azonenberg_work has quit [Ping timeout: 266 seconds]
tecepe has joined ##openfpga
doomlord has quit [Ping timeout: 260 seconds]
<cyrozap>
rqou: "oflag=direct" skips the cache, then "bs=4M" or some other size for performance, then "status=progress" to watch the status of the transfer.
digshadow has quit [Quit: Leaving.]
amclain has quit [Quit: Leaving]
doomlord has joined ##openfpga
doomlord has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
doomlord has joined ##openfpga
<azonenberg>
cr1901_modern: read the rest of the thread
<azonenberg>
turns out there is literally a kernel component called the bowser
<azonenberg>
b/c having a usermode and kernel component with the same name messes up debuggers or osmething
<azonenberg>
Why the computer browser is located in the kernel at all is another question...
carl0s has joined ##openfpga
<carl0s>
hi cyrozap, are you giving PSoC support on libopencm3 already or it's a near future plan? would love to contribute if it's not a problem.
digshadow has joined ##openfpga
digshadow has quit [Quit: Leaving.]
<cyrozap>
carl0s: I've started adding support, but I don't have anything buildable yet.
<carl0s>
Somewhat similar to the support for EFM32 micros (core only support)?
digshadow has joined ##openfpga
m_w has quit [Quit: leaving]
m_w has joined ##openfpga
azonenberg has quit [Ping timeout: 256 seconds]
azonenberg has joined ##openfpga
JvD__ is now known as JvD_
azonenberg has quit [Ping timeout: 256 seconds]
DocScrutinizer05 has quit [Disconnected by services]
<cr1901_modern>
azonenberg: I am going to reject that reality and substitue my own!
<rqou>
i wonder what angler ek will do on this machine? :P
<azonenberg>
rqou: i would laugh if modern malware won't run
<azonenberg>
b/c it is missing DLLs or API functions
<rqou>
um... no "DLLs" on this system
<azonenberg>
oh yeah thats a mac
<azonenberg>
well, whatever the equivalent is
<rqou>
i'm not sure os 9 even had shared libraries at all
<rqou>
other than the system apis
<rqou>
also, I hope you're aware this is a PPC machine :P
<rqou>
not x86
<azonenberg>
Lol oh
<azonenberg>
yeah for some reason i thought it was win 9x
<azonenberg>
so i assumed x86
<azonenberg>
You're trivially pwnable if someone triees
<azonenberg>
but... nobody is likely to try :p
<rqou>
yes definitely
<rqou>
os 9 barely even has the concept of "passwords"
<rqou>
btw protip for those people dealing with ancient flaky cd rom drives: you really do need to decrease your disc burner speed
<rqou>
also, you probably shouldn't use cd-rws
<azonenberg>
you're saying burning at a lower speed makes the disks more readable?
<azonenberg>
like, the dots are darker or something?
<rqou>
yeah, it really does seem to
kuldeep has joined ##openfpga
<azonenberg>
would be cool to SEM one of the chips
<azonenberg>
one of the disks*
<azonenberg>
and see how they look
<azonenberg>
or even just high res optical imaging
<rqou>
i don't know about the effect of speed, but cd-rws are significantly less reflective than normal cds
<azonenberg>
whitequark: around?
<azonenberg>
wanted to discuss HiL testing in more depth
<rqou>
according to wikipedia, faster cd drives also switch from CLV to CAV
<azonenberg>
That makes sense
<azonenberg>
accelerating a disk is time consuming
<rqou>
this ancient mac's drive is right at the cutoff
<rqou>
so forcing the burner to burn slow may kick it back to CLV
<azonenberg>
Fun
<cr1901_modern>
yeah for some reason i thought it was win 9x <--- really ._.?!
<cr1901_modern>
azonenberg: REALLY?! T_T
<cr1901_modern>
I suppose even back then you were using Linux :P?
<azonenberg>
cr1901_modern: no i just didnt look at the pic too closely
<azonenberg>
i saw IE in the title bar
<azonenberg>
now that i look its obviously a mac
<azonenberg>
i forgot ie for mac was even a thing
<rqou>
it's part of the normal os install even
<azonenberg>
and forgive me, i never actually ran win 9x :p
<rqou>
as well as stuffit expander
<azonenberg>
i used 3.1 and dos 5
<rqou>
apple apparently wasn't as NIH before :P
<azonenberg>
then windows mistake edition
<azonenberg>
then XP, and vista
<cr1901_modern>
mistake edition was the best Windoze
<azonenberg>
jumped to linux around that time, kept one win7 box in the lab before moving to 100% virtualized
<azonenberg>
havent run anything newer than w7 natively
<azonenberg>
and what?
