mcc111[m] has quit [Quit: Idle timeout reached: 172800s]
<esden[m]>
Ok... I made a thing: https://glasgow.1bitsquared.com (yes it is not very pretty, I am not a web designer unfortunately... but it is a start 😅 )
pelrun[m] has joined #glasgow
<pelrun[m]>
got my glasgow shipping notification, yay! It's going to my old address, aaaarrrgh
<esden[m]>
If someone has better CSS skills and idea how to make it prettier I would really appreciate some help. 😄 If you want to see my code and burn their retinas out with it you can find it here: https://github.com/esden/glasgow-cs-analytics 😄
<esden[m]>
pelrun (@_discord_301993647123988481:catircservices.org) congrats and I am sorry 😦
<pelrun[m]>
no-one's fault but my own, but it doesn't help that crowd supply maintains separate addresses for every pledge, and I had other CS packages arrive around the time you sent out the notification in march :p
<josHua[m]>
this I can do tonight, it is the least I can do. I don't really have great skills but I know how to, like, import bootstrap
<jof__[m]>
Same 😭 , but lessons learned: analog lab PSUs are not super trustworthy, make sure you touch the right dial! Also, the Glasgow protections are not infinite 😆
<jof__[m]>
Those level shifters and ESD protections for port A look pretty toasty (and maybe that PU/PD?), but I'm left wondering that if it isn't even showing up on USB that it maybe fried too many parts to even consider fixing.
<jof__[m]>
Since I'd imagine the FX2 would sit much further upstream and hopefully work even if the frontend got zapped
<esden[m]>
I guess it was a bit of a jolt of voltage that things did not have time to give way and protect the core circuits
<whitequark[cis]1>
it might just have a short on 5B
<whitequark[cis]1>
s/5B/5V/
<whitequark[cis]1>
I'd try to repair it
vegard_e[m] has joined #glasgow
<vegard_e[m]>
some parts tends to fail short and bring down the supply
<whitequark[cis]1>
first remove the dead parts
<whitequark[cis]1>
then run selftedt
<whitequark[cis]1>
* then run selftest
<vegard_e[m]>
so it might not be enumerating because the microcontroller is simply not getting power
<jof__[m]>
I suppose its worth trying to remove those and see what happens. Couldn't make it worse! :p
<jof__[m]>
thanks for the dose of confidence
<whitequark[cis]1>
you might also be able to use some of the spare pins to reroute away stuff from dead FPGA pins if there are any
<whitequark[cis]1>
requires you to stay on a fork but you save money and time
omnitechnomancer has joined #glasgow
<omnitechnomancer>
Gotta run an x86 emulator and display stuff through that somehow, maybe have a windows 95 application
<esden[m]>
omnitechnomancer (@_discord_157742445801504768:catircservices.org) you mean it looks extra oldskool? 😄 I am glad you are not suggesting C64 or a Centurion... 😛
<esden[m]>
I gave the date match a +-1 day wiggle room to help with timezone issues. And I check if the order was fulfilled...
<esden[m]>
Sidenote, rust is fun! let diff = od.date.signed_duration_since(query_date.unwrap()).abs().num_days();
<omnitechnomancer>
I mean we run windows 95 in the emulator on the browser and have an emulated network interface so that a win 95 application can make the query
<omnitechnomancer>
I do enjoy how many convenience things there can just be
<esden[m]>
Ahh ok, yeah, that is pretty much what I feel wasm web apps are (see Leptos). (I had a false start on this website using Leptos, it is a bit too much of a machine to build a small lookup page with, but for a bigger project I would consider using it.)
<esden[m]>
I agree, I like how generally well thought out the crates are.
galibert[m] has joined #glasgow
<galibert[m]>
I don't have mine yet, but I was wondering, can a glasgow sniff USB traffic, to find out where I got the implementation wrong?
<galibert[m]>
or is the differencial-but-not-quite nrzi just not sane enough?
