* Revert "Fix modeld dropping frames if processing takes over 50 ms"
This reverts commit fb908877a50d72e20a047bd4138316b12c8a30cc.
* track frame drop
* fix unused
* alert on frame drop perc
* reduce thresholds
* posenet is invalid on non-consecutive frames
* Refactor filter time constant
* Use vipc_dropped_frames for posenet valid check
* Actually set frame drop percentage
* Add explicit cast just to be sure
* Prevent frames dropped to go up to quickly on startup
* bump cereal
* reduce soft disable threshold to 1%
Co-authored-by: Willem Melching <willem.melching@gmail.com>
old-commit-hash: 58c3b5ba84f6c3da3c660770148cea98c708fce5
* start building release2 in CI
* build script
* bring over changes from test branch
* fix build release test
* remove references to test branches
* ucnomment push
* fix duplicate remove
* two commits for dashcam
* need the key
old-commit-hash: c7156616568ce2fa780db3eafc5f912ab2ead14e
* separate scons cache by branch
* debug
* guess you can only call CacheDir once
* remove another call
* pass git branch and commit through
* copytree
* set CI
old-commit-hash: d8ff1865450fe341f05dc6d8b6a8580eccb24f86
* WIP sound test
* it does something
* refactor
* phone only
* update release files
* test sound card init
* add to CI
* check writes
* increase time
* unused
* only build cereal
* small tolerance
Co-authored-by: Comma Device <device@comma.ai>
old-commit-hash: f1afb3e3ae4b95a1de0262d1408c49093c5ada6f
* let's see if this works
* fix build_release actions job
* does jenkins like this config
* separate jenkinsfile for release build
* fix devel build
* devel build should work
* always pass that for now
* run modeld replay
* release2 build will be a separate PR
* pass env to phone shell
* force checkout
* run on real jenkins eons
* add timeout
* rsync
* more timeout
* trailing slash
* fix branch detection
* debug
* not sure why paramiko doesn't pass it through
* newline
* CI_PUSH
* still not passing it
* test branch
* should be good now
old-commit-hash: 32f03ec8a52c7c5fa56d5cce19fbbd7ecf24b7dd
* pull CPU usage test out of manager
* remove that
* add to release files
* cleanup
* executable
* this should work
* check if car started procs are running
* debug
* add min cpu usage
* remove debug prints
* adjust min cpu threshold
Co-authored-by: Comma Device <device@comma.ai>
old-commit-hash: 6db7fa8c033841c82d1139119aad8a20645449f4
* Add Hyundai BSM signals
to take advantage of BSM to block lane change when BSM active.
* bug, added LCA11 to canparser
previous check failed, fixed by adding LCA11 to canparser
* update refs
Co-authored-by: Adeeb Shihadeh <8762862+adeebshihadeh@users.noreply.github.com>
old-commit-hash: 57f29968a2662bb49807f15165fd69254da0cc93
* Add initial Forester 2019 support
* Add dgranger15 Forester 2019 FPv1
* fingerprint syntax fix
* FPv1 syntax fix 2
* Add Forester 2019 to CAR
* fix indent
* Add Forester 2019 to test_car_models ignore list
* Update Forester 2019 steering tuning (dgranger15)
* Add Forester 2019 to README.md
* Change lists to single value
* Add test route for Forester 2019
* Remove model year
* Remove Forester model year in test_car_models
* Remove usernames, add Forester ACC speed mph unit value
old-commit-hash: 09e7d4ab3a044fd5279cce43853f69465eac876f
* locationd at 20hz
* update ref
* bump cereal
* dont modify global state
* add scons files
* ecef2geodetic and geodetic2ecef
* Finish local coords class
* Add header file
* Add orientation.cc
* cleanup
* Add functions to header file
* Add cython wrapper
* y u no work?
* This passes the tests
* test rot2quat and quat2rot
* Teste euler2rot and rot2euler
* rot_matrix
* test ecef_euler_from_ned and ned_euler_from_ecef
* add benchmark
* Add test
* Consistent newlines
* no more radians supported in geodetic
* test localcoord single
* test localcoord single
* all tests pass
* Unused import
* Add alternate namings
* Add source for formulas
* no explicit tests needed
* remove benchmark
* Add release files
* Typo
* Remove print statement
* no access to raw transform matrix
* temporarily add tolerance
* handcode quat2euler
* update ref
old-commit-hash: c18e7da3c25c59a19efda8ec3a1b716c2571cc2f
* kind of works
* move that
* hack to get camerad to reliably terminate
* not sure why SIGTERM wasn't working before
* compare bytes
* clean up some hacks
* gitignore
* fix that
* WIP
* no reboot
* comparison works
* pretty print
* fix build
* run in jenkins
* python path
* space
* raise timeout
* new eon
* skip the copy
* spinner
* spin less
* update model ref commit
* reenable that
* clean up
* fix jenkinsfile
* parallel
* wrap it in a stage
* fix linter
* better progress
* lower timeout
Co-authored-by: Comma Device <device@comma.ai>
Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com>
old-commit-hash: 63c8e8439b6d19020055f922340d8e33d906363a