<azonenberg>
it BSOD'd constantly
<azonenberg>
least stable os i've ever used
<cr1901_modern>
Poe's Law :P
<rqou>
btw this imac i have isn't a configuration that is supposed to exist
<rqou>
it's a 333mhz imac in bondi blue
<rqou>
my father got it from one of his friends who was a tester at apple
<cr1901_modern>
And tbh, I used to think Windows 95 was the best thing since sliced bread, b/c it could play Math Blaster and other edutainment games that didn't really help much. It was actually pretty bad in retrospect.
<azonenberg>
lol
<azonenberg>
what about number munchers?
<rqou>
lol i played that in elementary school
<rqou>
on a mac
* azonenberg
fondly remembers long nights in the basement of the rpi e-club
<cr1901_modern>
Never played those T_T
<azonenberg>
playing that on an apple 2
<azonenberg>
in... 2009?
<cr1901_modern>
Prob will have to go to archive.org and try it
<rqou>
i remember in elementary school i was playing it on a slightly newer imac g3
<cr1901_modern>
I mean, I do still play oregon trail from time to time :3
<azonenberg>
i wonder what happened to that when the eclub went down the toilet
<azonenberg>
hope it ended up with one of the members and not the trash
<cr1901_modern>
I most certainly hope so :'(
<azonenberg>
pretty sure the sgi onyx ended up getting junked
<azonenberg>
but that wasn't in bootable conditoin anymore
<azonenberg>
people were trying to restore it
<azonenberg>
but it was too big and powrer hungry for someone to take home
<azonenberg>
In other news i'm considering acquiring (either by buying them or by getting silego to donate) a couple of greenpak devkits for the rpi electronics club
<azonenberg>
i mean, embedded hardware club
<rvense>
the onyx was like.. small closet, wasn't it?
<azonenberg>
if i donate a couple i might get people beta testing :)
<azonenberg>
and it was a full height rack
<azonenberg>
so, yeah about the size of a broom closet
<rvense>
imagine it made a nice little racket when you switched it on... never seen it run, but there was one in a corridor(!) under uni, about ten years ago
<rvense>
seemed like it'd basically been left there by someone who'd just given up on moving it any further and then pushed it against the wall
<azonenberg>
lol
<rqou>
anyways, I wonder how possible it is to get NT/PPC on the imac
<rqou>
if it works (highly unlikely) then that might be the most unholy combination possible :P
<azonenberg>
cr1901_modern: my experience has been that you rarely run out of ram by just a little bit
<azonenberg>
and if you go into swap, you're doing so many Gbps of disk read/write that the whole system is unusable
<azonenberg>
if you're lucky you can get enough of a shell to kill -9 the offending process
<azonenberg>
and not have to power-cycle to recover control
<azonenberg>
much better to just segfault one process
<azonenberg>
and recover
<cr1901_modern>
azonenberg: Well yes, but it still seems kinda wasteful to keep everything in RAM at all times when you're only using 5% of a program :P
<cr1901_modern>
Also, I would like to stem the tide of "f*** it, just add more RAM"
<azonenberg>
cr1901_modern: well i hate situations where i havent used an app for a while
<azonenberg>
and it takes 30 sec to page in before i can use it
<cr1901_modern>
Sure. Not every application is interactive. I was thinking more along the lines of BG processes. Though I suppose code segment sharing mitigates this
<azonenberg>
I would rather see exokernels take over
<azonenberg>
and have apps explicitly swap out data they don't plan to use for a while
<azonenberg>
say, a web browser can page out tabs that havent been used for 10 minutes
<azonenberg>
by serializing the DOM and freeing the ram
<cr1901_modern>
Exokernels seem really nice. I'm not sure I'd want to use one as a general purpose OS
<azonenberg>
then if the tab is selected you wake it back up
<azonenberg>
i may be a bit biased as my phd work was on an exokernel-y system
<azonenberg>
but they seem far superior to the current state of the art :p
<azonenberg>
you can implement a compatibility layer on one that looks like a normal OS
<azonenberg>
then selectively bypass that layer per app to get much better performance
<azonenberg>
