azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | https://github.com/azonenberg/scopehal-apps | Logs: https://libera.irclog.whitequark.org/scopehal
someone--else has quit [Quit: Connection closed]
<_whitenotifier-7> [scopehal] azonenberg pushed 4 commits to master [+0/-0/±7] https://git.io/JGRdj
<_whitenotifier-7> [scopehal] azonenberg 54dd420 - Oscilloscope: implemented FMA version of Convert16BitSamples()
<_whitenotifier-7> [scopehal] azonenberg fe8e933 - Unrolled FMA implementation to 64 samples
<_whitenotifier-7> [scopehal] azonenberg 1a827a1 - Interleaved instructions to improve ILP
<_whitenotifier-7> [scopehal] azonenberg 1e713f8 - Revert "Interleaved instructions to improve ILP" This reverts commit 1a827a143f961a953e25f9ef91c71d071ca1e385.
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 264 seconds]
Degi_ is now known as Degi
<_whitenotifier-7> [scopehal] azonenberg opened issue #490: Save/restore gain/offset config for filters - https://git.io/JG0qt
<_whitenotifier-7> [scopehal] azonenberg labeled issue #490: Save/restore gain/offset config for filters - https://git.io/JG0qt
<_whitenotifier-7> [scopehal] azonenberg assigned issue #490: Save/restore gain/offset config for filters - https://git.io/JG0qt
<_whitenotifier-7> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JG0m4
<_whitenotifier-7> [scopehal] azonenberg 66f89f3 - Filter: save/load range and offset if applicable. Fixes #490.
<_whitenotifier-7> [scopehal] azonenberg closed issue #490: Save/restore gain/offset config for filters - https://git.io/JG0qt
<_whitenotifier-7> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±4] https://git.io/JG0aO
<_whitenotifier-7> [scopehal-apps] azonenberg b633ea7 - Statistic configuration is now saved in scopesession files. Fixes #87.
<_whitenotifier-7> [scopehal-apps] azonenberg closed issue #87: Save/load statistic config - https://git.io/JfWkE
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg commented on issue #12: Allow triggering on MSO channels - https://git.io/JG0yi
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg closed issue #12: Allow triggering on MSO channels - https://git.io/JsR7H
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg pushed 1 commit to master [+0/-0/±3] https://git.io/JG07D
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg b6190f9 - Catch SIGPIPE and don't abort if we see it. Also catch "no channels enabled" error during shutdown gracefully.
<_whitenotifier-7> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/azonenberg/scopehal/compare/66f89f38dfc0...5d43c2e18556
<_whitenotifier-7> [scopehal] azonenberg 5d43c2e - PicoOscilloscope: clear any pending waveforms in the queue when changing trigger configuration
<azonenberg> Ok so at this point i still have to get the bandwidth limiter working
<azonenberg> (nothing wrong with it, just never tried)
<azonenberg> and the signal generator
<GyrosGeier> dang`r`us, exactly, cl_khr_gl_sharing is the extension
<GyrosGeier> 22:03 < ericonr> intel has their own CPU opencl impl, are there any others?
<GyrosGeier> amd also has one
<GyrosGeier> this is basically a good idea: a cheap API to a compiler that is optimized for the exact CPU model that is present, with no system knowledge required
<_whitenotifier-7> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±4] https://git.io/JG0js
<_whitenotifier-7> [scopehal-apps] azonenberg e2923cd - Fixed potential use-after-free in right click menu
<dang`r`us> GyrosGeier, I suppose that like in GL, living 100% without extensions is going to be hard
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JGEtE
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg 9a9daea - Fixed memory depth changes not applying sometimes
<GyrosGeier> yes
<GyrosGeier> but AFAIK every CL implementation supports GL sharing
<GyrosGeier> although I'm unsure if it's enabled for Beignet
<dang`r`us> GyrosGeier, any idea about the windows linker error my PRs (apparently) caused btw? I think it has to do with yaml being statically linked *only on windows* but I couldn't see anything fishy in the cmakelists, *especially* not with scopeprotocols where it seems to fall over
<dang`r`us> regarding beignet, a grep for cl_khr_gl_sharing in their repo indicates that it is conditionally enabled
<dang`r`us> that document renders a bit poorly. Oh well, *strictly* speaking it's still markdown, eh
_whitelogger has joined #scopehal
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JGEsG
<_whitenotifier-7> [scopehal-pico-bridge] azonenberg 9784cc7 - Allow bridge to operate in MSO-only mode with no analog channels enabled
<GyrosGeier> libscopehal re-exports the libyaml-cpp symbols because it is linked against a static library, and for some reason these are marked as dllexport
<GyrosGeier> that means libyaml-cpp was built wrong
<GyrosGeier> if you build a library on Windows that is supposed to be used as a static library, symbols should not be marked for export
<GyrosGeier> the other thing is that the include_directories(${yamlcpp_INCLUDES}); target_link_libraries(${yamlcpp_LIBRARIES}) approach is not how it works in modern CMake
<GyrosGeier> normally there'd be a magic project name, and linking against it pulls in both the headers and the libraries
<azonenberg> libscopehal should not be linking statically to yaml-cpp either
<azonenberg> On linux that's dyanmic
<azonenberg> dynamic*
<GyrosGeier> exactly
_whitelogger has joined #scopehal
bgamari has quit [Quit: ZNC 1.8.1 - https://znc.in]
bgamari has joined #scopehal
<dang`r`us> great
<dang`r`us> so I suppose the change I made here fixed it for macos but broke it for windows https://github.com/azonenberg/scopehal/pull/489/files#diff-0c1fc4781215d8612038a1668d272725fc7f805904755eeeae35a22f2ed5af9cL107-R114
<dang`r`us> FWIW: on macos I have a homebrew/lib/cmake/yaml-cpp directory, which is probably preferable over the FindYaml.cmake included with scopehal-apps? (I'm still a total cmake noob)
<azonenberg> That I'm not sure of. I've never done anything too advanced build system wise either, this project is the most complicated i've done to date in terms of build chains and multiple target platforms
<azonenberg> Bird|ghosted: any thoughts?