<galibert[m]>
Oh, nitpicking, Glasgow Case with the queue ID 1377, we have to ship 241 Glasgows to Mouser before your order can be filfilled. -> we have to ship 241 Glasgows *cases*, right?
<nekrondev[m]>
we have to send 87 cases back to Mouser 😉
<Darius>
nice
<nekrondev[m]>
nah just kidding... the query form is just great and helps checking your current queue position
<Darius>
I got a notification about my case shipping
<nekrondev[m]>
ah me should RTFM
<nekrondev[m]>
```Note: There is a possibility that some of the numbers are negative. If for example the "At Mouser" value is negative it most likely means that some items were returned and Mouser ended up shipping them to another customer. This results in more Glasgows being shipped to customers than the amount we sent to them. If the number in "Remaining" is negative, this indicates that we sent more items to Mouser than the amount needed to
<nekrondev[m]>
fulfill the outstanding orders. ```
bvernoux has joined #glasgow
<marcus_c>
galibert[m]: USB 1 (Full speed and Low speed) can be sniffed as two single ended 3.3V signals. No voodoo needed. :-)
<galibert[m]>
you don't need some kind of rebuilt clock?
<marcus_c>
The clock is slow enough that you can rebuild it with LUTs and DFFs in the FPGA.
<galibert[m]>
I see, just a hand-made pll
<galibert[m]>
Annoyingly the one I have problems with is high speed 2.0
<marcus_c>
Yes, high speed is going to be more of a challenge.
<galibert[m]>
Dunno if I can make the (linux) host dial it down to full, since the protocol is mostly the same
FFY00 has quit [Read error: Connection reset by peer]
FFY00 has joined #glasgow
<whitequark[cis]1>
<galibert[m]> "Dunno if I can make the (linux..." <- full is the default, high has to be negotiated
<whitequark[cis]1>
getting down to full can be done with a low pass filter
<whitequark[cis]1>
cut off everything above 50 MHz, done
<whitequark[cis]1>
or use a 1.1 only hub which is functionally similar
<whitequark[cis]1>
(hard to find nowadays or even 10y ago)
miek__[m] has joined #glasgow
<miek__[m]>
i'm not sure that a low pass would work, the chirp sequence used for negotiation doesn't actually run at high-speed - it's only about 10 kHz
<whitequark[cis]1>
oh, I didn't realize that!
<whitequark[cis]1>
I think it would *probably* fall back to HS after not being able to up-speed but that is uninformed speculation; I haven't looked at the FSMs
<whitequark[cis]1>
(either way this is quite easy to test)
redstarcomrade has quit [Read error: Connection reset by peer]
<whitequark[cis]1>
a proper USB capture device is an ULPI transceiver glued to the bus, like in OpenVizsla, and this can be done for Glasgow too but ~requires HyperRAM
<galibert[m]>
well, with any luck I'll have the code fixed even before the glasgow gets here :-)
<nekrondev[m]>
yep HyperRAM will be a useful add-on when it comes to FIFO buffer limits (like LA , USB and other high speed protocols). What are the plans on this... going after this once the campaign had been fulfilled?
jan2642[m] has joined #glasgow
<jan2642[m]>
infinitum.
<jan2642[m]>
It will most likely not fallback to FS. It's indeed true that USB enumeration doesn't happen at HS and the HS capability is just flags in the device descriptor. We have a device at work where USB is routed through a multiplexer which only supports up to 20 MHz. If we don't limit the device to FS it will happily enumerate fine to HS but fail dramatically once you try to send data. Then it will just reset and re-enumerate again to HS, ad
<galibert[m]>
Huh?
<galibert[m]>
The HS/FS decision is done before the SET_ADDRESS message from what I understand
<whitequark[cis]1>
<jan2642[m]> "It will most likely not fallback..." <- ah, good to know. one more annoying usb behavior. thanks for confirming!
