<openfpga-github>
[openfpga] rqou commented on issue #121: Alright, updated the cmake file. Apparently the simple solution really does work. https://git.io/vFK1Z
soylentyellow has quit [Ping timeout: 264 seconds]
<azonenberg>
Welp
<azonenberg>
replaced the 115 ohm resistors with 100 in my ethernet PHY
<azonenberg>
and diff amplitude is now -975 mV
<azonenberg>
~975*
<azonenberg>
now smack in the middle of the 950 to 1050 mV allowed range
<azonenberg>
rise time is 3.03 to 3.32 ns, well within the 3-5 ns range
<azonenberg>
Peak jitter is probably somewhere between 1.25 and 1.5 ns (the spec limit is 1.4), but my scope is only 4 GSa/s so it's hard to be super accurate there
Bike_ has joined ##openfpga
<awygle>
Gotta improve jitter by another 0.25ns so you can claim compliance with your equipment :-P
<awygle>
More seriously, what's the primary source of jitter do you think? Just need to clean up the inout clock a bit more or what?
<azonenberg>
awygle: tbh? i think some of the jitter is my hack of a clock recovery postprocessor in my eye rendering code
<azonenberg>
lol
<azonenberg>
but the bigger issue is, my scope can only measure 250 ps timebase
<azonenberg>
so i may be in compliance already
<azonenberg>
i just cant measure one way or another
<awygle>
Well yeah but you can solve that problem by killing it. If you're in compliance it's not by more than 150ps currently (ignoring the clock recovery issue because lol)
<azonenberg>
well honestly i think i probably have 250+ ps of jitter in the clock recovery
<azonenberg>
what i do now is, i trigger on a rising edge
<azonenberg>
then dead-reckon based on my interpolated baud rate
<azonenberg>
for X number of samples
<azonenberg>
then resync
<azonenberg>
like a uart :D
<awygle>
Can you output the source clock and trigger the scope on that? Sync the timebases and avoid recovery altogether?
<azonenberg>
not exactly... among other things the scope can't take an input clock
<azonenberg>
i could trigger off an edge of the tx clock
<azonenberg>
but the scope isnt sampling at exactly 250ps per sample
<azonenberg>
it might be 249 or 251
<azonenberg>
so there will be drift
<azonenberg>
over a half million points or so a few samples of drift is almost assured
<awygle>
Sure, makes sense. I thought those fancy scopes had sync inputs but I've never actually used one
<azonenberg>
This is not a "fancy" scope
<azonenberg>
it's the cheapest one lecroy makes
<azonenberg>
that isnt like a relabeled rigol or something
<azonenberg>
(they have a really low end one that looks like it might be, lol)
* awygle
has a rigol
<azonenberg>
This one is only 350 MHz 4 Gsa/s and sells for around $5K
<azonenberg>
The line of scopes i really WANT to have are in the $20-50k class
<azonenberg>
Also, all of that fancy eye rendering you saw?
<azonenberg>
that's my own code
<awygle>
Of those numbers the only one that's not fancy to me is the 350 MHz
<azonenberg>
not the scope
<cr1901_modern>
only 350 MHz 4 Gsa/s, "only" $5k
<azonenberg>
cr1901_modern: list price
<azonenberg>
i paid decently less
<cr1901_modern>
I see
<awygle>
cr1901_modern++
<cr1901_modern>
And what's funny is this is still not enough to do modern protocol work
<azonenberg>
i think 3.2k for the scope itself
<cr1901_modern>
USB 2.0 is out. PCIe is out.
<azonenberg>
then a bunch more for the LA probe and some protocol decode options
<azonenberg>
you could probably do usb2 decoding, but probably not signal integrity
<azonenberg>
there'd be barely enough bandwidth to read the bits
<azonenberg>
480 Mbps = 240 MHz
<cr1901_modern>
480/2 ?
<azonenberg>
so the bits would look like sinusoids
<azonenberg>
but they'd be distinguishable
<azonenberg>
pcie/sata/usb3 are out though
<awygle>
I am sitting in front of a 20 Gsps scope whose software is so bad it is nearly unusable
<azonenberg>
My next scope will be 1+ GHz class, depending on my budget at the time
<azonenberg>
awygle: who makes it?
