azonenberg changed the topic of #scopehal to: ngscopeclient, libscopehal, and libscopeprotocols development and testing | https://github.com/ngscopeclient/scopehal-apps | Logs: https://libera.irclog.whitequark.org/scopehal
Degi has quit [Ping timeout: 246 seconds]
Degi has joined #scopehal
preamp-org has joined #scopehal
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±371] https://github.com/ngscopeclient/scopehal/compare/f51f4be6eedb...96fb51f92bb8
<_whitenotifier-e> [scopehal] azonenberg 96fb51f - Removed all usage of deprecated GetInputLocation(), as well as non-Vulkan FlowGraphNode::Refresh(). Fixes 1054.
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/3772bd082364...76a3b6c56783
<_whitenotifier-e> [scopehal-apps] azonenberg 76a3b6c - Updated to latest scopehal
preamp-org has quit [Ping timeout: 252 seconds]
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal/compare/96fb51f92bb8...07bbe206a426
<_whitenotifier-e> [scopehal] azonenberg 07bbe20 - Added missing override specifier
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/76a3b6c56783...e280fe4cb4b3
<_whitenotifier-e> [scopehal-apps] azonenberg e280fe4 - Updated to latest scopehal
preamp-org has joined #scopehal
preamp-org has quit [Read error: Connection reset by peer]
preamp-org has joined #scopehal
preamp-org has quit [Read error: Connection reset by peer]
preamp-org has joined #scopehal
pirate has quit [Ping timeout: 248 seconds]
pirate has joined #scopehal
pirate has quit [Remote host closed the connection]
pirate has joined #scopehal
pirate has quit [Ping timeout: 248 seconds]
pirate has joined #scopehal
pirate has quit [Ping timeout: 262 seconds]
pirate has joined #scopehal
pirate has quit [Ping timeout: 245 seconds]
pirate has joined #scopehal
pirate has quit [Ping timeout: 256 seconds]
pirate has joined #scopehal
pirate has quit [Ping timeout: 246 seconds]
pirate has joined #scopehal
pirate has quit [Ping timeout: 246 seconds]
preamp-org has quit [Ping timeout: 264 seconds]
preamp-org has joined #scopehal
pirate has joined #scopehal
pirate has quit [Ping timeout: 248 seconds]
preamp-org has quit [Remote host closed the connection]
pirate has joined #scopehal
ablu has quit [Ping timeout: 264 seconds]
ablu has joined #scopehal
gruetze_ is now known as gruetzkopf
d1b2 has quit [Remote host closed the connection]
d1b2 has joined #scopehal
d1b2 has quit [Remote host closed the connection]
d1b2 has joined #scopehal
jj5 has quit [Remote host closed the connection]
<_whitenotifier-e> [scopehal] blazra commented on issue #334: CAN FD protocol decode - https://github.com/ngscopeclient/scopehal/issues/334#issuecomment-4614181327
<_whitenotifier-e> [scopehal] azonenberg closed issue #1054: Deprecate GetInputLocation() - https://github.com/ngscopeclient/scopehal/issues/1054
<_whitenotifier-e> [scopehal] azonenberg commented on issue #1054: Deprecate GetInputLocation() - https://github.com/ngscopeclient/scopehal/issues/1054#issuecomment-4614606224
<_whitenotifier-e> [scopehal] azonenberg commented on pull request #1017: WIP: Rigol oscilloscope refactor - https://github.com/ngscopeclient/scopehal/pull/1017#issuecomment-4614617302
<_whitenotifier-e> [scopehal] azonenberg commented on issue #965: Somewhat random timebase control / detection & returned sample order with Rigol MSO2302A - https://github.com/ngscopeclient/scopehal/issues/965#issuecomment-4614624947
<_whitenotifier-e> [scopehal-apps] azonenberg commented on issue #997: Newly added measurement filters don't auto refresh - https://github.com/ngscopeclient/scopehal-apps/issues/997#issuecomment-4614654972
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±3] https://github.com/ngscopeclient/scopehal-apps/compare/e280fe4cb4b3...0a06a81df9ae
<_whitenotifier-e> [scopehal-apps] azonenberg 0a06a81 - Mark newly created filters as dirty even if not adding to a plot, so measurements get correctly updated too. Fixes #997.
