#linuxcnc Logs
Mar 17 2024
#linuxcnc Calendar
02:20 AM Deejay: moin
03:17 AM Loetmichel_ is now known as Loetmichel
04:58 AM Tom_L: morning
05:01 AM JT-Cave: morning
06:15 AM travis_farmer: just setup a shopping cart on my web (un published so far) to sell my CNC made crafts when i make them, using OpenCart and PayPal. my question for JT, how much fraud do you get on your store? i am worried about "opening" my store, because i worry about fraud and scams.
06:25 AM travis_farmer: JT-Cave?
06:29 AM JT-Cave: I've had no issues with paypal
06:29 AM travis_farmer: ok, thank you
06:32 AM travis_farmer: now i just have to make stuff and figure out shipping :-)
06:36 AM JT-Cave: I use woocommerce and use weight based shipping
06:37 AM travis_farmer: ok, thank you.
07:03 AM JT-Cave: Tom_L, clear mdi history works for me
07:20 AM JT-Cave: Tom_L, any mouse click will now copy a mdi item to the mdi entry
07:29 AM JT-Cave: Tom_L, I "think" I fixed everything you broke
08:10 AM Tom_L: i'll test after a shower
08:12 AM Tom_L: oh, 'step' doesn't move the yellow line in gcode view here
08:13 AM Tom_L: haven't tested that much yet
08:33 AM JT-Cave: it might take 3 presses of step to move the line
08:35 AM JT-Cave: I'm not sure what emc does for the first two steps if your on line 0
08:40 AM Unterhaus_: I found some aluminum that will work to hold up my parts organizers. It's going to be even more labor intensive than Adam Savages method, but I guess I have the time
08:54 AM solarwind: YESSSSSSSS
08:54 AM solarwind: Got PREEMPT_RT to decent latency on raspberry pi 5 with custom linux-6.8 kernel config
08:55 AM solarwind: 250µs interval worst case jitter is < 30µs over the past few days
08:56 AM solarwind: had to dig through a lot of code since linx-6.1
08:56 AM solarwind: so many changes... the kernel developers and maintainers should be getting a salary for this work
08:58 AM solarwind: it's been great not having to deal with 40 years of x86 legacy stuff and SMI and other NMI
08:58 AM lcnc-relay: <skunkworks8841> nice!
08:58 AM solarwind: The interrupt list actually fits in one console window so I can keep an eye on it
08:58 AM lcnc-relay: <skunkworks8841> Whatever image and kernel I have on my was around <100us iirc (which was fine for mesa..)
08:59 AM lcnc-relay: <skunkworks8841> have it on for days at a time without any missed deadlines
08:59 AM solarwind: yeah, I just didn't want to use that official image since it's doing weird things and doesn't play nice with the imager
08:59 AM solarwind: I wanted to use stock raspbian and be able to just install my own kernel
09:00 AM solarwind: at first, the latency was much worse, now it's much better
09:00 AM solarwind: than the linuxcnc image that is
09:00 AM lcnc-relay: <skunkworks8841> did you isolate cores and forceturbo?
09:01 AM solarwind: that's pretty much a hard requirement
09:01 AM lcnc-relay: <skunkworks8841> I am also overclocking it a bit.
09:01 AM lcnc-relay: <skunkworks8841> right
09:02 AM solarwind: isol_cpus=managed_irq,domain..., rcu_nocbs=..., rcu_nocb_poll, kthread_cpus=..., irqaffinity=...,
09:02 AM solarwind: nohz_full=...
09:02 AM solarwind: that's just off the top of my head
09:03 AM solarwind: kernel.sched_rt_runtime_us=-1 kernel.timer_migration=0
09:03 AM lcnc-relay: <skunkworks8841> lol - I just have isol_cpus and idle=poll
09:04 AM solarwind: I can have it compile a kernel in the background and it'll have zero impact on the realtime threads
09:04 AM lcnc-relay: <skunkworks8841> I should try all of those.. you should post your actual kernel line somewhere
09:05 AM solarwind: yeah and kernel config for 6.8 and compile instructions and all that for raspberry pi 5
09:05 AM Tom_L: JT-Cave, on touchoff, do you use the spin box or the current value of the coordinates?
09:05 AM lcnc-relay: <skunkworks8841> oh - you are running a 5?