<azonenberg>
name and shame :p
<awygle>
Keysight
<cr1901_modern>
I wonder how BK Precision stacks up. I love their multimeters
<awygle>
Since I sent that text the UI has crashed.
<azonenberg>
lol
<azonenberg>
wow
<azonenberg>
i've crashed the ui on my lecroy too
<azonenberg>
but not super often
<awygle>
I just want to go home :-(
<awygle>
Lol
<azonenberg>
awygle: any experience with rohde & schwarz?
<azonenberg>
i have two R&S power supplies arriving tomorrow
<azonenberg>
awygle: if you're not writing home in anger, thats a good thing :p
<awygle>
I am about to order one or more BK power supplies
<azonenberg>
I considered bk and rigol
<azonenberg>
but the r&s had ethernet
<awygle>
I have used a programmable load from them that was great
<azonenberg>
and a few other nice options
<azonenberg>
i have a bk multimeter that i like
<azonenberg>
havent used any of their high end stuff
<azonenberg>
i got very close to buying the 9130 series
<azonenberg>
i forget what pushed me to the r&s
<awygle>
Iirc the BK I was looking at was Ethernet, and triple output. Can't remember tho.
<azonenberg>
oh i remember, it was just the voltage/current options
<cr1901_modern>
triple output?
<azonenberg>
i couldnt find quite the range i wanted
<cr1901_modern>
I have an old multimeter that's lasted me pretty faithfully on the same batteries the past 5+ years lol
<cr1901_modern>
BK multimeter*
<awygle>
cr1901_modern: three independent voltage outputs which can be paralleled for more current
<azonenberg>
i cant remember the last time i changed the battery in my bk multimeter
<awygle>
I have a uni-T that I picked up in Singapore
<awygle>
Again, fine. But very cheap.
<cr1901_modern>
I almost picked up a Fluke oscilloscope a long time ago, but decided against it
<cr1901_modern>
their portable oscilloscopes kinda skeeve me
<awygle>
I had a fluke 6 digit multimeter at a former job, it was good
<cr1901_modern>
How do you even use an analog scope without a repeating waveform (you don't?)?
<azonenberg>
cr1901_modern: lol, you dont
<azonenberg>
either that, or you get one with a really slow phosphor
<azonenberg>
and look at it quickly :p
<azonenberg>
they might have made ones with polaroid cameras for that
<awygle>
I think I need to become the azonenberg I want to see in the world and unify my test equipment into a useful software package
<cr1901_modern>
right
<cr1901_modern>
Not gonna help when I need a trigger
<cr1901_modern>
So the search continues
<azonenberg>
awygle: you know about STARSHIPRAIDER right? lol
<awygle>
I am
<cr1901_modern>
Well I may buy STARSHIPRAIDER depending on the price
<awygle>
*I do
<cr1901_modern>
and use it as an oscilloscope
<azonenberg>
There needs to be an analog capture board first
<azonenberg>
but it will happen
<awygle>
I'm more interested in the software than the hardware currently
<cr1901_modern>
how'd you capture your ethernet phy stuff then (it's using your custom software, no?)?
<awygle>
We just do pretty different things hw wise atm
<azonenberg>
cr1901_modern: my lecroy scope over ethernet
<azonenberg>
using my own software
<azonenberg>
i use the same app for my fpga-based internal LA
<azonenberg>
and starshipraider's LA will do the same as well
<awygle>
FUCK this scope okay I'm going home.
Bike_ is now known as Bike
<awygle>
azonenberg: is the code for scopey mcscopeface on your github?
<cr1901_modern>
Oh ffs, _that's_ what it's named?
<awygle>
No
<awygle>
I don't know its NSA code name tho
<cr1901_modern>
TRAGICBACKSTORY or something
<azonenberg>
no
<azonenberg>
tragiclaser is the ethernet bitbang project
<azonenberg>
the app itself actually doesnt have a name
<awygle>
Lmao 10/10 name
<azonenberg>
it lives in the antikernel repo
<azonenberg>
i've been meaning to fork it out to be standalone
<azonenberg>
one of the holdups is coming up with a name :p
<cr1901_modern>
It lives in the drawer made out of steak
<awygle>
Please call it TRAGICBACKSTORY
<azonenberg>
awygle: lol
<awygle>
That's amazing
* cr1901_modern
seconds that name
<cr1901_modern>
how do you open a drawer made of steak?