<vegard_e[m]>
kinda like how ethernet will happily negotiate gigabit even when only two pairs are available
hardkrash[m] has joined #glasgow
<hardkrash[m]>
I used to use the LTM2894 as a way to force full or low speed modes
balrog has quit [Quit: Bye]
balrog has joined #glasgow
cr1901 has quit [Read error: Connection reset by peer]
cr1901_ has joined #glasgow
cr1901 has joined #glasgow
cr1901_ has quit [Ping timeout: 246 seconds]
cr1901 has quit [Ping timeout: 246 seconds]
<omnitechnomancer>
full speed or high speed?
<omnitechnomancer>
It is a shame that the downshift type features are not more standard
cr1901 has joined #glasgow
farhan_009c[m] has joined #glasgow
<farhan_009c[m]>
Hello, don't know if this is the right time or the channel but I have a question I am new to cyber security can anyone tell me where should I start.
dx has joined #glasgow
<whitequark[cis]1>
I don't think this is the right channel really
<dx>
hi! i got my glasgow. i'm happy.
<dx>
now i am facing the problem i've been fearing of not having any idea how to attach probes to things on boards. generally.
<whitequark[cis]1>
oh, so there are a few different ways
<whitequark[cis]1>
the most universal one is to get thin lacquer coated wire ("magnet wire") and solder it to Whatever. you can do it with really tiny packages, like down to 0.25 mm pitch, with some skill and a lot of flux
<whitequark[cis]1>
the other end you solder to a normal 2.54 mm pin header that you can connect to glasgow, which in turn you epoxy to the board so that it doesn't move. make sure to put on enough epoxy to immobilize the actual pins or they can slide out
<whitequark[cis]1>
you can also get micro-hooks. they are expensive, pain in the ass to obtain even in the west (were you from South America?) and don't work that well
<whitequark[cis]1>
the aliexpress ones are cheap and basically completely worthless; at best they're low-single-digit-uses
<whitequark[cis]1>
the pomona ones for example are like 50 times more expensive and still don't stay in place very well, and are very frustrating to handle, but at least they mostly work as advertised, it's just that many modern targets are way too high density for this to be a reasonable strategy
<whitequark[cis]1>
there is also a strategy where you have a long, thin, rigid probe that is fixed to a gooseneck or similar (made out of some sort of wire or rod you can bend and position) where the probe is merely touching the target, and the contact is maintained by gravity and/or spring force. I think this is the highest end option that testing labs opt for for low-speed signals, and coincidentally you can make it out of a wire hanger and fishing
<whitequark[cis]1>
weights if you really want to do it cheaply
<whitequark[cis]1>
ah, and for the really high speed signals that I assume you won't be handling yet, there are solder-on flex-PCB probes which let you attach to transmission lines like those used in high speed protocols like USB 3 while disturbing them sufficiently little for the whole contraption to work. you can do this with magnet wire too if you're sufficiently motivated or tolerant of failure, it's just a more reliable version of the same concept
<whitequark[cis]1>
hope this helps
<dx>
intense. it does give me a bunch of leads. thank you
<dx>
(i was indeed from south america but i'm in germany nowadays)
<whitequark[cis]1>
that helps
SnoopJ has quit [Remote host closed the connection]
SnoopJ has joined #glasgow
SnoopJ has quit [Client Quit]
SnoopJ has joined #glasgow
Eli2_ has joined #glasgow
Eli2 has quit [Read error: Connection reset by peer]
gsuberland has quit [Remote host closed the connection]
gsuberland has joined #glasgow
<esden[cis]>
<galibert[m]> "Oh, nitpicking, Glasgow Case..." <- ok, fixed the typo and a bunch of the copypasta mistakes in there. (there is a bunch of spaghetti template stuff going on there...)
<esden[m]>
The fact that it even told you that there are outstanding cases to send to Mouser was wrong... but it is fixed now.
SnoopJ has quit [Quit: "Entropy always comes for its due, and that's what even lobsters must accept."]