09:05 AM Tom_L: i got a warning the touchoff spin box wasn't installed
09:05 AM Tom_L: it used to use the current coordinates iirc
09:06 AM solarwind: skunkworks8841, yes so I don't have to touch x86
09:06 AM solarwind: I prefer it for embedded
09:06 AM JT-Cave: spin box
09:06 AM Tom_L: recent seems ok now
09:06 AM lcnc-relay: <skunkworks8841> what are your thoughts on the pi5 is it that much faster?
09:07 AM solarwind: there's too much to know about x86 since it's usually designed for throughput, not embedded/low latency
09:07 AM JT-Cave: touch off has always used a spin box
09:07 AM Tom_L: can it be set as an either/or?
09:07 AM solarwind: pi 5 is indeed using a newer CPU
09:07 AM solarwind: and yes it is much faster
09:08 AM solarwind: but that doesn't matter for embedded
09:08 AM Tom_L: you jog your wobbler to position and push touchoff. as is that's not possible
09:08 AM solarwind: it has a PCIe FPC connector which is very convenient
09:08 AM Tom_L: unless you match the value with the spin box
09:08 AM solarwind: for Pi 4, you needed the compute module to access its PCIe lane. On the Pi 5, you just need a breakout board
09:10 AM solarwind: I just hit the reset stats button and restarted latency test with default options, it's sitting pretty with < 10µs on both threads while it's compiling a whole kernel in the background
09:11 AM solarwind: yes it's completely unnecessary for linuxcnc with mesa cards, but I do a lot of other embedded/realtime things with it
09:11 AM Tom_L: jog increments seem off here
09:11 AM Tom_L: .1 doesn't move .1
09:11 AM solarwind: Tom_L you must be using freedumb units. Set to metric where .1 is .1
09:13 AM solarwind: there's very occasional lag when moving the mouse cursor but that's _exactly_ what I wanted to see - it means it's properly giving full priority to realtime tasks and all other bullshit including USB and display is getting deprioritized
09:16 AM solarwind: Also rpi 5 is in stock everywhere it seems
09:16 AM solarwind: I appreciate how well it's documented and supported soooo much and how can you not love the pricing
09:16 AM solarwind: Will support raspberry pi all day over sum-ting-wong clones / odroid bullshit
09:17 AM solarwind: it pisses me off to see fuckin chinese characters in the extremely limited documentation/code they have
09:19 AM Tom_L: where do you set the precision on the spin box? all i'm getting are whole numbers 1,2,3...
09:20 AM solarwind: Tom_L https://doc.qt.io/qt-6/qdoublespinbox.html
09:20 AM solarwind: Tom_L decimals, stepType, singleStep
09:32 AM Unterhaus_ is now known as Unterhausen
09:36 AM Unterhausen: they say I can clear data from just one site, but they got rid of that option
09:37 AM Tom_L: spindles must still be broke
09:38 AM Unterhausen: nice to slip that into production with zero testing
09:42 AM JT-Cave: what error are you getting running the mdi command
09:42 AM Tom_L: <Tom_L> 'invalid params in linear command'
09:42 AM JT-Cave: that's you then
09:42 AM Tom_L: <Tom_L> G0 X0 Y0 Z0
09:43 AM Tom_L: <Tom_L> or G0 X.25 same error
09:44 AM JT-Cave: they all work for me... are you outside the bounds or something
09:44 AM Tom_L: i'll try again
09:45 AM JT-Cave: iirc I got that error trying to move outside the limits in mdi
09:45 AM Tom_L: i can jog past that point
09:45 AM Tom_L: G0 X.1 errors
09:46 AM Tom_L: aparently you have to not only home but touch off
09:46 AM Tom_L: it should work in world coordinates as well as fixture coordinates
09:48 AM Tom_L: do your active codes display work yet?
09:49 AM Tom_L: mdi works here but you have to touch off first along with homing
09:51 AM JT-Cave: all axes must be homed to run mdi touch off does not matter except for limits and that's emc not flex
09:52 AM Tom_L: lemme try another test
09:52 AM Tom_L: run program is enabled with no file loaded
09:52 AM Tom_L: wasn't the test i was going for...
09:53 AM JT-Cave: and yes the g and m code labels work
09:53 AM Tom_L: i was looking for them in _lb
09:53 AM Tom_L: haven't found it just yet
09:54 AM JT-Cave: run is disabled if no file is loaded
09:54 AM Tom_L: i pushed 'run program' just to see and 'pause' was enabled
09:58 AM Tom_L: is 'tool_table_zoffset_0_lb' the current tool?