<azonenberg>
the steak goes in the drawer
<azonenberg>
its not the structure of the drawer
<cr1901_modern>
I see
<azonenberg>
also i really need to set up antikernel.net at some point for my email
<azonenberg>
i want to finally retire drawersteak.com
<azonenberg>
the company closed up shop 7 years ago and i held onto the domain ever since just for email
<awygle>
That name always reminds me of a webcomic I read in college by someone called tailsteak
<azonenberg>
most of my internal stuff is already under antikernel.net
<awygle>
Do you self host email?
<azonenberg>
I kinda inherited the godaddy account from drawersteak and have been looking to move off
<azonenberg>
i think i moved the DNS out already
<azonenberg>
but the mail is still there
<azonenberg>
During the migration to antikernel.net i want to either find a better host or set up a self-host
<azonenberg>
but in either case i'll be moving
<azonenberg>
been meaning to do it for years and just havent had time
<cr1901_modern>
I wonder what's the fastest LA I could get from my FPGAs alone (prob would need to take transmission line effects into account at that stage tho)
<cr1901_modern>
and Idr how to make an FPGA input "look like" 50 Ohms
<azonenberg>
first question: why the heck would you want to do this?
<rqou>
why not? e.g. gp4par on a webpage
<azonenberg>
could you do yosys there too?
<azonenberg>
if not, not much use
<rqou>
yes, but there might be issues
<azonenberg>
also it seems like a horrible idea in general :p
<rqou>
i'll work on that later
<azonenberg>
That said
<azonenberg>
it doesnt look like it will break anything we have now
<awygle>
cr1901_modern: for an input you could probably just parallel terminate 49R9
<openfpga-github>
[openfpga] azonenberg pushed 4 new commits to master: https://git.io/vFKSV
<openfpga-github>
openfpga/master 5f01ad6 Robert Ou: build: Add an option to disable building hardware-interfacing code
<openfpga-github>
openfpga/master 10ebdeb Robert Ou: emscripten: Mount NODEFS on /x if running under Node
<openfpga-github>
openfpga/master e836f41 Robert Ou: Update to latest logtools
<cr1901_modern>
parallel terminate?
<awygle>
49R9 to ground near the input
<awygle>
The input is high Z so the resistor dominates
<cr1901_modern>
Oh a shunt resistor
<awygle>
Yeah
<cr1901_modern>
and cool, makes sense
<cr1901_modern>
Now if I could remember that math for more than 2 seconds (I got an A in the course somehow, so clearly at one point in my life I understood it)...
<cr1901_modern>
I could work thru examples instead
<awygle>
I literally Google a voltage divider calculator every single time
<awygle>
My brain has better things to do
<awygle>
(in theory at least)
m_w has quit [Ping timeout: 240 seconds]
m_w has joined ##openfpga
ZipCPU|Laptop has quit [Ping timeout: 240 seconds]
<rqou>
not sure if that's an asm.js problem or a browser problem
<rqou>
azonenberg: do you have any idea if your code happens to have some global state such that invoking main n times causes all output to duplicate itself n times?
<openfpga-github>
[openfpga] rqou pushed 1 new commit to master: https://git.io/vFKFO
<openfpga-github>
openfpga/master 19eede5 Robert Ou: contrib: Add a demo of using emscripten-compiled gp4par
<azonenberg>
rqou: not that i know of?
<rqou>
ok, well i'll have to hunt it down :P
<rqou>
azonenberg: it's g_log_sinks
<rqou>
that... needs some fixes anyways for emscripten, so i'll ignore this problem for now
Bike has quit [Quit: Lost terminal]
<azonenberg>
rqou: invoking main more than once is generally not something most c programs expect :p
<azonenberg>
long term librarizing gp4par is definitely planned, i think i already have a ticket open from years ago for that
<rqou>
yes, i know that that usually doesn't work
<rqou>
at least we found another potential gotcha for library-izing
<rqou>
anyways, did you at least try the demo?