if you're so inclined
<cr1901_modern>
Maybe an OS that provides a browser and not much else would be a good candidate for an exokernel
<azonenberg>
in general i think modern systems are waaay too resource hungry
<azonenberg>
and excessively complex / failure prone
<cr1901_modern>
azonenberg: I'm not arguing there :P
<azonenberg>
So simplifying is the way to go
<azonenberg>
also, the discussion probably belongs in #antikernel :p
<cr1901_modern>
My bias is that I still don't quite know how a modern OS works and I'd like to figure out how it works (without spending months in BSD or Linux source code) BEFORE it inevitably gets replaced so I don't feel left out
<rqou>
you have too many different irc channels :P
<cr1901_modern>
"All the cool kids know how an OS works"
* cr1901_modern
isn't cool :(
<azonenberg>
rqou: well i don't want to spam one chan with github notifications or discussion of another project
<azonenberg>
i like modularity
<cr1901_modern>
azonenberg: I have 11 IRC tabs open already
kuldeep has joined ##openfpga
<azonenberg>
I only have 12 right now
<azonenberg>
not counting PMs
kuldeep has quit [Ping timeout: 244 seconds]
carl0s has quit [Quit: Leaving]
kuldeep has joined ##openfpga
cr1901_modern1 has joined ##openfpga
cr1901_modern has quit [Ping timeout: 245 seconds]
<azonenberg>
Sooo i just set up a test VM for running hardware-in-loop nightly builds
<azonenberg>
whitequark: ^
<azonenberg>
it's not actually doing anything yet
<azonenberg>
PM me to discuss setup
kuldeep has quit [Ping timeout: 245 seconds]
kuldeep has joined ##openfpga
massi has joined ##openfpga
kuldeep has quit [Ping timeout: 250 seconds]
kuldeep has joined ##openfpga
qu1j0t3 has quit [Ping timeout: 256 seconds]
<azonenberg>
I can set the board up on it easily, but i have to figure out how to share nightly build jobs vs interactive debug
<azonenberg>
i could always throw on something like SLURM but it seems like overkill for a single node
qu1j0t3 has joined ##openfpga
kuldeep has quit [Ping timeout: 245 seconds]
<openfpga-github>
[openfpga] azonenberg pushed 1 new commit to master: https://git.io/vPp8M
<openfpga-github>
openfpga/master 139eb51 Andrew Zonenberg: greenpak4: refactoring of Greenpak4LFOscillator for SLG46140 support
<openfpga-github>
[openfpga] azonenberg force-pushed gh-pages from 0be1b45 to 3d9a493: https://git.io/v6vmV
<openfpga-github>
openfpga/gh-pages 3d9a493 Travis CI User: Update documentation
<travis-ci>
azonenberg/openfpga#149 (master - 139eb51 : Andrew Zonenberg): The build passed.
massi has quit [Remote host closed the connection]
<whitequark>
hum, where did azonenberg go
digshadow has joined ##openfpga
azonenberg has joined ##openfpga
m_w has joined ##openfpga
azonenberg_work has joined ##openfpga
dalias has quit [Read error: Connection reset by peer]
X-Scale has quit [Ping timeout: 256 seconds]
X-Scale has joined ##openfpga
X-Scale has quit [Read error: Connection reset by peer]
X-Scale has joined ##openfpga
tecepe has quit [Ping timeout: 244 seconds]
tecepe has joined ##openfpga
X-Scale has quit [Ping timeout: 256 seconds]
<cr1901_modern>
azonenberg: What would cause Kicad to decide to refuse to let me place certain components in a loaded library (when I know damn well the part exists)?
<azonenberg>
this is one of many reasons i use my own libraries exclusively
<rqou>
welcome to kicad levels of stability :P
<cr1901_modern>
JFC... the entire EDA industry needs to burn to the ground. It's unusable
<azonenberg>
cr1901_modern: why do you think this channel exists?
<azonenberg>
:p
<azonenberg>
also i like where kicad is going
<azonenberg>
but... it's not there :p
<rqou>
not "to be acquired by Cadence in the future?" :P
<rqou>
lots of EDA software goes to die in Cadence :P
<azonenberg>
if cadence eats kicad i'm forking it :p
<cr1901_modern>
All I wanted to do was move a goddamn git repository into another directory. Should've taken 5 mins at most. Now I'm on hour 2 of "I just want to fucking work on my schematic"
<azonenberg>
fuuun
<cr1901_modern>
The component is literally gone o.0; I didn't update anything, or change any settings. The last_modified date for ftdi is July 2015. What. The. Actual. Fuck?!