10:03 AM JT-Cave: pause is enabled when a program is running and not paused
10:03 AM Tom_L: but i did that with no file loaded
10:04 AM JT-Cave: run program is not enabled if no file is loaded... are you sure no file is loaded in linuxcnc?
10:04 AM JT-Cave: The Tool Table status contains status items for all the tools in the tool table. Replace the _n_ with the position in the tool table.
10:05 AM Tom_L: i read that about the tools. i wanted the offset for the currently loaded tool
10:05 AM JT-Cave: so to answer that question I don't know what 0 might be
10:05 AM Tom_L: used to be tool_offset_lb
10:05 AM Tom_L: i think that one is gone now
10:06 AM JT-Cave: it's still there
10:07 AM JT-Cave: it returns all the offsets for the current tool
10:07 AM JT-Cave: it's in status labels and it needs to be fixed up by axis letter I think
10:07 AM Tom_L: start flex, enable estop, power on, home x, home y, home z, run program, pause is enabled
10:07 AM Tom_L: no file loaded
10:08 AM Tom_L: does throw a 'no file loaded' error in status
10:08 AM Tom_L: but it still enables it
10:08 AM JT-Cave: hmm I see that now
10:09 AM Tom_L: i've noticed converting jet, i have to re select the labels in order for them to become active for some odd reason
10:11 AM JT-Cave: re select?
10:11 AM Tom_L: reenter the widget objectName
10:11 AM Tom_L: i've done that on several now
10:13 AM Tom_L: tool_offset_lb returns a tuple of offsets
10:14 AM Tom_L: must be called something else
10:15 AM Tom_L: tool_in_spindle_lb works but i can't figure out where the offsets for it are
10:18 AM Tom_L: tool_offset_lb should be it but it returns a tuple
10:19 AM JT-Cave: yeah I've not programmed tool offset stat yet to have the axis
10:19 AM Tom_L: k
10:20 AM JT-Cave: found that run enable bug and squashed it
10:20 AM Tom_L: :)
10:20 AM * JT-Cave heads to town
10:28 AM Tom_L: you didn't push it
11:05 AM Unterhausen: oshcut stocks O1
11:23 AM JT-Cave: telepathy not working today :)
11:45 AM Tom_L: if you load the file after estop,power, home x y z run program is not enabled otherwise it is
11:46 AM Tom_L: with 'file open' or 'recent'
12:02 PM lcnc-relay: <skunkworks8841> lol - inheritance machining... oof. I think it would be funny to make a video - where I walk into the shop - turn on the lights and walk over to the cnc and take that shape out of the cnc...
12:21 PM Unterhausen: in the early days of additive machining, I wanted to make a video where I pulled a cellphone out of the moore
12:21 PM Unterhausen: and said our lab had advanced capabilities in manufacturing
12:22 PM Unterhausen: I'm pretty sure I saw someone predicting such things for additive manufacturing
12:29 PM lcnc-relay: <big_kevin420> add a flipper zero in there, and you have youtube gold
12:46 PM Roguish: well, well, warm and sunny day.....
12:48 PM Roguish: geeeze..... just files taxes..... what a fracking nightmare. why can't we have a simple flat tax rate. no loopholes, no exemptions, no exceptions.
12:49 PM Unterhausen: it would be funny to rig the x axis to throw the cellphone out of the enclosure
12:49 PM Unterhausen: I was going to have a hidden grad student stand at the back
12:50 PM Unterhausen: youtube finally loading, I really didn't want to wipe everything out for all websites
12:52 PM Unterhausen: never mind, still isn't working
12:55 PM Unterhausen: porn spammers took over the facebook page of a local farm. Facebook just told me they disregarded my report because it didn't violate community standards
12:55 PM Unterhausen: there is no human there that can actually make the simplest of decisions, it's awful
01:17 PM Unterhausen: I just decided to report FB to the state AG, zuck needs some governmental assistance
03:14 PM * JT-Shop patiently waits for Tom_L to break something in flex
04:04 PM Tom_L: JT-Shop, you get my note about the file open order?