<azonenberg>
No, i've been busy debugging the RX side of my 100M PHY
<azonenberg>
i had to tweak a few passive values but it seems like the input buffers are working, now to try actually demodulating
digshadow has quit [Ping timeout: 252 seconds]
sgstair has quit [Ping timeout: 260 seconds]
sgstair has joined ##openfpga
<awygle>
azonenberg: you're going to formally verify the core when you're done right? :-P
<azonenberg>
awygle: might be tricky as there's a lot of analog stuff
<azonenberg>
most of the potential failures are there
<azonenberg>
i mean it's rtl
<azonenberg>
but it's rtl directly controlling analog behavior through external passives
<azonenberg>
the MAC i could certainly verify if i had time
<azonenberg>
the PHY would be harder
<azonenberg>
i might even be able to verify the PMA
<azonenberg>
the PCS*
<azonenberg>
but the PMA is the hard part
sgstair has quit [Ping timeout: 248 seconds]
sgstair has joined ##openfpga
sgstair has quit [Ping timeout: 248 seconds]
sgstair has joined ##openfpga
laintoo__ has quit [Ping timeout: 260 seconds]
sgstair has quit [Ping timeout: 240 seconds]
digshadow has joined ##openfpga
sgstair_ has joined ##openfpga
sgstair has joined ##openfpga
sgstair_ has quit [Ping timeout: 268 seconds]
soylentyellow has joined ##openfpga
soylentyellow has quit [Remote host closed the connection]
sgstair_ has joined ##openfpga
sgstair has quit [Disconnected by services]
digshadow has quit [Ping timeout: 258 seconds]
sgstair_ has quit [Ping timeout: 260 seconds]
sgstair_ has joined ##openfpga
laintoo__ has joined ##openfpga
sgstair has joined ##openfpga
sgstair_ has quit [Ping timeout: 240 seconds]
digshadow has joined ##openfpga
sgstair has quit [Ping timeout: 240 seconds]
sgstair_ has joined ##openfpga
sgstair_ is now known as sgstair
laintoo__ has quit [Ping timeout: 248 seconds]
laintoo__ has joined ##openfpga
sgstair has quit [Ping timeout: 268 seconds]
sgstair has joined ##openfpga
laintoo__ has quit [Ping timeout: 250 seconds]
sgstair_ has joined ##openfpga
laintoo__ has joined ##openfpga
sgstair has quit [Disconnected by services]
sgstair_ has quit [Ping timeout: 250 seconds]
sgstair_ has joined ##openfpga
sgstair_ has quit [Ping timeout: 240 seconds]
laintoo__ has quit [Ping timeout: 240 seconds]
laintoo__ has joined ##openfpga
sgstair_ has joined ##openfpga
sgstair_ is now known as sgstair
soylentyellow has joined ##openfpga
sgstair_ has joined ##openfpga
sgstair has quit [Disconnected by services]
sgstair_ is now known as sgstair
sgstair_ has joined ##openfpga
sgstair has quit [Disconnected by services]
sgstair_ is now known as sgstair
<azonenberg>
Sooo in case anybody is wondering why i have UPSes and generators... :P
sgstair has quit [Remote host closed the connection]
Hootch has joined ##openfpga
pie_ has joined ##openfpga
laintoo__ has quit [Ping timeout: 260 seconds]
pie_ has quit [Ping timeout: 248 seconds]
laintoo__ has joined ##openfpga
pie_ has joined ##openfpga
laintoo__ has quit [Ping timeout: 240 seconds]
pie_ has quit [Ping timeout: 248 seconds]
sn00n_ is now known as sn00n
laintoo__ has joined ##openfpga
eduardo_ has quit [Remote host closed the connection]
laintoo__ has quit [Ping timeout: 240 seconds]
laintoo__ has joined ##openfpga
xdeller has quit [Ping timeout: 250 seconds]
laintoo__ has quit [Ping timeout: 248 seconds]
laintoo__ has joined ##openfpga
xdeller has joined ##openfpga
xdeller has quit [Quit: Leaving]
laintoo__ has quit [Ping timeout: 264 seconds]
laintoo__ has joined ##openfpga
laintoo__ has quit [Ping timeout: 258 seconds]
sn00n has quit [Ping timeout: 248 seconds]
laintoo__ has joined ##openfpga
xdeller has joined ##openfpga
<gruetzkopf>
the power backup solution i have is totally overkill and totally unneccessary
laintoo__ has quit [Ping timeout: 260 seconds]
laintoo__ has joined ##openfpga
<Zorix>
yea power here has brownouts almost every day and goes out a few times a month.. its crap.. i have a battery on everything
<qu1j0t3>
buy candles
laintoo__ has quit [Ping timeout: 248 seconds]
laintoo__ has joined ##openfpga
