<_whitenotifier-f>
[scopehal] azonenberg pushed 4 commits to master [+2/-0/±17] https://git.io/JUhhm
<_whitenotifier-f>
[scopehal] azonenberg 4b60d54 - Initial work on SD card command bus decode. See #169.
<_whitenotifier-f>
[scopehal] azonenberg 0e26669 - CreateMergedHeader now takes index field
<_whitenotifier-f>
[scopehal] azonenberg 012598d - CanMerge now takes first/cur/next arguments
<_whitenotifier-f>
[scopehal] azonenberg 201cc51 - Finished basic implementation of SD card command bus decode. See #169.
<_whitenotifier-f>
[scopehal-docs] azonenberg pushed 1 commit to master [+0/-0/±3] https://git.io/JUhhs
<_whitenotifier-f>
[scopehal-docs] azonenberg 3642f68 - Skeleton headers for demo driver and SD command bus decode
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±3] https://git.io/JUhhn
<_whitenotifier-f>
[scopehal-apps] azonenberg 27e0e5f - ProtocolAnalyzerWindow: fixed some merge bugs and a potential segfault if data column isn't shown
<azonenberg>
lain awygle: my thought at the moment is to have the command bus be standalone, then the data bus will take in command bus + clock/data as inputs
<azonenberg>
so you can map LBAs on the command bus to read data
<azonenberg>
i cant think of a reason to allow decoding data without commands
<azonenberg>
but commands without data could be useful if you're constrained by # of probes
alexhw has quit [Ping timeout: 244 seconds]
alexhw has joined #scopehal
juli965 has joined #scopehal
bvernoux has joined #scopehal
bvernoux1 has joined #scopehal
bvernoux is now known as Guest65046
bvernoux1 has quit [Read error: Connection reset by peer]
<Katharina>
azonenberg: For hierarchical preferences, we need an internal way to reference them
<Katharina>
are you fine with using periods for this? like scope.bla.foo
<azonenberg>
Periods or slashes, no preference. Periods are probably better
<Katharina>
ok, i did periods
<azonenberg>
How's it progressing? any progress screenshots to share etc? :)
<Katharina>
currently reworking the internal representation to support nesting
<azonenberg>
Are you gonna be able to reuse the FilterParameter class or was it too different for that to work?
<Katharina>
Ill look at that as soon as the tree is done :) i can just extend the Preference class to support enums
<azonenberg>
Also, it would be nice if you had the option of tagging fields with a unit like i do for filter parameters. This allows pretty printed display and parsing of SI prefixes for numeric quantities
<Katharina>
that sounds good
<azonenberg>
I have a generic unit "counts" which describes "abstract things" ant not SI units
<_whitenotifier-f>
[scopehal-apps] nshcat assigned issue #221: The preference system should support units - https://git.io/JUjQG
<_whitenotifier-f>
[scopehal-apps] nshcat labeled issue #221: The preference system should support units - https://git.io/JUjQG
<_whitenotifier-f>
[scopehal-apps] nshcat opened issue #221: The preference system should support units - https://git.io/JUjQG
<Katharina>
azonenberg: I hope you dont mind me working on the gui-dev branch directly, and then doing a PR to master when its done?
<Katharina>
its much more ergonomic for me this way
<azonenberg>
no that's the intended workflow
<Katharina>
got it
<azonenberg>
If you have push access to the main repo by all means use it rather than a private branch and PRing
<azonenberg>
For the time being i'm the only one who should be pushing to master without approval except for trivial typo fixes of already-approved code. but anyone with access to the repo is encouraged to do dev in a feature branch there
<Katharina>
yea i think thats important
<_whitenotifier-f>
[scopehal-apps] nshcat opened issue #222: The preferences system should support enum-style prefs - https://git.io/JUj5c
<_whitenotifier-f>
[scopehal-apps] nshcat labeled issue #222: The preferences system should support enum-style prefs - https://git.io/JUj5c
<_whitenotifier-f>
[scopehal-apps] nshcat assigned issue #222: The preferences system should support enum-style prefs - https://git.io/JUj5c
<Katharina>
azonenberg: do you have any premade unit class or something?
<Katharina>
is scopehal::Unit meant for this?
<azonenberg>
Katharina: yes
<Katharina>
k
<azonenberg>
PrettyPrint() and ParseString() are the methods of interest
<Katharina>
I think we also need a UNIT_NONE member, for example a double might be dimensionless, but a string isnt even dimensionless since it isnt scalar
<azonenberg>
So for the moment I use UNIT_COUNTS for that. but you also look at the type field
<azonenberg>
for a value of type string i don't even look at the unit
<azonenberg>
units only apply to numbers
<Katharina>
thats true, i was mostly looking for a unit to "default to" if a property isnt a number, if that makes sense
<azonenberg>
Yeah. That's UNIT_COUNTS
<Katharina>
i guess i can use COUNTS for that, k
<azonenberg>
it basically means "do nothing to it"
<azonenberg>
I might change the name but that's an easy refactoring
<Katharina>
alright :)
<azonenberg>
it was originally just for histogram counts
<azonenberg>
also that was fast. Shapeways shipment with the flex pcb strain relief is already on the way
<azonenberg>
the boards just went to fab today lol
<azonenberg>
Also I just arranged VPN access to a LeCroy WaveSurfer 3034, my old scope from a few years ago, now living in the lab at $dayjob
<azonenberg>
This will allow me to develop a UI for the integrated function generator which most of lecroy's higher end scopes do not have
<Katharina>
will it be part of glscopeclient?
