* not sure why this is a test
* Revert "not sure why this is a test"
This reverts commit c13e16311e.
* no ruff for this pattern :(
* this is always caught
* should fail
* simpler
* Revert "should fail"
This reverts commit 21bd15275c.
* fix
* cleanup this
* better comment
* better
* this can be more localized
* better names
* better names
* should fail
* Revert "should fail"
This reverts commit c58495e8b3.
---------
Co-authored-by: Justin Newberry <justin@comma.ai>
old-commit-hash: 7f398e1cf9
* fails if no skip function
* Revert "fails if no skip function"
This reverts commit ff3d697d38.
* Revert "test_models: fallback to public route when route isn't uploaded to CI bucket (#30794)"
This reverts commit f60e70d4729efe7db36888a0a750a0e4706cc943.
old-commit-hash: 8c4930d2d3
* pre-hypothesis
* some hypothesis junk
* this kinda works but is really slow due to counter check
* choose addrs from fingerprint
* stash
* honda nidec brake pressed mismatches fixed
* bump panda
* stash
* tesla: use DI_torque2 (panda msg)
* run
* run
* ah this honda mismatch too
* no more multi can msgs
* clean up, remove old file
* add todo
* prob can remove urandom
* stash, huge examples
* fix pq standstill mismatch
* yuge
* yup there's a leak somewhere
* try to find leak
* skip dashcam (pq and tesla)
* PR comments
* bump
* draft stash
* fix alt brake hondas
* bump
* bump
* bump
* some clean up
* minor clean up
* more clean up
* stash
* fix honda bug
* more
* 100 examples
* revert tesla
* no memory leak any more?
* bring back tests with skips
* parameterize max_examples
* skip interceptor
* is jenkins on my branch?
* ooh that's fast
* 50 is not bad for GH CI
* 300 might be better with rest of test_models
* no more detection
* bump
* need CS_prev to catch bugs where openpilot changes and panda doesn't (eg. not setting interceptor safety mode)
* need to simplify all this
* need a warm up first, since some signals are 1 by default (toyota's gas_released!=1)
* changes
* set honda safety param
* set toyota safety param
* bump panda
* clean up honda
* rm interceptor
* thought interleaving addrs might help, but we can fine tune later
* Revert "thought interleaving addrs might help, but we can fine tune later"
This reverts commit 153301384b.
* get size from dict
* what
* add nocapture marker
* clean up
* try to raise logging level
* need to run last as pytest_runtest_call, since it starts capturing
* get capman conditionally
* mark
* type fingerprint
* should use gen_empty_fingerprint
* no longer needed
* draft
* no longer need gc
* clean that up
* test everything!
* more clean up
* more
* no point
* fix that
* fix errors
* bump
* nice even 300 examples for 300 segs
* final bump :fingers_crossed:
* better import order
* remove debugging prints
* warm up kinda works
* Revert "warm up kinda works"
This reverts commit 7fc77b07d5.
* random seed
* revert
* strat
strat
* add expl comment
* cmt
* check controls allowed
* Revert "check controls allowed"
This reverts commit e82a0e5396.
* not unittests
* run tests!
* run tests 2!
* run tests 3!
* seed unused
* revert
* add shrink phase, and remove health check suppression
* hello
* oncemore
* Update selfdrive/car/tests/test_models.py
old-commit-hash: 5052b55c44
* sanity check on velocity
* import from there
* use raw
* just check last
* ignore body
* reduce tolerance and bump panda
* its not average anymore
* slightly higher tolerance
old-commit-hash: 1ca1cf3922
* randomize internal seg list segments
* random
* pytest-randomly sets random.seed to a consistent value for all the workers/processes
* noeol
* update
* Revert "update"
This reverts commit aff9a69c4e.
* lock
* don't randomize by default
* remove random-order
* strict
* random
* one fix
* test
* does nothing
* rm tests
* Revert "rm tests"
This reverts commit b548e3fcd4.
* (can't repro locally) just athena should be fine
* bs1
* bs2
* bs3
* bs4
* bs5
* wrong way
* no controls
* no car
* no board
* controls?
* crazy -common
* Revert "crazy -common"
This reverts commit 02365d712b.
* test athena
* test athena 2
* test athena 3
* test athena 4
* test athena 5
* test athena 6
* test athena 7
* test athena 8
* test athena 9
* ??
* in one commit
* common?
* car and board
* -controls -board
* random-order
* no board
* revert
* car/tests
* least likely
* try
* try 2
* draft
* draft
* so much better
* cmt
* use randomly
* not needed here
* directly modify option works
* bb
* test time
* Revert "test time"
This reverts commit 2c5caabe2b.
* tmut
* i concur
* revert
old-commit-hash: ac83318ac4
* test_moedls: check controlsd events
* clean up
* sort
* controlsd modifies CP
* fix
* can use any segment now
* final fix
* use constant
old-commit-hash: 1410a1150c
* wait to leave elm mode before checking relay malfunction
* clean up
* clean up
* pycharm is best
* clean up
* more
* use pandaStatesDEPRECATED
* fix
* use constant
* only radar
* cmt
* set inline
* comment and re-arrange
* enable for relay malfunc check
* down here
* Update selfdrive/car/tests/test_models.py
* rm
* don't affect other tests
* up here
* one line
* update cmt
* no tolerance
* Revert "no tolerance"
This reverts commit 41b1c7e8be.