laintoo__ has quit [Ping timeout: 248 seconds]
laintoo__ has joined ##openfpga
laintoo__ has quit [Ping timeout: 264 seconds]
laintoo__ has joined ##openfpga
sn00n has joined ##openfpga
pie_ has joined ##openfpga
m_t has joined ##openfpga
laintoo__ has quit [Ping timeout: 268 seconds]
laintoo__ has joined ##openfpga
laintoo__ has quit [Ping timeout: 240 seconds]
laintoo__ has joined ##openfpga
digshadow has quit [Ping timeout: 248 seconds]
Hootch has quit [Quit: Leaving]
<azonenberg>
gruetzkopf: i got my UPS after the first outage at my old place in five years
<azonenberg>
here? i had six outages in six months after i moved here
<azonenberg>
and that's only counting the 8+ hour ones
<azonenberg>
not the brief glitches
<azonenberg>
Power reliability here leaves something to be desired :p
<azonenberg>
although, that was a particularly bad storm season
<azonenberg>
its not quite as bad now
sgstair has joined ##openfpga
<awygle>
Although work didn't lose power yesterday apparently my apartment did, briefly. Long enough to reset clocks and upset PCs but not long enough to defrost the freezer.
digshadow has joined ##openfpga
<shapr>
In my lifetime I've had the most problems with brownouts, I've often thought a five minute UPS for the entire house would make sense
Zorix has quit [Ping timeout: 246 seconds]
<azonenberg>
shapr: i've never had a sustained brownout
<azonenberg>
i've had glitches where power went out then on again after a couple of cycles
<azonenberg>
very brief brownouts when a line was shorting to a tree or something but not hard enouhg to pop the upstream breaker (yet)
<azonenberg>
most of the time it's trees either shorting to a wire or tugging on it
<azonenberg>
the sustained outages are usually like a tree falling across a road and taking out an entire section of line
<azonenberg>
and sometimes some poles
<azonenberg>
I'm actually looking at extending my UPS coverage when i move from ~10 minutes to 6-8 hours (my rough numbers say that would be around 12U or $3500 of batteries)
<azonenberg>
This would give me enough batter to run through a full night of no power before starting the generator first thing in the morning
<shapr>
I lived in Alabama for a bunch of years, that's where I experienced the brownouts
<azonenberg>
Would be less of an issue if i got a less noisy generator that wouldnt wake up the neighbors :p
<qu1j0t3>
why would you need to run things over the whole outage period? freezers?
<azonenberg>
Computers running ongoing tasks, freezers, ideally some HVAC but that would need more battery than i can afford
<azonenberg>
Communications (a lot of telephone service here is VoIP that goes down if you lose local mains power)
<azonenberg>
Anything i have running for work
<azonenberg>
i work from home a lot and being forced to go into the office during a power outage (with the asscoaited traffic jams etc) would be annoying
<awygle>
Brownouts feel quaint and old timey to me because I experienced them in California when I was very young and haven't really since
<awygle>
azonenberg: are you a ham? Seems like a radio or two would be a good investment for comms in that situation
<azonenberg>
awygle: KD2HKV
<azonenberg>
i hop on the 9 o'clock net on the PSRG every now and then during my ride in to work
<azonenberg>
But i'm not too active other than that and emcomm nets
<azonenberg>
i find analog voice kinda boring
<azonenberg>
i'm a lot more interested in e.g. radar :p
<azonenberg>
But i havent had the time or budget to build that 8x8 element 5 GHz active-scanning phased array i've always wanted to do :p
<awygle>
Yeah same
<awygle>
I don't even have a radio currently
<awygle>
Not counting SDRs I guess, could be pressed into service in an emergency
<awygle>
But I live in a metro area :-P
<azonenberg>
Yeah out in kitsap things are ... different
<azonenberg>
So lets see, 5 GHz is 6 cm so 5.8 GHz is probably closer to 5 cm wavelength
<azonenberg>
so i'd need about 40x40 cm of space for an 8x8 phased array
<azonenberg>
and then 64 power amps, 64 microwave delay lines...