04:04 PM JT-Shop: https://youtu.be/0rWvM_OGaHo?list=PLH4J0b3KfGaNED27Y1kCvS_mSQtyXsiwC&t=433
04:04 PM Tom_L: Tom_L> if you load the file after estop,power, home x y z run program is not enabled otherwise it is
04:04 PM JT-Shop: I've never seen scotch brite roller before
04:04 PM Tom_L: <Tom_L> with 'file open' or 'recent'
04:04 PM JT-Shop: and I've fixed it
04:05 PM Tom_L: this was after the first fix
04:05 PM Tom_L: the and clause you added
04:05 PM Tom_L: the order sequence matters
04:05 PM JT-Shop: https://github.com/jethornton/flexgui/commit/6c2f62185ebd54ad5164086d5505aade245adf15
04:05 PM Tom_L: i must have been napping
04:05 PM * Tom_L clocks back in
04:06 PM JT-Shop: :)
04:06 PM JT-Shop: also while you were napping I added current tool offsets
04:07 PM Tom_L: thanks
04:07 PM JT-Shop: thank you for testing
04:07 PM Tom_L: i'll thank you by breaking it
04:07 PM JT-Shop: :)
04:08 PM Tom_L: for some reason i had to clone the repo instead of fetch. it failed to bring in the changes a couple times
04:09 PM JT-Shop: I use get pull --rebase
04:09 PM JT-Shop: but you could just use git pull
04:09 PM Tom_L: git fectch origin; git reset --hard origin/master is what i usually do
04:10 PM JT-Shop: if you have not changed anything local a git pull works
04:10 PM Tom_L: ok
04:10 PM Tom_L: if i can remember that
04:11 PM JT-Shop: I added it to git gui
04:11 PM Tom_L: i just use the terminal
04:13 PM Tom_L: run program works with the file open sequence now
04:13 PM JT-Shop: yippie
04:13 PM Tom_L: i'm not done testing it but that sequence works
04:13 PM Tom_L: the tool_offset_lb works now?
04:14 PM JT-Shop: https://gnipsel.com/linuxcnc/flexgui/labels.html#special-labels
04:15 PM Tom_L: is current tool 0 then?
04:15 PM JT-Shop: for?
04:16 PM Tom_L: tool_offset_lb_0
04:16 PM JT-Shop: that's the x axis
04:16 PM Tom_L: oh i see now
04:16 PM JT-Shop: https://gnipsel.com/linuxcnc/flexgui/controls.html
04:16 PM JT-Shop: see the axis index on that page
04:16 PM Tom_L: maybe post an order of what each represent
04:17 PM JT-Shop: there is a note...
04:17 PM Tom_L: i don't see it
04:18 PM Tom_L: see the control page for...
04:18 PM Tom_L: ok
04:18 PM JT-Shop: https://gnipsel.com/linuxcnc/flexgui/labels.html#special-labels
04:18 PM JT-Shop: later I'll put a link to that
04:18 PM Tom_L: is there one for tool_diameter_lb?
04:19 PM JT-Shop: hmm I'd have to look
04:19 PM Tom_L: there used to be
04:20 PM JT-Shop: that's in the tool table
04:21 PM JT-Shop: I don't see anything that has current tool diameter
04:22 PM Tom_L: tool_offset_lb_2 doesn't load the Z offset value
04:22 PM Tom_L: of the current tool
04:22 PM JT-Shop: it did for me
04:23 PM Tom_L: lemme open the tool table and see what it says there :)
04:23 PM Tom_L: why is MDI disabled?
04:24 PM Tom_L: i estop, power, home, touchofff
04:24 PM JT-Shop: do you have all the mdi items?
04:24 PM Tom_L: yes, i've used it
04:24 PM Tom_L: and it just worked a sec ago
04:25 PM Tom_L: trying to find the sequence that trips it
04:25 PM JT-Shop: how did you break it?
04:26 PM Tom_L: now it's enabled
04:26 PM Tom_L: i dunno what sequence breaks it but one does
04:26 PM JT-Shop: more testing is needed :)
04:26 PM Tom_L: of course the tool i tested had a 0 offset :)
04:27 PM Tom_L: nope it's not working
04:27 PM JT-Shop: it?
04:27 PM Tom_L: T11 has an offset of .5195 and shows 0
04:27 PM Tom_L: Z offset
04:28 PM JT-Shop: ok
04:28 PM Tom_L: yeah tried T1 and no update
04:29 PM * Tom_L gets out the spell checker too
04:29 PM JT-Shop: what about 0 and 1 x and y
04:29 PM Tom_L: lemme add those real quick
04:30 PM Tom_L: none update
04:31 PM JT-Shop: what if you restart flex?