<azonenberg>
Yes. It has to be, becuse the scope only supports one connection to the socket
<azonenberg>
also lovely, i come back from lunch and the scope isnt responding over VICP anymore
<azonenberg>
i can telnet to it (it's running win CE, not full windows)
<azonenberg>
but havent yet found a way to reboot it from the CLI
<Katharina>
do you have shell access?
<azonenberg>
I have shell access via telnet. but it seems to be somewhat sandboxed
<azonenberg>
shutdown.exe didn't work
<azonenberg>
but then again i dont know if that even exists on winCE
<azonenberg>
It looks like i can start the vxi-11 server from telnet
<azonenberg>
i think that should allow me to launch the lecroy remote UI and reload the VICP server
<electronic_eel>
I suggest a remote controlled robot arm that presses the power button ;)
<azonenberg>
lol
<azonenberg>
the funny thing is, i wasnt even using glscopeclient at the time
<azonenberg>
i just left the lecroy remote UI running
<electronic_eel>
hehe
<azonenberg>
i wonder if the siglent firmware is just that unstable? i've never had a real MAUI scope do that
<electronic_eel>
ah, that is one of the scopes done by siglent?
<azonenberg>
yeah
<azonenberg>
wavesurfer 3000/3000z/4000hd are all siglent-lecroy collaborations
<azonenberg>
i dont know the exact details. when ziggggggy was at lecroy he did factory test code for them
<azonenberg>
and the UI is based on MAUI
<azonenberg>
i dont know who did how much on what part
<electronic_eel>
I have a siglent function generator. maybe the scpi for the func gen part is similar
<azonenberg>
is that unstable too?
<azonenberg>
this thing implements the lecroy scpi protocol - imperfectly it seems :p
<electronic_eel>
I have never done any serious work with it over scpi yet
<electronic_eel>
is the func gen in this scope fully arb or is it just fixed function with some modulation?
<azonenberg>
of note: the wavesurfer responds as "VXI11PROTO" rather than the model name over LXI
<azonenberg>
That's a bit unclear :)
<electronic_eel>
heh
<azonenberg>
the UI makes it look like fixed function but the command variables make it seem like it might actually be an arb under the hood
<azonenberg>
Hmmm
<azonenberg>
seems like right now the lxi server is running, but not talking to the scope GUI
<azonenberg>
since that is expecting it in VICP
<azonenberg>
so i'm back to telnet as the sole UI
<electronic_eel>
for many devices a switchable power outlet is enough. but scopes usually have a soft power button these days and miss a autostart function
<azonenberg>
in my case the issue is the remote control server is frozen
<azonenberg>
but it's still running
<electronic_eel>
yeah, but you can't reset it. that is what I usually suggest the remote controlled power outlet for
<azonenberg>
Yeah
<electronic_eel>
but with scopes the problem is getting them to switch back on
<azonenberg>
true
<electronic_eel>
that is where the robot arm comes in
<azonenberg>
i may hve to give up on this for the short term
<miek>
just need an intern in the office :)
Katharina has quit [Read error: Connection reset by peer]
Katharina has joined #scopehal
<azonenberg>
i also cant touch program files or a bunch of other "fun" locations
<azonenberg>
My one hope is that something just hung
<azonenberg>
and the socket will die on its own and it will become available
<azonenberg>
but its probably gonna need a reset
<Katharina>
azonenberg: the property system supports arbitrary nesting, but do you plan on using many levels of nesting?
<Katharina>
this is mostly relevant for GUI design - if you want things nested at least two levels deep (like device.rendering.color etc) then maybe a customized tree view control would be appropiate for the settings gui
<azonenberg>
I'm thinking 2-3 levels of nesting is likely
<azonenberg>
Tree view at left for navigating hierarchy, dialog at right for showing settings within that level?
<Katharina>
oh thats a good compromise
<azonenberg>
that seems to be a pretty standard way of displaying settings
<Katharina>
like having settings like device.rendering, device.connection, device.whatever, and when i click on rendering in the treeview it would list all the preferences in that category
<Katharina_>
also, sorry for all the disconnects my ISP is fucking about with my router remotely
<Katharina_>
its so fucking annoying
<azonenberg>
No worries
<Katharina_>
another design question - gtk tree views can use a "shadow" root node and hide it, so while the tree might be one whole tree, it will look like two separate ones
<Katharina_>
would you like that for the topmost settings categories, like it is in vivado=
<azonenberg>
Yes. I see no reason to allow collapsing the top level of the tree
Katharina has quit [Ping timeout: 260 seconds]
<Katharina_>
alright
<azonenberg>
(15:42:31) azonenberg: Yes. I see no reason to allow collapsing the top level of the tree
<azonenberg>
In case you missed that
<Katharina_>
Saw it, but thanks :) its just my alter-ego that timed out
<azonenberg>
ah ok
<Katharina_>
im going to replace this router soon. its one of these modern ones (new contract) that can be completely remote controlled by the ISP
<Katharina_>
they are constantly updating it (it has a remote update LED that blinks if they do so)
<Katharina_>
im going to replace it with a older, much dumber router