SnoopJ has joined #glasgow
GregNGM has joined #glasgow
redstarcomrade has joined #glasgow
redstarcomrade has joined #glasgow
redstarcomrade has quit [Changing host]
Eli2_ has quit [Remote host closed the connection]
Eli2 has joined #glasgow
<josHua[m]>
the brand name version of this, in the west at least, I believe is Sensepeek
<whitequark[cis]1>
i waaaaaant
<josHua[m]>
for less permanent uses, hot melt adhesive can be pretty good for this too. (or blu tak, for extremely temporary uses.) hot melt adhesive has the benefit that you can hit it with your hot air rework station (turn the temperature down!!) which has way less of a stringy mess, and also heats the substrate well enough that you get a better bond, and also you can shape it with the air
<josHua[m]>
I am glad it is not just me that can't make microhooks work, though. I have a bunch of them that came with my Saleae and they are nto good
Attie[m] has joined #glasgow
<Attie[m]>
I'd quite like to try one of those "long nail" fixutres some time... but I imagine they'll be frustrating for anything that isn't "small"
<Attie[m]>
I generally find the micro hooks are okay... but I did manage to melt a few the other day
<Attie[m]>
I'd hooked up to a number of points on a jet engine, started the engine, and noticed a few probes fall off... on inspection 3x had fallen off - on one each of the BLDC phases
<Attie[m]>
turns out I'd got the harness upside down, and had effectively shorted the BLDC phases through my logic analyser (😱) ... the ground clip had also melted (even more so)
<Attie[m]>
the pen + hole setup looks interesting too - have you used / got feedback Catherine ?
<whitequark[cis]1>
ouch
notgull has joined #glasgow
notgull has quit [Ping timeout: 264 seconds]
<josHua[m]>
it sounds like you were lucky to have some micro hook shaped fuses
<esden[m]>
hah! yes, you are one of the skipped orders, it would have been your turn a while ago...
<esden[m]>
I should probably add some logic to indicate that...
<esden[m]>
I need to make that variable into a signed value I guess...
<ebb>
Awesome stuff esden!
<K900>
Edge case: war
<K900>
(ugh)
<esden[m]>
(I am learning that I don't like Askama template system very much... it is annoying to debug, and it has a bunch of shortcomings when you need to deal with calculations... I might look into something else... :/ )
<K900>
Can recommend minijinja
<K900>
It's basically just Jinja
<esden[cis]>
ebb: Thanks! :D
<galibert[m]>
Glasgow Case 1377 Need to ship 18446744073709551245 to Mouser before ready.
<galibert[m]>
not only that one variable :-)
<ebb>
This actually spurred me to find out where mine is
<esden[m]>
that does not make sense because it should say it is ready to ship...
<esden[m]>
gaaaahhh again copypasta BS... >_<
<galibert[m]>
huhuhu
<galibert[m]>
How fast are you producing them lately, roughly?
<esden[cis]>
K900: Ok i will check it out. Askama also says it is jinja inspired... curious if minijinja is better...
<ebb>
Minor UX thing, using placeholder= for the yyyy/mm/dd in date inputs will avoid having to empty the fields by oneself
<K900>
esden[cis]: minijinja is written by Armin, the same guy that wrote the original Jinja/Jinja2
<esden[cis]>
ebb: ahh! great! Thanks! for my defense I am not a web dev :P
<esden[cis]>
I appreciate the input!
<ebb>
lol
<ebb>
10/10 pun
<esden[cis]>
:D
<esden[m]>
ok, should be better now
<esden[m]>
slightly better... I still need to deal with "skipped" oders... I guess I should have a message there, "Your order was skipped for some reason. If you have questions contact CrowdSupply" or something like this?