04:31 PM Tom_L: i rerun the sim each time
04:31 PM JT-Shop: ok I'll have to test and see
04:31 PM Tom_L: it changed i know because it added 2 more labels
04:32 PM Tom_L: for 0 1 & 2, x y & z
04:33 PM Tom_L: lemme ask you this.
04:33 PM Tom_L: if you restrict the colums shown in the tool table does that affect this any?
04:33 PM JT-Shop: no
04:33 PM Tom_L: the file should be the same, just the display difference
04:33 PM Tom_L: that's what i thought
04:33 PM JT-Shop: that's just for the tool table editor
04:33 PM Tom_L: right
04:36 PM Unterhausen: good job google, they might have just gotten me to switch to firefox because youtube is broken on chrome for me
04:40 PM Tom_L: ok now to figure out the broken mdi sequence
04:42 PM Tom_L: estop, power, home xyz, touchoff xyz, no mdi
04:43 PM Tom_L: estop, power, file open, home xyz, mdi works
04:43 PM Tom_L: no touchoff xyz
04:43 PM Tom_L: needed there
04:43 PM Tom_L: looks like file open is the culprit for mdi
04:44 PM Tom_L: but
04:45 PM Unterhausen: keyboard/chair interface problem
04:45 PM Tom_L: if you estop, power, home xyz, then file open, no mdi either
04:45 PM Tom_L: it depends on the sequence of events
04:46 PM Tom_L: if you open a file before homing xyz, mdi seems to work
04:46 PM Tom_L: after, it doesn't
04:47 PM JT-Cave: estop, power, home xyz is all that "should be" needed
04:47 PM Tom_L: right
04:47 PM Tom_L: the file open is messin with it
04:47 PM Tom_L: but you can force it to work if you open the file before homing xyz
04:48 PM JT-Cave: ah home all works but home a joint does not
04:48 PM Tom_L: ie: estop, power, open file, home xyz: mdi works
04:48 PM Tom_L: i didn't add home all back yet since it was broke earlier
04:49 PM Tom_L: i can add that back in now i suppose
04:52 PM XXCoder: lol jank tools https://youtu.be/pnq1ZuIjGgQ
04:55 PM JT-Shop: I see that reload tool table is broken in emc
04:56 PM Tom_L: :/ home_all_pb unhome_all_pb broke it
04:57 PM -!- #linuxcnc mode set to +v by ChanServ
04:57 PM Tom_L: and that would be an operator error
04:57 PM Unterhausen: I bought a hf grinder, and it eventually caught fire. Turns out the power cord was 32gauge or some such
04:58 PM Unterhausen: I keep meaning to replace the power cord, grinder is still fine and I have some concrete projects I want to kill it with
04:59 PM Tom_L: yes, home all enables mdi without a file
05:00 PM JT-Shop: file has nothing to do with mdi
05:00 PM Tom_L: home x y z separate does not
05:00 PM JT-Shop: right that is broken
05:00 PM Tom_L: maybe not but it did trigger something to enable mdi
05:01 PM Tom_L: estop, power, file open, home x y z allowed mdi to work
05:01 PM XXCoder: unter insane on cheap wiring
05:01 PM JT-Shop: home all is a different function than home some axis
05:01 PM Tom_L: ok
05:01 PM Tom_L: i'm just saying ^^ that sequence enables mdi
05:02 PM Tom_L: oh
05:03 PM Tom_L: on status, the 'Error' that shows at the bottom of the main window should clear for the next error
05:03 PM Tom_L: as should the message
05:03 PM Tom_L: or make a new line for the new message
05:04 PM Tom_L: maybe click on the 'Error' to clear it similar to axis
05:05 PM JT-Shop: I've barely got error to work so lots of room for improvement
05:05 PM Tom_L: i think it works good other than not being cleared
05:05 PM JT-Shop: the status bar message?
05:06 PM Tom_L: on axis it will scroll the errors
05:06 PM Tom_L: yes
05:06 PM Tom_L: both actually
05:06 PM Tom_L: there and in the status window
05:06 PM Tom_L: if there are no errors shown, clear the status bar
05:07 PM Tom_L: maybe click on each error to clear it
05:07 PM Tom_L: i dunno
05:07 PM JT-Shop: https://github.com/jethornton/flexgui/blob/master/flexgui/src/libflexgui/status.py#L361
05:07 PM Tom_L: so you use it for error and info
05:08 PM JT-Shop: yes and you can only check errors once per error
05:09 PM Tom_L: i _can_ manually clear the error in the status window with the keyboard and then the status window is blank
05:09 PM JT-Shop: keyboard with axis?