<_whitenotifier-e> [scopehal-apps] azonenberg closed issue #997: Newly added measurement filters don't auto refresh - https://github.com/ngscopeclient/scopehal-apps/issues/997
preamp-org has joined #scopehal
<d1b2> <helaslo> What is the cutoff date for 0.2? I would like to finish my tarball PR by then
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±5] https://github.com/ngscopeclient/scopehal/compare/07bbe206a426...93b1fb53edf4
<_whitenotifier-e> [scopehal] azonenberg 93b1fb5 - Fixed a bunch of casts and format string issues found by static analysis
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/0a06a81df9ae...c7eb0c77ab68
<_whitenotifier-e> [scopehal-apps] azonenberg c7eb0c7 - Updated to latest scopehal
<_whitenotifier-e> [scopehal] azonenberg pushed 3 commits to master [+0/-0/±4] https://github.com/ngscopeclient/scopehal/compare/93b1fb53edf4...3e862cd9a22b
<_whitenotifier-e> [scopehal] azonenberg 06e95d5 - Updated to latest xptools
<_whitenotifier-e> [scopehal] azonenberg 5ec15ad - Added some checks for empty/overflowing file sizes
<_whitenotifier-e> [scopehal] azonenberg 3e862cd - PulseWidthTrigger: avoid naming collision
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/c7eb0c77ab68...675032a9e888
<_whitenotifier-e> [scopehal-apps] azonenberg 675032a - Updated to latest scopehal
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±5] https://github.com/ngscopeclient/scopehal/compare/3e862cd9a22b...a35d4cf94138
<_whitenotifier-e> [scopehal] azonenberg a35d4cf - Static analysis fixes, missing initializations, suppressions for enumeration functions
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/675032a9e888...1b6f1f697ada
<_whitenotifier-e> [scopehal-apps] azonenberg 1b6f1f6 - Updated to latest scopehal
<d1b2> <helaslo> @azonenberg Sorry for the ping, just don't know if you are seeing this 😄
<d1b2> <azonenberg> We don't have a formal cutoff but thunderscope dev units are going through calibration and test now
<d1b2> <azonenberg> so likely "not too much longer"
<d1b2> <helaslo> Also what is the workflow for the new CI? If I understand correctly just an sh script and copy to the artifacts folder?
<d1b2> <azonenberg> Yeah the actual build executes test-scripts/test-driver-whatever.sh
<d1b2> <azonenberg> anything you dump in artifacts/ gets copied off and uploaded to the artifact server
<d1b2> <helaslo> What determines the OS that is runs on?
<d1b2> <azonenberg> this is the script that launches all of the build jobs
<d1b2> <helaslo> Ohh sorry didn't see this repo
<d1b2> <azonenberg> https://github.com/ngscopeclient/scopehal-ci-scripts/blob/main/api/github-hook.php this is the entry point github's webhook invokes
<d1b2> <azonenberg> it parses out the few fields we care about from the json then invokes batch-launcher to kick off the build jobs
<d1b2> <azonenberg> right now it's unidirectional, we do not report status back to github
<d1b2> <azonenberg> so there's no pretty green check on the repo anymore, you have to open the dashboard to see results
<d1b2> <azonenberg> i may eventually fix that but if we are thinking of migrating off github to codeberg or something it might make sense to delay until then
<d1b2> <azonenberg> the migration to the new CI had to happen regardless to enable gpu/scope in the loop testing
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±2] https://github.com/ngscopeclient/scopehal/compare/a35d4cf94138...b3c049dc5baf
<_whitenotifier-e> [scopehal] azonenberg b3c049d - DigilentOscilloscope: don't allocate memory if we arent going to use it
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/1b6f1f697ada...b92b0a101d4a
<_whitenotifier-e> [scopehal-apps] azonenberg b92b0a1 - Updated to latest scopehal
<d1b2> <helaslo> Also if You will eventually tag the release will you be please be able to sign the tarball with GPG, and include the public key on the project page? Suse seems to prefer this: https://en.opensuse.org/openSUSE:Package_source_verification Also I think Fedora prefers a similar approach
<d1b2> <azonenberg> Yeah release tags will go through an extra manual process to sign mac/windows binaries anyway
<d1b2> <azonenberg> So we can do the same for the linux ones
<d1b2> <azonenberg> i'll have to make a new gpg signing key for that purpose, i think the one i used to use for mail crypto is long expired and i want to use a separate one for this anyway
<d1b2> <helaslo> Thanks, in the summer I plan to get ngscopeclient to Factory, I looked at the process in the past The nice thing about that is that we get daily CI runs around every external dependency update, so it will scream as soon as any dependency breaks the compilation/unit tests 😄
<d1b2> <azonenberg> i mean running every ci build on arch is pretty good at that too :p
<d1b2> <azonenberg> thats one of the reasons we have arch as a ci runner it's the canary in the coal mine for such problems
<d1b2> <helaslo> I kinda like SUSE's approach to CI Test everything For the main OS they even "click" through the whole install process (and everything else) step by step and do some OCR to check for errors See: https://openqa.opensuse.org/tests/5993035 Despite being a rolling distro, it breaks only subtly and very-very rarely, and if that happens there is rollback
<d1b2> <azonenberg> I want to eventually have the ngscopeclient gui scripted for testing
<d1b2> <azonenberg> but we are a long ways from that being a thing
<d1b2> <azonenberg> we dont even have 10% test coverage on the filters
<d1b2> <azonenberg> and ~0% on drivers
<d1b2> <helaslo> Maybe that can go together with the decoupling of scopehal?