<galibert[m]>
Yep, says "Ready to ship" now
<esden[m]>
Ok great! 🙂
<galibert[m]>
interesting, 681 orders before mine but 739 to send to mouser before I'm reached.... oh, that means multiple glasgows in some orders
<esden[m]>
exactly
<esden[m]>
I should probably also add a disclaimer. "This information is calculated, it might diverge from Mouser operations and might not reflect physical reality."
<esden[m]>
damn, where is all that traffic to the page coming from? Ahh yes... Catherine posted the link on masto... 😛
redstarcomrade has quit [Read error: Connection reset by peer]
<josHua[m]>
feel free to grab whatever bits and pieces of that you like. I am also not a webdev but I have heard of bootstrap before. so now it looks like every other piece of content on the internet
<whitequark[cis]1>
may I suggest cutting the table in two
<josHua[m]>
King Solomon accepts pull requests
<josHua[m]>
(actually, just wget it and modify it, I messed with it statically)
<esden[m]>
Yeah I will do that josHua (@_discord_256468461818085377:catircservices.org)
<esden[m]>
whitequark (@_discord_182174208896401419:catircservices.org) you mean the "shipped" and "orders" should be two tables. That is a good idea.
<endrift[m]>
(It took me several minutes to realize this was the aluminum cases for my Glasgows
<endrift[m]>
* my Glasgows)
<whitequark[cis]1>
niiiiiiiiice
<whitequark[cis]1>
endrift: oh hi i didn't realize you bought one :D
<endrift[m]>
I bought two!
<endrift[m]>
I haven't used them yet
<whitequark[cis]1>
whoa
<endrift[m]>
was hoping to chain them together to use as a ~28 channel protocol analyzer
<endrift[m]>
that seems like it may be a bit optimistic though
<josHua[m]>
ooh. UPS has told me that mine, also, is arriving tomorrow. 24 hours before I get to start in on glasgow.applet.interface.wafflehouse!
<josHua[m]>
(glasgow.applet.interface.fauxtdi?)
<whitequark[cis]1>
fauxtdi?
<josHua[m]>
my harebrained idea to expose a glasgow as what looks to OS software like a pair of FT4232s with a very versatile level shifter front end over USBIP
<whitequark[cis]1>
I approve of wafflehouse
<whitequark[cis]1>
depending on how cursed the code is, I might even accept it upstream, on the basis of it being extremely funny
<josHua[m]>
use all your pyftdi or other tools (or flashrom or whatever), and turn four tigards on your desk into a single glasgow
<josHua[m]>
I mean, it brings 'glasgow' full circle.
<josHua[m]>
I don't generally write code that is cursed for the sake of being cursed. I generally try to write good code. sometimes the idea enforces some level of cursedness, but in this case, I imagine that roughly the glasgow gateware would implement a command FIFO of 'things that look like simplified versions of the FTDI commands, just not wrapped in USB headers', and the applet side would expose a aio socket interface. these bits need not be
<josHua[m]>
cursed; there would be some out-of-tree (or contrib) apparatus that takes the socket swiss army knife interface and exposes a usbip interface (this is where the tradeoffs are made from 'non-bullshit code that avoids breaking abstraction layers of FTDI protocol, USB protocol, and USBIP protocol' to 'actually writing programs in a timely fashion')
<whitequark[cis]1>
oh no what i mean is that if it's too cursed i don't want to maintain it
<whitequark[cis]1>
so you making an effort to not make it cursed is good in this case
<whitequark[cis]1>
and i think wafflehouse is enough of a deterrent / indicator to people that they should absolutely not rely on this thing too much
<whitequark[cis]1>
> would implement a command FIFO of 'things that look like simplified versions of the FTDI commands
<whitequark[cis]1>
there is a gateware doing exactly this in the git log
<whitequark[cis]1>
you don't even need to implement it
<whitequark[cis]1>
> this is where the tradeoffs are made from 'non-bullshit code that avoids breaking abstraction layers of FTDI protocol, USB protocol, and USBIP protocol' to 'actually writing programs in a timely fashion'
