<sorear>
is "do something different from every other piece of software the wraps a tty" the goal here?
<Darius>
personally I would have tried ~. but I have used a lot of cu over the years..
<Darius>
just put the terminal in raw mode, no SIGQUIT then ;)
<whitequark[cis]>
<sorear> "is "do something different..." <- what do everyone else do
<sorear>
not a lot of consistency. ^a for qemu and screen, ^b for tmux, <newline>~ for ssh, ^^ for mosh, ^] for telnet, not sure about the serial port focused tools, emacs gets a few commitment to the bit points for setting VINTR to ^G while running
<Darius>
cu uses ~. like ssh
<Darius>
(or ssh uses it like cu to be more historically accurate)
<whitequark[cis]>
ohhhh, you meant to ask whether i want to make inconsistency slightly worse or leave it at the incredibly bad level it is
<Darius>
qemu really uses ^a? talk about violence against emacs users
<sorear>
I suppose the most consistent thing here is that all of the major ones have a configurable escape character. I wonder how much of this is colored by "tool X needs to use a different escape from tool Y so that the developer of tool X, who is primarily only considering their own needs, can run it under tool Y for debugging"
<whitequark[cis]>
OH
<whitequark[cis]>
I remember now why I was using SIGQUIT
<whitequark[cis]>
so that I could use SSH over the TTY
<whitequark[cis]>
so, yes, I think it actually is the goal
redstarcomrade has quit [Read error: Connection reset by peer]
<theorbtwo[m]>
Vaguely relatedly, I've noticed that the baud rate autobaud reports isn't quite what I expect. Is this likely to be Glasgow's clock being off, my hard drive's clock being off, the hard drive's implementor not caring all that much, or ...?
<whitequark[cis]>
the Glasgow's clock is accurate to less than 50ppm, I think usually within 20ppm
<theorbtwo[m]>
Right, that effectively rules it out.
<whitequark[cis]>
probably the hard drive has some weird frequency as its base, like 200M, and the baud is something like 38400 that doesn't properly divide it after a pre-divider
<whitequark[cis]>
so you get like a quantization error in the divider or something like that
<Darius>
theorbtwo[m]: is it a consistent error or random?
<theorbtwo[m]>
Consistent, I think. I've made the error of asking while not being in front of an easy way to poke at it. Will fix that shortly, I think.
<theorbtwo[m]>
I should stop saying"I think" so much, I think,
<Darius>
not very surprising it would be off IMO, as whitequark said
<Darius>
not many people putting xtals with frequencies derived from the NTSC colour burst in their designs anymore
<theorbtwo[m]>
I would have thought they'd only let me set rates they could get quite close to, though.
<Darius>
the general spec for a UART is 10% difference
<theorbtwo[m]>
It's one of those things I know doesn't matter, but bugs me anyway.
<theorbtwo[m]>
OK, I'm just a madman, the error maxes out at about 3.5%.
<whitequark[cis]>
did you look up the 16550 datasheet?
<theorbtwo[m]>
Not yet.
<josHua[m]>
🌌🧠: the UART applet should emulate a PCI 16550 connected with warp pipe
<whitequark[cis]>
cursed
<whitequark[cis]>
wait, warp pipe?
<josHua[m]>
https://github.com/antmicro/warp-pipe , though I thought there was a linux kernel driver for it to run natively, which makes my idea less useful
<theorbtwo[m]>
Ah. I have sometimes wondered if PCIe over ethernet could be a thing. After all, there's just about everything else over ethernet...
<galibert[m]>
So, pcie over http next?
<Darius>
damn and I thought USB over IP was cursed
<galibert[m]>
well, usbip is what I think I'm going to use to make Mame pretend it's a usb device
<whitequark[cis]>
huh
<josHua[m]>
I think I have expressed the cursed idea previously of a Glasgow applet that pretends to be a FTDI using usbip
redstarcomrade has quit [Read error: Connection reset by peer]
Wanda[cis] has quit [Quit: Idle timeout reached: 172800s]
fibmod has quit [Ping timeout: 255 seconds]
jstein has joined #glasgow
fibmod has joined #glasgow
redstarcomrade has joined #glasgow
fibmod has quit [Ping timeout: 268 seconds]
fibmod has joined #glasgow
<JanIbrahim[m]>
Hi all, what about the LVDS port ?
<whitequark[cis]>
what about it?
<JanIbrahim[m]>
How to use it ?
<whitequark[cis]>
it's not something we really support the use of; it's there and if you have the requisite skills and knowledge you can choose to use it at your own risk
<JanIbrahim[m]>
Ok
<whitequark[cis]>
among other things it's completely unprotected (no ESD protection, no overcurrent/overvoltage protection)
<JanIbrahim[m]>
Watching the main video clip on Crowdsypply site and on the sec 00:20 there is a EPROM adapter ( HAT) that fits on Glasgow , is there any way to get this HAT ?
