<asjackson>
i guess a lot of the setups where you'd be running multiple applets are probably things you might be working on quite often. could be quite nice if you could do `glasgow multi --config blah.whatever`
<whitequark[cis]>
now that would be quite a ways off (unless you mean like using @options to read command line options from a file, that's easy enough to add)
<asjackson>
but as a cli i think ++ works pretty nice
<whitequark[cis]>
however you can always just use HardwareAssembly to instantiate absolutely anything you want
<asjackson>
yeah i guess using argparse at the moment makes that a bit tricky
<whitequark[cis]>
in a script
<asjackson>
yeah
<whitequark[cis]>
you can instantiate *Interface directly
<asjackson>
i need to think of something to try with it, but i will definitely have a play once it's ready D:
<asjackson>
* :D
<whitequark[cis]>
note that you can only really use two simultaneous applets right now
<asjackson>
whats the behaviour with applets which exit once they've finished, when used along side long running things like openocd and uart?
<asjackson>
ah actually you'd probably rarely do that haha
<whitequark[cis]>
it just waits for all of them to finish
<asjackson>
it sort of works, but i can't boot unix with that RAM installed. i need to spend some time looking at it, maybe i can add support for DMA
<whitequark[cis]>
very nice
<asjackson>
jtag-probe makes me so happy by the way haha
<whitequark[cis]>
-probe or -pinout?
<asjackson>
ah pinout, yeah!
<asjackson>
been a while since i did that
<whitequark[cis]>
yeah it's super handy
<asjackson>
the level of confusion i can get myself in when hand wiring something is crazy (hence the very odd pinout in the gpib applet at the moment :P)
<asjackson>
although the cable i was using (and cut in half) didn't help much with that. the handshaking lines are supposed to be twisted with a ground, but they were randomly twisted with eachother. good quality cable....
<rabagliati[m]>
Using spi-analyzer between a Rpi and FPGA - buffer sizes 24 bit and 1024 bit. Only 3Mhz (Rpi insists on being controller) - works great, I appreciate getting a file dump of the transaction
<whitequark[cis]>
very nice, I'm always happy to hear field reports
<whitequark[cis]>
we ought to have a better I2C analyzer too
<_whitenotifier-5>
[glasgow] github-merge-queue[bot] created branch gh-readonly-queue/main/pr-832-37a121cd022b38b7a8b38d2266ca1cd4e8be0a4a - https://github.com/GlasgowEmbedded/glasgow
<whitequark[cis]>
yes, except it's not going to be a limit for much longer: I'm going to reimplement the USB comms code with nusb and PyO3, which should make it possible to use complex flow control mechanisms needed for multiplexing
<gruetzkopf>
i still owe you a few performance measurements on uncommon-but-not-totally-absurd platforms
<whitequark[cis]>
what I want is a high-level API where the Rust code exports an interface that is per-applet-pipe, not per-USB-whatever
<whitequark[cis]>
so the Rust code doing COBS encoding/decoding somewhere off the main Python thread is a key part of why I even bother with this refactor
<whitequark[cis]>
having the main Python async loop respond to every URB is unsustainable for higher data rate applets like qspi-analyzer even with minimal amount of processing, it's basically impossible to structure our library code in such a way where there is a straightforward way to write your applet code without running into overflows
<fridtjof[m]>
<whitequark[cis]> "fridtjof: okay, i've implemented..." <- wow, thank you very much :) I was previously looking at having to desolder the flash because I couldn't read it out in-system. I'm spared from that now :D
<fridtjof[m]>
<whitequark[cis]> "there is a... FAT image inside..." <- yup, for the web management interface assets i'm assuming
<whitequark[cis]>
fridtjof[m]: very happy to hear that this applet helped you that much
<whitequark[cis]>
desoldering flashes is a pain, I hate having to do it
<fridtjof[m]>
hence why this switch has just been sitting there, open with test hooks attached, for roughly a month
<fridtjof[m]>
fully agree
<whitequark[cis]>
(also, the dump does have your configuration, including an $1$encrypted password)
<fridtjof[m]>
oops, yeah :D
<fridtjof[m]>
it might actually just be "admin", not sure if i ever changed the password
<fridtjof[m]>
ah, yeah
<fridtjof[m]>
just admin:admin
<_whitenotifier-5>
[glasgow] github-merge-queue[bot] created branch gh-readonly-queue/main/pr-838-3e8c35391a59713ce36da6b2ac84cf8f380da0a0 - https://github.com/GlasgowEmbedded/glasgow
<galibert[m]>
Cat: it was supposed to be a joke, too bad it landed flat
<gruetzkopf>
it is a 3.5" disk, but that's an aftermarket mod. machine came with a 5.25mm drive. the manual for the FDC says it supports 8" and 5.25" disks
<gruetzkopf>
ill run downstairs and see how the FDC card is jumpered
<galibert[m]>
I guess it’s a 3.5 that pretends to be an 8
<gruetzkopf>
if i interpret this right, yep this is Single density
danieli has quit [Quit: Ping timeout (120 seconds)]