<azonenberg>
yeah wouldnt be cheap :p
digshadow has left ##openfpga [##openfpga]
<awygle>
Why 5GHz?
digshadow has joined ##openfpga
xdeller has quit [Ping timeout: 240 seconds]
<rqou>
azonenberg: cells_sim_wip.v doesn't parse in upstream yosys, i hope you know this?
nrossi has quit [Quit: Connection closed for inactivity]
<azonenberg>
rqou: :o
<azonenberg>
i thought i had fixed that?
<azonenberg>
sec
* azonenberg
merges upstream
<openfpga-github>
[yosys] azonenberg pushed 15 new commits to master: https://git.io/vFi1B
<openfpga-github>
yosys/master dd46d76 Clifford Wolf: Fix a bug in yosys-smtbmc in ROM handling
<openfpga-github>
yosys/master c672c32 Clifford Wolf: Capsulate smt-solver read/write in separate functions
<azonenberg>
Pretty sure that's actually a yosys bug, i forget if i filed that upstream
<azonenberg>
Clifford had some bugs related to floats a while back
<azonenberg>
But let me check the LRM
<azonenberg>
to see if this is legal
<azonenberg>
OK, yes
<azonenberg>
The LRM confirms delay_value can be an unsigned, a real, or an identifier
<azonenberg>
So yosys is broken here
teepee has quit [Ping timeout: 252 seconds]
<awygle>
I have a very stupid question. When I was learning Verilog in school I used to get a warning pretty frequently that was basically "hey, this inferred a latch, you probably don't want that". But since I started up again recently I never get that warning. Did I get good or did something change? What infers a latch?
<azonenberg>
I have a temporary fix in azonenberg/yosys that works around this (by changing the delay to an integer) but yosys upstream really needs to just fix that
<azonenberg>
awygle: a combinatorial block that doesn't assign all cases
<azonenberg>
always @(*) if(foo) bar <= baz;
<azonenberg>
if you have an "else" case, this turns into a 2:1 mux with "foo" as the select
<azonenberg>
If you do NOT have one, it turns into a latch with "foo" as the latch enable
<awygle>
azonenberg: ah thanks. So I'm just doing enormously less combinatorial stuff than I did in school :-P
<awygle>
rqou: I support your madness. Add a text editor.
<awygle>
And that svg visualizer that somebody posted
<awygle>
And get iverilog to build under emscripten too
<awygle>
:-P
<rqou>
doesn't iverilog have a jit? that would be a lot harder or slower
<awygle>
It compiles to byte code ahead of time iirc
<rqou>
it doesn't jit vvp to native code?
<awygle>
Uhh it might actually idk. I thought vvp was interpreted
<rqou>
hmm, looks like it is
<rqou>
vpi.js? (only half kidding)
xdeller has joined ##openfpga
<awygle>
Whoops. don't set your chipscope to distributed ram
<rqou>
anybody know what the tool du jour is for converting "nodejs stuff" to "in-browser stuff"?
* jn__
heard it's called "browserify"
<azonenberg>
awygle: lol
<azonenberg>
i imagine that would get big
<jn__>
(but i really *really* don't know much about javascript…)
<awygle>
rqou: I thought you could just automatically use everything because Javascript *shrug*
<awygle>
azonenberg: yeah PAR was still running when I got back from lunch and a meeting. 16k bits of distributed ram. Ooooops
<rqou>
no, there's at least three javascript "universes" that i know of with various degrees of overlap
<rqou>
nodejs, browser, and browser add-ons
<rqou>
and also the actionscript universe that nobody likes or cares about
<azonenberg>
actionscript is js?
<rqou>
it's ecmascript with extensions
Zorix has joined ##openfpga
<awygle>
eczemascript
<awygle>
if you will
* qu1j0t3
snerks
<rqou>
it feels like every time i step away from the web ecosystem and then look back in, everything gets 2x more insane
<awygle>
My current working theory is that there are sane things in the web ecosystem that are just swamped by all the idiocy. There's a functional system in there if you can extract it"
<awygle>
But I barely do web so I'm probably wrong.
<qu1j0t3>
no, i think that's true.
<qu1j0t3>
there may be disagreement on what those things are, but I'd nominate things like Elm, Purescript, Bucklescript/jsoo etc
<qu1j0t3>
still in flux but well engineered
<awygle>
I like the idea of things like WebAssembly and Service Workers
<awygle>
Languages generally are not what I'm looking for. Especially not languages that compile to JS (an irrational bias)
<qu1j0t3>
I'll do a lot not to have to write plain JS.