<JanIbrahim[m]>
And examples to use EPROM too?
fibmod has quit [Ping timeout: 264 seconds]
fibmod has joined #glasgow
<whitequark[cis]>
<JanIbrahim[m]> "Watching the main video clip..." <- esden: is that a product you have planned to put in stock or just a prototype?
<JanIbrahim[m]>
Prototype , just for fun to revive my 27c16 eprom!
<whitequark[cis]>
sorry that was a question to Esden
<whitequark[cis]>
I'm the project founder but I am not running sales or marketing, so I'm not sure of all the details
<JanIbrahim[m]>
Ok przepraszam , nie zauważyłem
Wanda[cis] has joined #glasgow
<Wanda[cis]>
there are plans to use the LVDS connector for addons specifically targetted for glasgow, like the ram-pak, but so far it's just a prototype and only a single-digit number of those exist
<Wanda[cis]>
if you just want to connect random stuff to glasgow, you should strongly prefer using ports A/B, since they are properly protected and can have arbitrary-ish VCCIO
<JanIbrahim[m]>
How to manage the voltages 12 v and 25 needed to program the eprom ?
<Wanda[cis]>
... oh, that kind of eprom
<Wanda[cis]>
yeahhh that needs some external circuitry
<Wanda[cis]>
glasgow can only supply up to 5V
<JanIbrahim[m]>
Yes that’s why I asked this question!
<whitequark[cis]>
Jan Ibrahim: glasgow only directly supports programming for 5V or 3.3V single supply PROMs normally, I think the device you see on the promo video does as well
<whitequark[cis]>
I don't even have any of the high voltage programming ones to test with tbh
<JanIbrahim[m]>
Ok
<Wanda[cis]>
heh
<Wanda[cis]>
I'd be kinda interested in playing with it some day
<Wanda[cis]>
though not PROMs, more like make glasgow + some addon into a GAL programmer
<whitequark[cis]>
yeah
<Wanda[cis]>
(also, fun fact: JTAG mode aside, Xilinx CPLDs can be programmed via external VPP as well, using some weird parallel protocol)
<whitequark[cis]>
fascinating
<Wanda[cis]>
curiously, this applies to both the Plus Logic-derived ones and Philips ones, suggesting it may be more universally applicable
esden[m] has joined #glasgow
<esden[m]>
EEPROM add on is on my todo list but RAM-Pak is higher priority for me at the moment. The EEPROM add on I made is fairly simple and the design files are on github if anyone wants to make one for their own use.
fibmod has quit [Ping timeout: 268 seconds]
fibmod has joined #glasgow
<JanIbrahim[m]>
Nice , thank you 👍
redstarcomrade has quit [Read error: Connection reset by peer]
fibmod has quit [Ping timeout: 256 seconds]
redstarcomrade has joined #glasgow
fibmod has joined #glasgow
fibmod has quit [Ping timeout: 268 seconds]
fibmod has joined #glasgow
<theorbtwo[m]>
So... I wanted a way to, from the UART applet, to control an extra pin to turn on and off an ATX PSU powering the device under test. https://github.com/theorbtwo/glasgow/tree/uart-extra-pins is my ham-handed attempt at that. I'm afraid it's baby's first HDL project, and... it doesn't work, and I don't really know how to debug it. I've verified with a multimeter, and the pin in question seems to stay floating. I'm reasonably secure
<theorbtwo[m]>
that the UI stuff is working as intended, but I'm not so certain about the more hardware / amaranth side of things.
redstarcomrade has quit [Ping timeout: 272 seconds]
redstarcomrade1 has quit [Ping timeout: 268 seconds]
<castaway[m]>
time to sleep on it (though if anyone has a brainwave, do share)
<castaway[m]>
btw am sat 1 desk to theorbtwo 's right ;)
redstarcomrade2 has quit [Quit: Cya.]
redstarcomrade has joined #glasgow
redstarcomrade has joined #glasgow
<chipb>
theorbtwo[m]: going out on a limb because I've never written amaranth or a glasgow applet, but should that be `target.m.d.comb` instead?
<whitequark[cis]>
none of that, in fact
<whitequark[cis]>
you don't ever manipulate m like that
<whitequark[cis]>
it's special syntax, not a variable for you to store wherever you feel like
<whitequark[cis]>
that's kind of the reason it's broken
<chipb>
yeah, that seemed a bit fishy too given the magic going on with Module.
<whitequark[cis]>
it's not actually magic
<whitequark[cis]>
it's just that the Module is modified after it's generated and the modification goes nowhere, I think
<chipb>
probably a stupid question as you've probably considered it already, but since it's almost certainly a mistake, I don't suppose it's possible to make a Module instance explode if modification is attempted after generation?
<whitequark[cis]>
i actually haven't; this is just not a common mistake, only beginners make it