01:17 <Error_404> azonenberg: have a small patch for xptools to add socket rx/tx timeouts, how do you want it?

01:18 <azonenberg> send a PR

01:21 <Error_404> done

01:27 <azonenberg> Error_404: are those linux only or portable?

01:55 <Error_404> oh shoot

01:55 <Error_404> pretty sure linux only

02:04 <azonenberg> I'm ok with you adding them as ifdef'd linux only for the short term i guess, as long as equivalent windows functionality can be added later

02:04 <azonenberg> that way it will still compile on windows if you dont use those functions

02:04 <azonenberg> and glscopeclient is currently linux specific, so a few more portability issues won't be the end of the world. I do want portability eventually though

02:16 <Error_404> https://docs.microsoft.com/en-us/windows/win32/api/winsock/nf-winsock-setsockopt actually looks supported nvm

02:16 <Error_404> fixing types and pushing that

02:18 <Error_404> done cc azonenberg

02:36 <azonenberg> Perfect

02:36 <azonenberg> Will look and merge as soon as i'm done soldering this probe

02:38 <azonenberg> also oooh i sized the pcb-to-shell fit PERFECTLY

02:38 <azonenberg> the little tail around the SMA basically friction-fits into the probe shell

02:38 <azonenberg> i might not even have to glue it

02:58 <_whitenotifier-3> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/Jvzf2

02:58 <_whitenotifier-3> [scopehal] azonenberg 7d9840b - Updated to latest xptools

02:58 <azonenberg> Merged

03:21 <azonenberg> also new speed record for scopehal

03:22 <azonenberg> 52 FPS with 4k points on a single channel from a hdo9204

03:22 <azonenberg> mind you that is 52 frames with completely new waveform data each frame

03:22 <azonenberg> so 52 WFM/s

03:23 <azonenberg> can't wait till we build hardware that can push 100k points waveforms or more at that speed

03:47 <azonenberg> https://twitter.com/azonenberg/status/1232871552959430656

05:51 <Error_404> azonenberg: pull request of shame submitted :p

05:51 <Error_404> thanks C, for providing type information about enums /s

05:52 <Error_404> also sigh, need to strip back and reimplement a lot of the virtualized lecroy classes, 'cause retrying reads isn't sufficient, we need to actually re-send the command

05:53 <Error_404> impossible to do w/o refactoring all instances of sendcommand unless I pivot to something like an optional expect/stored result

07:11 <azonenberg> hmmm

07:11 <azonenberg> great

07:12 <azonenberg> maybe might make sense to just make it be a separate driver rather than a subclass in the end? or what

07:13 <_whitenotifier-3> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JvzI2

07:13 <_whitenotifier-3> [scopehal] azonenberg aef5ecf - Updated to latest xptools

07:20 <azonenberg> https://www.antikernel.net/temp/probe-risetime.png - my probe

07:22 <azonenberg> https://www.antikernel.net/temp/ta061-risetime.png vs Pico TA061 (commercial 1.5 GHz transmission line probe, $329 MSRP)

07:23 <azonenberg> The Pico claims 2 pF input capacitance, i think mine is a fair bit less but don't have instruments sensitive enough to measure single digit pF accurately

07:42 <Error_404> yes that's what I am thinking

07:42 <Error_404> well

07:43 <Error_404> to be fair, it shouldn't hurt anything if used for a lecroy target

07:43 <Error_404> but I will probably just kill the inheritance to LecroyOscilloscope to keep things reasonable

07:43 <Error_404> just will take some tinkering

07:43 <Error_404> very nice

07:43 <Error_404> multech?

07:44 <azonenberg> my probe? yeah. On Rogers RO4350B

07:51 <monochroma> Error_404: siglent's network interface code: https://i.imgur.com/A8qOQ3L.png

07:52 <Error_404> t.t

10:56 <Error_404> so, basically ended up implemented a layer of one-way commands and commands with an expected resp. could probably re-inherit the lecroy driver if it used a similar interface

10:56 <Error_404> and I likewise think it'll be a better interface going forward, but I'm not gonna try and convert everything to it

10:57 <Error_404> has resulted in much cleaner code for the siglent impl so far

10:57 <monochroma> we probably should have some better communication mechanisms, i don't think it's acceptable for the whole UI to lock up because of a communications fault

10:58 <monochroma> communication and UI seperation i mean

10:58 <azonenberg> monochroma: yeah right now the data acquisition thread is pretty separated

10:59 <azonenberg> but a few of the context menu functions etc do call methods on the scope that could potentially block

10:59 <monochroma> okay so we just need some timeouts i suppose

11:00 <azonenberg> yes agreed. We should try to avoid ever having unbounded blocking

11:09 <Error_404> I think there are some timed events too? basically anything from gtkmm that fights for the scope mutex will result in a lock if comms get blocked

11:09 <Error_404> would be nice to have a way to indicate up the stack that that has occured

11:09 <Error_404> but for now drawing more coherant boxes around communication layers is a start

11:10 <Error_404> will pay dividends when abstracting away transport

11:10 <azonenberg> Yes. One of the challenges i had early on was not even knowing what assumptions were valid

11:10 <azonenberg> because i hadn't looked at enough hardware yet

11:10 <Error_404> oh for sure

11:10 <Error_404> I think another thing is there is just necessarily a large bulk of code that isn't...very exciting. makes sense you'd want to get something working

11:11 <azonenberg> remember, scopeclient (not glscopeclient, the original GTK-only app) was originally a viewer for RED TIN

11:11 <Error_404> hehe

11:11 <azonenberg> and eventually i added support for my old ds1102d

11:11 <azonenberg> then dropped it because it was such a pain in the butt to get waveforms off of

11:11 <azonenberg> :p

11:11 <azonenberg> Some of this code goes that far back

11:12 <azonenberg> hence the 2012-2020 copyright header :p

11:13 <azonenberg> almost everything has been rewritten, and some of the original subversion history probably got squashed when i migrated to git circa 2015

11:13 <azonenberg> But that is the original legacy of it

11:13 <Error_404> hah, had no idea it talked to the ds1102