebin.com/ELw4XFL0
But I do have limit burst of an average of 3 ICMP per minute a per normal.
If IA used an ESTABLISHED/RELATED connection as per normal it likely wouldn't of crashed. And dozens of machine? You're servers really want to know I'm here. hahaha.
[14:01]
baithe SRC on those packets is within your own network, they're coming from your router [14:05]
JohnTalenPerhaps I can fix the IA code for you.
bai: Yes, if you look closer you'll notice those are incoming signals to my gateway, then to end node (127).
[14:05]
bairight, which to me says that something on your .127 host is trying to establish connections and being refused [14:07]
JohnTalenthe 127 SRCs (if you look) are the ICMP wanting to respond. It's always a response from my servers. [14:07]
bainot that the other end is launching an all-out attack on you :) [14:07]
JohnTalenGreen is not blue. okay.
:)
[14:07]
baisome greens are blue [14:07]
JohnTalenI'm not saying it's an attack. it just looks like one. [14:08]
SketchCowHEY SO HOW IS THAT EMULATION [14:10]
JohnTalenAnyway should be a easy fix. Just take out all the ICMP that cause denial of service using dozens of IA servers that attack their clients. [14:11]
SketchCowbai: Did you see the db48x wasm work [14:11]
baiyeah, great stuff [14:11]
JohnTalenAs you can see, one serer tries to attack. and then after a long wait, dozens of IA servers follow suit. [14:11]
baithe jaguar build seems to be running relatively smoothly, it's still not hitting 100% during gameplay, just in menus, but it's a big improvement in playability [14:12]
JohnTalenI can fix it for you. [14:12]
db48xhow can the wasm work be great? they don't even run [14:16]
SketchCowIt's in a direction [14:18]
baiwell, dfjustin's jaguar wasm build ran well, I was a bit unclear about whether that was a new build or an old one, I assumed new [14:18]
SketchCowI was getting 95%-100% on that wasm on DFJustin's [14:18]
db48xhis was 186, mine is 192
his does at least run though
[14:18]
baiah ok, so that's not even the latest, good to know [14:18]
SketchCowIt'll be neat to try all the platforms after the code stuff
Moogs has a deep interest in the code changes now
[14:27]
On my side, well, I've got this backlog of software to upload
People are sending me thousand-disk collections now
[14:32]
...... (idle for 27mn)
In fact, it's overwhelming, just moving CD-ROMs people are uploading to the general software means hundreds [14:59]
baiit's like you're an alcoholic who's trying to quit, but people keep inviting you to booze-themed parties [15:07]
SketchCowWell, the processes for ingestion are very fast now though [15:12]
bai"it's okay, the processes for ingestion are very fast now - they've got a beer bong and they're filling it with liquor"
:D
[15:16]
db48xa few more discs can't hurt [15:17]
SketchCowfff you I'm keeping my keys [15:23]
baiI should probably go to bed
merry christmas, you crazy fuckers :D
[15:23]
SketchCowYeah what with that maniac [15:31]
***bennett_ has joined #jsmess [15:40]
......... (idle for 42mn)
i0npulse has quit IRC (Ping timeout: 248 seconds)
Rai-chan has quit IRC (Ping timeout: 248 seconds)
[16:22]
i0npulse has joined #jsmess
Rai-chan has joined #jsmess
[16:31]
............... (idle for 1h14mn)
db48x has quit IRC (Read error: Connection reset by peer) [17:46]
................ (idle for 1h17mn)
i0npulse has quit IRC (Ping timeout: 248 seconds)
Rai-chan has quit IRC (Ping timeout: 248 seconds)
Rai-chan has joined #jsmess
i0npulse has joined #jsmess
[19:03]
.... (idle for 18mn)
JohnTalenhey i0npulse. [19:23]
SketchCowI've made a floppysoftware collection. [19:30]
..... (idle for 24mn)
DFJustinniiiice
mind if I populate it
[19:54]
SketchCowSure
FOR the moment I'm just doing ones where they scanned the floppy
I intend to generate filler images with the floppy disk info in them
[20:05]
DFJustinI need to scan mine, that's something I was hoping to get into over the holiday break [20:06]
JohnTalenokay, nice place for ignorance. noted. [20:15]
***JohnTalen has quit IRC (Quit: leaving) [20:15]
SketchCowI got super tired of looking in the software stacks, I'm going to try and knock them down to something intelligent
Hi, gang.
I have no idea what is the story with John.
I assume life is rough so he's taking it out on the world.
Regardless, his Vice work was interesting to watch.
[20:18]
bai¯\_(ツ)_/¯ [20:25]
DFJustinI uploaded a fun christmas thing https://twitter.com/DopefishJustin/status/945169114493943809 [20:26]
bainice [20:27]
.... (idle for 18mn)
SketchCowI'm happy to say the manga nightmare is almost over.
almost.
[20:45]
baioh god, you didn't grab that pile from the side of the road when you were in japan, did you? :D [20:46]
SketchCowHa
No, someone gave me 4.4tb of manga
That's... a lot of manga
[20:46]
baiindeed [20:46]
***db48x has joined #jsmess [20:46]
SketchCowSomeone has given me hundreds of podcast episodes from 2004-2005
Another core sample saved!
[20:51]
............ (idle for 57mn)
***binji has joined #jsmess [21:49]
..................... (idle for 1h44mn)
binji has quit IRC (Read error: Operation timed out) [23:33]

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)