someone--else has joined #scopehal
<xzcvczx> azonenberg: i assume you have both the 3k and the 6k sdks installed?
<xzcvczx> (pico)
<noopwafel> on linux the pico gui pulls in everything
<xzcvczx> i am thinking i should start by cleaning it up so it will only use the ones it finds
<xzcvczx> with a way to query to see which ones were builtin
<noopwafel> why not just require them all?
<xzcvczx> why require them all? if you only have 1 or 2 models of scope it seems rather excessive to require everything to build
<noopwafel> I guess I don't know how it works on other OSes :)
<xzcvczx> well i am on linux just not on debian so i have to install the debs one by one
<xzcvczx> and dpkg is vommiting at one of the debs which doesn't help
<noopwafel> the picoipp one?
* xzcvczx gives noopwafel a cookie
<noopwafel> I should probably have reported that as a bug
<Bird|ghosted> azonenberg, I'm with Gyros on this one, static libyaml-cpp shouldn't be dllexporting its symbols
<Bird|ghosted> it sounds like the static/dynamic switch doesn't turn off/on the dllexport machinery
<dang`r`us> >_>
<xzcvczx> noopwafel: how did you workaround it?
<noopwafel> I rebuild the deb without the broken line
<noopwafel> I didn't keep the deb apparently, but I can do it again
<xzcvczx> nah nvm, i just hoped their might be an easier way
<noopwafel> you can just install an older deb also
<noopwafel> maybe. hm.
<noopwafel> I can't make dpkg complain at all with the debs in their repo, wha.
someone--else has quit [Quit: Connection closed]
<noopwafel> Oh, wrong machine. Nope, still broken.
<xzcvczx> lol
<noopwafel> sent a mail to support@ just in case they're not aware
<xzcvczx> not that only installing a single sdk would have stopped this from vommiting btw
<noopwafel> yeah indeed
* xzcvczx sorta wishes the sdk was a bit more of a hal
<GyrosGeier> # Required for Qsys to work
<GyrosGeier> quartus_sh --tcl_eval set_user_option -name HDL_PREFERENCE VHDL
* GyrosGeier lights things on fire
<dang`r`us> GyrosGeier, so that was the root cause of your docker issue?
<GyrosGeier> the root cause is that this software is broken, but at least I can now work around it
someone--else has joined #scopehal
<xzcvczx> azonenberg: ok i got about 90% of the way to implementing 5000 support, but i think i am going to start over with a large refactor of picobridge with putting each model into its own class
<xzcvczx> the code was just getting a bit too ugly for my tastes
_whitelogger has joined #scopehal
<GenTooMan> xzcvczx, trying to make your code more classy I see... pun intensive.
<monochroma> ow.
<xzcvczx> GenTooMan: you are banished to the naughty step for that one
esden has quit [*.net *.split]
_whitenotifier-7 has quit [*.net *.split]
sorear has quit [*.net *.split]
dang`r`us has quit [*.net *.split]
esden has joined #scopehal
sorear has joined #scopehal
dang`r`us has joined #scopehal
<Kliment> xzcvczx: the naughty STEP
<GenTooMan> well I would rather have the nice step but such is life.
dang`r`us has quit [Quit: WeeChat 3.1]
dang`r`us has joined #scopehal
<azonenberg> There is no nice STEP
<azonenberg> AP203 and AP214 are equally painful
someone--else has quit [Quit: Connection closed]