<d1b2> <helaslo> What is the plan for that? hardware-in-the-loop or mock devices with test cases?
<d1b2> <azonenberg> Hardware in loop for thunderscope, ignore everything else to start
<d1b2> <azonenberg> thunderscope we can easily get dev units and they dont take up much space or use much power
<d1b2> <azonenberg> so it makes sense
<d1b2> <azonenberg> other stuff, we may eventually want to make software mocks for but not sure yet
<d1b2> <helaslo> Sounds good
<d1b2> <azonenberg> but there is so much other refactoring and architecture cleanup needed before we think about that
<d1b2> <helaslo> Wow the builds are a bit faster on the new CI (looking at the latest windows one with its 11 minutes 😄 )
<d1b2> <azonenberg> and this is with no caching too
<d1b2> <azonenberg> each build is from a clean image
<d1b2> <azonenberg> (although i did bake some of the dependencies into the images so there's less downloads)
<d1b2> <azonenberg> but they do an apt install/whatever at the start to grab every dep
<d1b2> <helaslo> Maybe consider setting up a local cache for the linux repos? Just to minimize that part of the scripts I think most mirrors are http for this exact reason Can save a few minutes
<d1b2> <azonenberg> the builders are also much beefier than github's, we have 8 vCPUs and 32GB RAM per runner (except the arm64 ones which have only 8GB due to capacity limits on the mac mini) and allow up to five to run simultaneously on the x86 box plus one on the mac
<d1b2> <azonenberg> I have a local apt-cacher-ng instance I use for all of my personal boxen/VMs plus i believe the debian runners
<d1b2> <azonenberg> i attempted to set it up for the ubuntu runners and had technical difficulties because of something weird canonical was doing and didnt have time to troubleshoot
<d1b2> <azonenberg> the non-deb based distros and msys are currently running uncached
<d1b2> <azonenberg> fixing that is a longer term todo
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±6] https://github.com/ngscopeclient/scopehal/compare/b3c049dc5baf...6f007d6f0623
<_whitenotifier-e> [scopehal] azonenberg 6f007d6 - More static analysis fixes, possible integer overflow, some bad casts
<_whitenotifier-e> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal/compare/6f007d6f0623...2254902d383f
<_whitenotifier-e> [scopehal] azonenberg 2254902 - Fixed typo
<_whitenotifier-e> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/b92b0a101d4a...d218cccc79b4
<_whitenotifier-e> [scopehal-apps] azonenberg d218ccc - Updated to latest scopehal
<d1b2> <azonenberg> ok i think that is basically all of the static analysis fixes that are not "we left X incomplete and didn't suppress the warning or mark the argument as unused because we want to be reminded to come back and finish the job"
<d1b2> <azonenberg> (54 open cppcheck + clang-analyzer findings down from multiple thousands)
jmz1993 has joined #scopehal
jmz1993 has left #scopehal [Leaving]
<_whitenotifier-e> [scopehal-apps] azonenberg commented on issue #523: ngscopeclient: use VK_EXT_memory_budget if available - https://github.com/ngscopeclient/scopehal-apps/issues/523#issuecomment-4617073755
preamp-org has quit [Remote host closed the connection]
azonenberg has quit [Read error: Connection reset by peer]
azonenberg has joined #scopehal
jj5 has joined #scopehal