* comment is real solution
---------
Co-authored-by: Kacper Rączy <gfw.kra@gmail.com>
old-commit-hash: 5d1550de87
* timing test
* test
* fix
* print
* loop
* clean up
* wider range
* Update selfdrive/car/tests/test_models.py
* Apply suggestions from code review
* run many times
* Update selfdrive/car/tests/test_models.py
* run for many its
* run with unittest and print as list
* Update .github/workflows/selfdrive_tests.yaml
* Update .github/workflows/selfdrive_tests.yaml
* total time is super inconsistent (body)
* Update selfdrive/car/tests/test_models.py
* clean up
* clean up
* clean up
* this works!
* draft
* test suite not as modular
* try something like this
* can do kb, but not too representative
* clean up
* remove kb? it depends on signals
* clean up
* more clean up
* rename
* just measure all CANParsers
* can do all this manually
* but this is way simpler
* comment
* fix mypy
* this is correct too
* 3 seconds instead of 1.5
old-commit-hash: e2910d0720
* check relay malfunction in test_models
* draft
* actual relay malfunction (or ASCM was turned on)
* rav4: not dashcam and op enabled, but we're not sending anything
* clean up
* spaces
* more space
* flip
* better?
* combine
* for
* move up
old-commit-hash: 1099b3486a
* move to openpilot
* draft
* support internal urls
* not used
* update seg list
* regen segment list
* regen seg list
* regen
* regen
* regen
* no dirty segments
* regen seg list with no fuzzy or fixed fp
* regen segments with hda2 exception
* regen with added filters
* regen without bad dongle
* regen
* regenerate!
* regenerate, only OP enabled
* regen
* regenerate!
* regenerate!
* stash
* use SegmentName
* new route list
* add temp comment
* remove comment
* let's see if this works
* comment out other tests to speed up
* fix agent
* ?
* hmm
* oh whoops...
* add matrix
* fix
* how many levels do we need?
* fix values
* matrix can't be in parallel
* how about this
* try this
* Revert matrix
Revert "try this"
This reverts commit 8d0d91fd70.
Revert "how about this"
This reverts commit a8e4fc5323.
Revert "matrix can't be in parallel"
This reverts commit daaa6fcc3c.
Revert "fix values"
This reverts commit df554b6a33.
Revert "how many levels do we need?"
This reverts commit 1a17320fa1.
Revert "fix"
This reverts commit e7eb6e4043.
Revert "add matrix"
This reverts commit a1b57e5725.
* use pytest instead! (5 jobs is ~150 mins)
* split lines, uncomment
* This Sienna seg has a fault SDSU that stopped forwarding/sending msgs
* picked a route with no PSCMStatus and no panda errors, recent routes are working
* this cadillac was dashcammed (no radar)
* opened an issue for this, it's 'expected' right now
* small clean up
* small clean up
* i don't think that worked
* is this needed?
* add to new PC tests
* cache
* draft
* Revert "draft"
This reverts commit 3b7f740dd4.
* probably should be function
* draft
* clean up
* add todo
* 600 random segments
* debug
* does this fix pythonpath issues?
fix
* try this
* mount?
* pytest again!
* no need for PYTHONPATH now
* Update Jenkinsfile
* ?
* convention
* clean up
* would be even more complex (have to unset ci which is class level)
* track
* is lfs pulled at all?
* ah no it's not
old-commit-hash: 70b1e9dd75
* draft
* assertrue
* this is way too complicated, even with unittest discover
* works
* test we can send resume button
* nice catches this
* fix
* fixed
* not needed
* draft
* works
* comment out to let catch
* clean up into a function
* clean ups
* final clean up
* no this is final
* bump
* simplify
* simplify
* pass
* bump
* globals don't work inside methods
* space
old-commit-hash: 80bbba14f7
* draft
* bump opendbc
* still draft
* that's not right
* superset of the changes, 33hz
* cleanup
* this should work
* remove line
* pass it in again
* actually no need to check updated now
* now_nanos
* consistent name
* fix replay
* one line isn't that bad
switch
switch
* fix CarController tests
* Update ref_commit
old-commit-hash: 8f5057ff2d
* panda/SConscript
* think this is right?
* bump panda
* fix test_models
* use a filter
* bump panda to mater
* more aesthetic ordering
* alphabetical :(
* alphabetical :(
old-commit-hash: 01dac5d06b
* add regen braking field that's just used to add a pedalPressed
* bump
* bump
* Update ref_commit
* we want the standstill check
we want the standstill check
* see what diff is now
* Update ref_commit
old-commit-hash: 74b6e22a7d
* gm: match panda standstill check
* fix
* needs to be <= 10 to avoid a fault, fix for safety tests
* fix
* fix
* bump panda to master
old-commit-hash: 826d8a8ae3
Justin Newberry
Shane Smiskol
Adeeb Shihadeh
Lee Jong Mun