<whitequark[cis]1>
i mean i am willing to merge that if it's bearable
<whitequark[cis]1>
there is nothing inherently wrong in adding arch/protocol definitions for USB and USBIP to Glasgow main
<whitequark[cis]1>
and a combination of all that may be written in a maintainable way
<whitequark[cis]1>
the issue i have is that usbip is inherently something that is likely to crash your kernel, so you should just not rely on it if you can avoid it
<whitequark[cis]1>
and i think glasgow run wafflehouse is a sufficient indication of "this may crash your kernel" to be fine with it
<whitequark[cis]1>
I guess one other consideration I have is that this almost completely removes the incentive to rebuild other tools on top of the Glasgow framework if they already work with an FX2
<whitequark[cis]1>
so I might still be undecided about it and request an out-of-tree applet in the end
<josHua[m]>
'the OS/2 problem'
<whitequark[cis]1>
yes
<josHua[m]>
in this case, though, your goal is to create a tool that works well for many purposes, rather than to supplant Microsoft
<whitequark[cis]1>
i didn't want ftdi's trademark lawyers on my ass
<josHua[m]>
whee
<whitequark[cis]1>
I think I'd like to have it out-of-tree for now for the 2nd reason, and re-evaluate the decision after the new applet UI lands which will simplify writing new ones and composing existing ones
<whitequark[cis]1>
* 2nd reason mainly, and
<whitequark[cis]1>
if people think we should have it as default then so be it
<whitequark[cis]1>
but I don't want this to be an out-of-the-box feature, if this makes sense
<josHua[m]>
I am not sure I can count the enumeration of reasons to identify the second reason, but also, a primary goal for this is 'for joshua to learn the Glasgow environment', and also 'to be vaguely useful to joshua', and if it makes it upstream, lovely
<whitequark[cis]1>
sorry, I mean I want to avoid the OS/2 problem
<josHua[m]>
o
<josHua[m]>
the MPSSE protocol is limited, slow, and annoying enough that I do not think that you will have people choosing to write MPSSE apps when Glasgow gateware is appropriate
<whitequark[cis]1>
yeah, this is by no means a final "you shall not pass" decision
<whitequark[cis]1>
it's just what I feel is probably better for the project at this time point
<josHua[m]>
that's fine
<josHua[m]>
we will see if I get to doing it or not. I have a fair bit of polyprojectitis right now with the x1 expansion board, my languishing probe program, and, like, paying work.
<whitequark[cis]1>
x1 expansion board?
<josHua[m]>
I have fabbed out an expansion board for the Bambu Lab X1 Carbon 3D printer that plugs into the AP board's camera connector (USB over 1.5mm pins), and has a USB hub + USB ethernet part and a FT2232. adds ethernet (the built in wifi chipset sucks), two USB ports, and a world of GPIO (some of which brought out to a Adafruit-compatible I2C header)
<whitequark[cis]1>
oh this X1
<josHua[m]>
yes
<josHua[m]>
in a week or so, it will be time to bring it up, finish making plastics for it, and frantically write some software to glue the proprietary bits to my python background daemon to make it do something interesting in response to onboard actions. that will be cursed code.
<esden[m]>
josHua (@_discord_256468461818085377:catircservices.org) thank you very much. This definitely looks better than it used to. 🙂 https://glasgow.1bitsquared.com
<whitequark[cis]1>
s-should the disclaimer not be centerd too
<whitequark[cis]1>
it looks weird <.<
<josHua[m]>
I just started trying to mess with it in the firefox element editor
<josHua[m]>
and then was like 'ah heck'
<josHua[m]>
I tried to put it in a <div class="row"> or something and then realized I had absolutely no idea how a bootstrap navbar worked, and then was like '?' and gave up
<esden[m]>
Yeah I tried to get the navbar to center too ... and failed.
<esden[m]>
But there is no good way to make something else stick to the bottom of the page. Have that thing just tacked at the end of the content looks strange too...