<d1b2>
src/ngscopeclient/CMakeFiles/ngprotoshaders.dir/all] Error 2 is this a known issue? Ubuntu 20.04 I have to sleep but will check it out myself later if nobody knows
<azonenberg>
@johnsel i thought that had been fixed. there was an issue some time ago that was a race condition when parallel building glscopeclient and ngscopeclient, one side would try to build the libscopeprotocols shaders but the other would copy them before they had been compiled
<azonenberg>
I do not know if this is the same root cause but it presented similarly
<d1b2>
<johnsel> Thanks @azonenberg that is helpful, I have to run some errands but it did indeed seem to behave differently when running single threaded. I didn´t have the patience to let it finish but I can let it do that while I am gone and investigate later.
<azonenberg>
(removing glscoepclient and having pure NG will also potentially fix some of this as we won't be buidling a bunch of stuff twice/redundantly)
esden has quit [Server closed connection]
esden has joined #scopehal
<d1b2>
<tiltmesenpai> oh lol that did actually set the webhooks off, I assumed it would T.T
_florent_ has quit [Server closed connection]
_florent_ has joined #scopehal
bvernoux has joined #scopehal
<d1b2>
<johnsel> @azonenberg doesn´t seem to be it, I'm upgrading cmake maybe it's something weird because it is an older version on 20.04 by default
<d1b2>
<johnsel> looks like that may have resolved the issue
<d1b2>
<johnsel> we do have to bump up the vulkan version to 250.1 instead of 250.0, it is no longer available for download
<d1b2>
<johnsel> I'll look at the patch notes in a bit to see if there is a reason why that shouldn't be done but I think I can make a PR for it
<azonenberg>
Also.... General forward looking thought
<azonenberg>
I think the idea of a single waveform associated with a given instrument/channel might be holding us back, since it seems to be creating some serialization points in which e.g. we can't render one waveform while doing other processing on a different one
<azonenberg>
it would be a lot of work to change this to a more pipelined model, and i dont yet know if it's a good idea or how to do it