05:09 PM Tom_L: no with flex
05:10 PM Tom_L: if an error shows up in the status window, i can backspace and remove it
05:10 PM Tom_L: so if i get the same error i will know it's a new one
05:11 PM Tom_L: get an out of bounds error on a gcode file and try running it twice and you'll see what i mean
05:14 PM Tom_L: does parent.errors_pte have a 'clickable' option?
05:14 PM Tom_L: that might work
05:14 PM JT-Shop: https://doc.qt.io/qt-6/qplaintextedit.html
05:14 PM JT-Shop: look at signals
05:15 PM Tom_L: how do i find that in designer?
05:16 PM JT-Shop: signals are not in designer
05:16 PM Tom_L: k
05:17 PM Tom_L: whatever it is, it needs to be user interacive
05:17 PM JT-Shop: https://github.com/jethornton/flexgui/blob/master/flexgui/src/libflexgui/startup.py#L527
05:18 PM Tom_L: generally i would think if you got an error or more than one, you could clear the one or all during that session
05:18 PM Tom_L: so clearing it like mdi seems ok to me
05:19 PM Tom_L: instead of one at a time
05:19 PM JT-Shop: clear errors button?
05:19 PM Tom_L: that or just clear... whatever
05:19 PM Tom_L: that would work
05:19 PM Tom_L: then update the status bar
05:20 PM Tom_L: or clear status - clear messages
05:20 PM Tom_L: something.
05:20 PM Tom_L: if you use it for errors and messages
05:21 PM JT-Shop: they all come from the error channel
05:21 PM Tom_L: ok
05:21 PM JT-Shop: dunno what an info message would be
05:21 PM Tom_L: clear errors sounds good
05:22 PM Tom_L: 'undo yer fckup'
05:22 PM JT-Shop: lol
05:23 PM JT-Shop: I should go fill the deer feeder up
05:24 PM lcnc-relay: <skunkworks8841> Well - I found a pile of chains up in the woods inclucing a cable.. Still missing a block and tackle...
05:25 PM lcnc-relay: <skunkworks8841> I bet it is sinched around a tree - somewhere..
05:25 PM JT-Shop: treasure hunt?
05:26 PM lcnc-relay: <skunkworks8841> yes - we are missing a bunch of chains.. but - I found atleast 2...
05:27 PM Tom_L: you may have to borrow a metal detector
05:27 PM Tom_L: if you wait too long
05:27 PM lcnc-relay: <skunkworks8841> I know - that is why I was walking through the woods - before it gets green..
05:27 PM lcnc-relay: <skunkworks8841> I still have a bunch of areas to look..
05:28 PM lcnc-relay: <skunkworks8841> it is a big block in tackle.. probably a good foot or more across
05:29 PM lcnc-relay: <skunkworks8841> it might also be somewhere around the buildings...
05:29 PM lcnc-relay: <skunkworks8841> so - who knows..
05:29 PM lcnc-relay: <skunkworks8841> abom got his vise assembled.
05:34 PM lcnc-relay: <skunkworks8841> https://jauriarts.org/_matrix/media/v1/download/jauriarts.org/BzBLqNltPXIKpaompJFiAABx
05:34 PM lcnc-relay: <skunkworks8841> looks so dinky from up here
05:34 PM Tom_L: is that a well cover?
05:34 PM JT-Shop: are you flying?
05:34 PM Tom_L: or shed
05:35 PM lcnc-relay: <skunkworks8841> yes
05:35 PM lcnc-relay: <skunkworks8841> I don't know why dad didn't do a pitless..
05:35 PM lcnc-relay: <skunkworks8841> attic window
05:36 PM JT-Shop: I thought you were talking about the crawler
05:36 PM lcnc-relay: <skunkworks8841> I was - it looks so small
05:37 PM lcnc-relay: <skunkworks8841> https://jauriarts.org/_matrix/media/v1/download/jauriarts.org/WGGtOnFncwGPVsJdYwcPxWQF
05:37 PM lcnc-relay: <skunkworks8841> wow - that is a dirty window
05:39 PM Unterhausen: armchair heavy industries has quite the discord