<supersat[m]>
the part where randos can get my computer to make arbitrary HTTP requests is kinda iffy tho
<whitequark[cis]>
this is soooooo heckin cool
<whitequark[cis]>
i never really expected anyone to use that applet except for me
<whitequark[cis]>
and seeing it in action like that? amazing
<whitequark[cis]>
supersat[m]: could add a command line option default-off maybe?
<whitequark[cis]>
as in, defaults to not providing that feature
<supersat[m]>
well it works! the only really confusing thing was the whole pin set thing. not sure if it was supposed to be 0-7 or 7-0 or A7-A0 or whatnot
<whitequark[cis]>
replace all the FTDI stuff with what's in --help text for jtag-openocd
<gojimmypi[m]>
I think the other reason I boight the Glasgow, it just for the FPGA. It's like buying a blank canvas for art. The possibilities are endless.
<gojimmypi[m]>
erg, except in this case, I need to be able to put back the original picture... the as-shipped RTL. 😉
<whitequark[cis]>
so Glasgow is actually a pretty bad general purpose FPGA devboard
<whitequark[cis]>
it's really not intended or designed for standalone use, nor does it enable it in any meaningful way
<whitequark[cis]>
sure, you can hack it to do so, always
<whitequark[cis]>
but it's not an FPGA devboard.
<gojimmypi[m]>
that's so mind bending, I'm still trying to understand that.
<gojimmypi[m]>
that's good to know.... but curious as to why?
<whitequark[cis]>
glasgow is designed for vertical integration because this is the only way in which i can provide the smooth experience i want to
<whitequark[cis]>
it's similar to why apple products are vertically integrated, except i'm not actually hiding anything
<whitequark[cis]>
it's developed in the open and is a bunch of reusable components, just the components go very well together and are not (and will not) be primarily designed for individual use
brolin has joined #glasgow
<whitequark[cis]>
i don't believe in things like "the unix philosophy"
<whitequark[cis]>
i believe in whole system design driven by cohesive principles on each scale and level of abstraction
<duskwuff[m]>
the most obvious "not a dev board" factors I see are that 1) the board has a bunch of hardware in between the FPGA and (most of) the IO pins which requires some nontrivial config to use (like configuring DACs); 2) the firmware running on the fx2 isn't a generic FPGA interface, it's going to expect the glasgow gateware to be there to talk to. if you want an ice40 eval board there's plenty of cheaper, more general-purpose options out there
<supersat[m]>
VGM files are just logs of register writes. The applet simply replays these register writes over a bus to a real Yamaha chip, and receives synthesized digital samples back. Then there's a web wrapper over all that, plus some client-side magic to play it back in the browser despite the wonky sample rates these chips are usually driven at
<supersat[m]>
honestly I haven't looked at how the whole thing works, but I'm fairly certain that's accurate
<whitequark[cis]>
the FX2 firmware is actually reasonably generic, it's just that we don't want to support people doing non-glasgow software because the scope is too much
<supersat[m]>
For some reason I just so happened to have a YM3812 sitting around on a breadboard from AGES ago (pulled from an old sound blaster card??) and so the existence of that applet made it the perfect thing to try my new Glasgow with 😂
<_whitenotifier-f>
[GlasgowEmbedded/glasgow] github-merge-queue[bot] f55645c - Deploying to gh-pages from @ GlasgowEmbedded/glasgow@c424d795e202ed45609b7795c9bb686daf59e7cb 🚀
<supersat[m]>
@_discord_1135453280001073254:catircservices.org (@_discord_1135453280001073254:catircservices.org) would it be useful to document how to get the Yamaha applet running in the manual? Or too esoteric? Also wondering what the scope of the glasgow repo is. For example, I'm thinking about adding both an interface to various video game consoles and T1 line interface ICs 😛
redstarcomrade has joined #glasgow
redstarcomrade has quit [Changing host]
redstarcomrade has joined #glasgow
<supersat[m]>
also would something like exposing an audio USB class device as a way to interface with I2S be in scope?
omnitechnomancer has joined #glasgow
<omnitechnomancer>
That's neat, listened to stan's theme from monkey island 2 just now 😄
<supersat[m]>
try it at 2.3 volts now 😂
<omnitechnomancer>
oh my that is glitchy
<supersat[m]>
ahaha
<omnitechnomancer>
I wonder what is actually glitching
<whitequark[cis]>
<supersat[m]> "also would something like..." <- Glasgow does not change how it presents itself on USB
<whitequark[cis]>
wrong type of device
<whitequark[cis]>
<supersat[m]> "@_discord_1135453280001073254:..." <- T1 is in scope, video game consoles are quite possibly in scope
<supersat[m]>
can't you change the FX2 firmware for that?
<whitequark[cis]>
theoretically yes, in practice it would be a support nightmare
<whitequark[cis]>
so we choose not to
<supersat[m]>
aha!
<supersat[m]>
okay
<supersat[m]>
I had this crazy idea of exposing T1s as 23x2 audio channels 😛
<supersat[m]>
plus a control channel
redstarcomrade has quit [Quit: Cya.]