<azonenberg>
did you load the wrong ftdi lib?
<rqou>
and this is why embedded/EDA loves to a) vendor everything b) never update anything
<cr1901_modern>
I don't know... how many are there?
<azonenberg>
well i mean
<azonenberg>
if you had one in the project dir / your home dir
<azonenberg>
and one in the kicad install
<azonenberg>
or something like that
<cr1901_modern>
azonenberg: The thing is, ftdi.dcm has the part information
<cr1901_modern>
But ftdi.lib doesn't think it exists
<azonenberg>
...
<rqou>
btw is kicad done with their weird format migration thingamajig yet?
<cr1901_modern>
azonenberg: My thoughts exactly
<azonenberg>
rqou: last i checked they were pretty done with pcb in that regard
<azonenberg>
but schematic is in the middle of a massive refactoring
<azonenberg>
i'm kinda waiting for that to finish, working on other stuff
* cr1901_modern
throws laptop out window
<azonenberg>
before i go do much
<rqou>
why does it seem like almost all software (OSS, proprietary, free, paid, etc.) fails to consider migration plans?
<balrog>
[19:38:48] <rqou>and this is why embedded/EDA loves to a) vendor everything b) never update anything
<balrog>
yay for security
<azonenberg>
rqou: fwiw
<azonenberg>
this is one of the reasons why i like to keep the number of users of a project i am working on small
<azonenberg>
early on
<azonenberg>
So that i can be agile and not have to worry about inconvenienceing users if i make a breaking chang
<azonenberg>
change*
<azonenberg>
But once people start using it, that's not an option
<azonenberg>
or well, it shouldn't be :P
<cr1901_modern>
Haha... hahaha!! HAHAHAHA!!!
* cr1901_modern
laughs maniacally
<cr1901_modern>
The part's ALSO in ftdi.lib
<cr1901_modern>
fuck you kicad
<cr1901_modern>
Why must life be so hard?! Why must I fail at every attempt at PCB design?!
<rqou>
my worst migration experience so far was ubuntu 14.04 to 16.04
<rqou>
these are both LTS versions
<rqou>
but 16.04 has systemd with basically no transition plan whatsoever from upstart
<cr1901_modern>
I'm done for tonight. Yet ANOTHER project I'm blocking on waiting for additional information.
<rqou>
the LTS upgrade path is by default to the next LTS
<rqou>
so i wouldn't have gotten systemd in 15.04
<azonenberg>
rqou: i dislike systemd just because it's trying to turn linux into windows
<azonenberg>
"windows kernel font parser" :p
<cr1901_modern>
azonenberg: I appreciate that you make your own parts. That's not the right solution from a UI perspective IMHO.
<rqou>
hey, win10 put parts of it back into userspace
<azonenberg>
cr1901_modern: that is the only solution in a serious eda team unless the library was acquired from a source that does proper QA according to IPC standards etc
<rqou>
also, FOSS has libfreetype which isn't really cause for celebration
<azonenberg>
rqou: i didnt say *nix was any better...
<rqou>
(e.g. the amazing iOS 4 jailbreak)
<rqou>
where assholes went through apple stores jailbreaking all the display iphones
<rqou>
forcing staff to reimage them all every day
<azonenberg>
lool
<azonenberg>
also freetype at least runs in userland right?
<rqou>
sure
<azonenberg>
or is it in the x server as root or something stupid
<rqou>
no, it's in the normal application
<cr1901_modern>
azonenberg: Found the problem. It's loading the wrong goddamn library
<azonenberg>
cr1901_modern: thought so
<rqou>
there was some other exploit to go into kernel mode
<azonenberg>
oh
<cr1901_modern>
Despite me adding the correct library
<rqou>
i was particularly impressed with that ios jailbreak though because it used the font postscript bytecode to defeat aslr
<cr1901_modern>
Looks like kicad has this "lovely" feature where the default libs can be shadowed by libs of the same name, even if the default lib has parts that the shadowing lib does not
<rqou>
scoping is hard
<rqou>
it's arguably a subset of naming things
<cr1901_modern>
Well, I'm talking to the alternate library maintainer right now, so at least that's one thing going right