<awygle>
qu1j0t3: are you involved in Clash or some other nuHDL? I feel like I remember that
<qu1j0t3>
i'm kinda HLHDL-curious but haven't really made it a priority yet
<qu1j0t3>
dabbling in analog for a while
<qu1j0t3>
awygle: not involved in Clash but yeah i'm open to using Scala or Haskell as a substrate language for it
<awygle>
I think language is not as important to me as other people? I regularly write Python which I hate, for example
<qu1j0t3>
well, saying you hate python kind of puts language on the table. (I'm not a fan either)
<awygle>
Just not super interested in HLHDL or BetterScripts
<rqou>
i quite like Python
<qu1j0t3>
you mean you 're content with verilog?
<awygle>
*shrug*
<awygle>
Yeah. I'm not very motivated to explore alternatives to it.
<qu1j0t3>
nobody will judge you for that.
<rqou>
best kitchen sink language for anything that doesn't involve GUIs
<awygle>
I would rather spend effort on improving compatibility of verilog implementations, and/or support for SV
<awygle>
On the other hand my code just failed (after synthesis and programming) because of a localparam being defined twice. So there may be a Stockholm element :-P
<rqou>
meanwhile, imo "html5+js" is the most portable and easiest to deploy GUI framework
<rqou>
even though it's a giant trashfire
<awygle>
rqou: and the one with the largest available talent pool on the design side
<rqou>
you mean a better design than my demo yosys.html? :P
<awygle>
Quite :-P
<awygle>
I was daydreaming about what "atom but not abusively slow" might look like the other day
<awygle>
I would like to try compiling like, xi with emscripten and embedding it in a web GUI thing
indy has quit [Ping timeout: 240 seconds]
indy has joined ##openfpga
m_t has quit [Quit: Leaving]
<rqou>
wait wait webusb replaced their idiotic "extra special usb descriptors" mechanism with just a normal security toaster
<rqou>
now it's actually usable
<rqou>
(except Chrome-only)
promach has quit [Ping timeout: 250 seconds]
<rqou>
oh hey azonenberg you're "infosec" right? what are your feelings on notifications like "this program would like to blah blah blah blah incomprehensible blah blah. do you want to allow this?"
<azonenberg>
First choice: don't provide dangerous functionality at all, so no need to warn about it
<azonenberg>
second choice: make the warning comprehensible
<rqou>
but what if the point of the feature is to do something dangerous?
gnufan has joined ##openfpga
<rqou>
e.g. the webex emergency fix
<rqou>
"the user must now click ok for arbitrary code to be executed"
<azonenberg>
I would then argue the feature should not have been implemented in the first place
* sgstair
awaits the inevitable report "Accessing the internet considered harmful"
<azonenberg>
Executing arbitrary code downloaded from an untrusted soruce
<azonenberg>
aka javascript
<azonenberg>
should be considered harmful IMO :p
<azonenberg>
sure you can limit the APIs and sandbox it
<azonenberg>
but it's fundamentally a high-risk operation
<rqou>
but i see you using websites with js enabled :P
<sgstair>
indeed
<azonenberg>
rqou: yes, unfortunately most websites dont work without js
<rqou>
e.g. yosys.html? :P :P :P
<azonenberg>
i didnt actually test that
<azonenberg>
But i whitelist where i have to, then run my browsers in a VM and hope for the best
<sgstair>
when the singularity comes, it will propogate through TLS and JS bugs
<rqou>
on light bulbs :P
<sgstair>
probably :)
promach has joined ##openfpga
xdeller has quit [Quit: Leaving]
m_w has quit [Quit: leaving]
Bike has joined ##openfpga
wpwrak has quit [Ping timeout: 248 seconds]
wpwrak has joined ##openfpga
soylentyellow has quit [Ping timeout: 248 seconds]
gnufan has quit [Quit: Leaving.]
soylentyellow has joined ##openfpga
m_w has joined ##openfpga
<lain>
:D
<lain>
there's a great part in Person of Interest <spoilers>where the two super-intelligent true AIs are battling it out, and one has to go into hiding... so it arranges for city contractors to install smart grid devices on all the electrical poles... except what it's doing is distributing itself among those devices, so it can't be taken down by a single datacenter being destroyed, for example :P</spoilers>