<whitequark[cis]>
hehe, yeah, that could work... but i doubt host audio stack will like it anyway
tomkeddie[m] has joined #glasgow
<tomkeddie[m]>
That sounds more like luna/cynthion territory perhaps?
<whitequark[cis]>
yeah
<supersat[m]>
that was the original plan 🙂
<supersat[m]>
iso endpoints are a PITA tho
<supersat[m]>
at the very least, there's no examples for how to build one in luna
<supersat[m]>
whereas cypress might provide some hints? idk
<supersat[m]>
my current plan is a common T1 LIU interface that can be used by both projects, but consumed in different ways
<supersat[m]>
I think I could probably put a full telephone switch on a luna/cynthion, for example, with a softcore and some dedicated switching logic
<supersat[m]>
while the glasgow interface would be purely an interface adapter
<whitequark[cis]>
right
joerg has quit [Ping timeout: 258 seconds]
trh has quit [Quit: weg]
trh has joined #glasgow
joerg has joined #glasgow
<_whitenotifier-f>
[glasgow] github-merge-queue[bot] created branch gh-readonly-queue/main/pr-434-c424d795e202ed45609b7795c9bb686daf59e7cb - https://github.com/GlasgowEmbedded/glasgow
<_whitenotifier-f>
[GlasgowEmbedded/glasgow] attie 879783e - meetings: add minutes from 2023-10-07
<whitequark[cis]>
<supersat[m]> "while the glasgow interface..." <- sounds good! could expose an audio device in the OS / python code for example
<whitequark[cis]>
pulseaudio lets you connect audio via rtp/tcp/etc
<supersat[m]>
The Saturn Action Replay interface feels like a natural fit because 1) it's 5V, 2) it is 1:1 bidirectional, 3) it has handshake signals, 4) transfers data a nibble at a time, and 5) the upper layer protocol is essentially a synchronous stream of bytes
<_whitenotifier-f>
[GlasgowEmbedded/glasgow] github-merge-queue[bot] e4efe61 - Deploying to gh-pages from @ GlasgowEmbedded/glasgow@879783e202c6d839b91c303e51be653b03836b3b 🚀
<supersat[m]>
if you're asking about the Action Replay, it's a cartridge that primarily acts like a "Game Genie", which patches RAM on every vertical blank, but you can also reflash the cart with new codes (or firmware) and upload/download RAM and optionally execute it via a PC interface, which only exists as an ISA card. I used it as a kid to try out some Saturn dev tools and examples that Sega of America accidentally leaked 😬
<supersat[m]>
Sega also has some weird controllers for the Dreamcast that are exposed over their "Maple Bus" like a keyboard, microphone, etc. Could emulate controllers, VMUs (memory cards with a screen and the ability to execute apps on-device), etc
<eigenform>
i cant tell if this is a bug or me being a little clueless: i tried using '--sck-idle 1' with spi-controller, but it doesn't look like the clock polarity is flipped?
<eigenform>
also hi :)
lle_bout[m] has quit [Quit: Idle timeout reached: 172800s]
brolin has joined #glasgow
brolin has quit [Ping timeout: 240 seconds]
brolin has joined #glasgow
<gojimmypi[m]>
I wonder if it would be better to have the human conversations in separate channel from the GitHub bot message?
brolin has quit [Ping timeout: 255 seconds]
<josuah>
it looks a bit less intrusive in IRC where the links are not expanded into big blocks
<esden[m]>
supersat (@_discord_342359737444204544:catircservices.org) the issue is also that if you ignore the bot you will not see what people are saying on the other platforms. So that is not helpful. One can disable url preview in the discord client. But that is for _ALL_ channels which is not helpful.
<supersat[m]>
Yeah I realized all the different "bots" have the same internal user number
<esden[m]>
I am also in the camp that would prefer the Git notifications to go to a dedicated channel and not disrupt human conversations. But I am not sure there is a way to fix that unless we create another group of git notification channels. That said, I would not be opposed to creating a dedicated vcs-glasgow channel on the discord server here, as I have it for black magic and icebreaker. Similar thing could be done on Matrix I would assume. But
<esden[m]>
Catherine used to have the in channel notifications for git since ages, so it might be her preference to have the stuff in the "feed"
<esden[m]>
it is one bot... that does the bridging
<esden[m]>
the commit messages are just relayed by the bot
<esden[m]>
just like any other message
<esden[m]>
it is coming from the IRC channel
<supersat[m]>
I would rather not interrupt developers' existing workflows
<supersat[m]>
Er, disrupt
<esden[m]>
supersat (@_discord_342359737444204544:catircservices.org) same 🙂
<whitequark[cis]>
im running a fork of notifico already; might as well hack that to silence merge queue studf
<whitequark[cis]>
s/studf/stuff/
<whitequark[cis]>
however don't expect me to have time for it any time soon, or ever
<whitequark[cis]>
thus far people here like to complain but do not like to set up, fix, and test the annoying infrastructure like matrix bridges and such; resulting in the infrastructure continuing to annoy them
<whitequark[cis]>
which to me feels fine; if you want to contribute to my projects you should have at least my level of willingness to maintain absolutely anything~
<whitequark[cis]>
the notifico instance needs to be moved to catircservices.org, at which time I might actually look into that (since that makes it a lot less annoying to deploy and test)