* cameraview: Scale the glViewPort by devicePixelRatio()
This fixes an issue seen in Cabana on a "hidpi" system with Wayland where
devicePixelRatio() != 1 and the video doesn't take up the full widget area.
On the recommended Ubuntu 20.04 install I wasn't able to reproduce, because
devicePixelRatio() was always equal to 1 even with scaling to 200% or 300%. It
might be different if "Fractional Scaling" is enabled in GNOME (I couldn't make
that option work in mv WM.)
Was going to enable just for Linux, but it appears to also be recommended for
Retina MacOS:
https://doc.qt.io/qt-5/scalability.html#high-dpi-scaling-on-macos-and-ios
... so have worked from the assumption that glViewport() always takes dimensions
in device pixels, never the "device independent pixels" of Qt.
* Update selfdrive/ui/qt/widgets/cameraview.cc
* Update selfdrive/ui/qt/widgets/cameraview.cc
---------
Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com>
old-commit-hash: fbcd0821c9
* speed up fingerprinting for CAN fp cars
* add comment to describe this behavior
* more explicit, no defaultdict
* do this later
* fix
* a little more clear I think
old-commit-hash: 9b1095a27e
* Update counter in carstate checks
* Rename es_lkas to es_lkas_state to match signal
* formatting
---------
Co-authored-by: Shane Smiskol <shane@smiskol.com>
old-commit-hash: 13c75dc138
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* wip
* remove compiler flag stuff
* force error for testing
* bump panda and fix mistake
* wip
* bump panda
* wip
* wip
* forgot checksum and stuff
* add to signals
* rename
* test
* test
* bump panda and opendbc
* bump cereal
* add blank filler
* test
* add steer required to infotainment
* fix missing bit
* fix missing bit
* cleanup for PR
* cleanup for PR
* bump submodules
* wip
* wip
* i wonder what 1 is
* seemed to also be obstacle detected?
* bumppanda
* only if global
* only global gen 2
* fix order
* detect and send
* detect in rest of places
* update refs
* bumpopendbc
* copy whole message (camera sometimes sets these signals)
* bumppanda
* update refs!
* bumppanda
* let's not change this
* same as all the other functions
* Revert "same as all the other functions"
This reverts commit e86dbb695ef84a18f3123720fed37a6b81391cc4.
---------
Co-authored-by: Shane Smiskol <shane@smiskol.com>
old-commit-hash: d59bfa5519
* can FP time
* add
* move the ref up
* assert function
* all brands
* Update selfdrive/car/tests/test_fw_fingerprint.py
* can this can vary a bit more
* or can it
old-commit-hash: 158c951d6f
* initial commit
* HR-V is diff enought to split from Civic22
* Add US HR-V 2023
* update CARS.md
* Update RELEASES.md
* Use a alternative brake signal
* US HR-V have a weird noise in the BrakeSwitch
* Shane tunning
* this is added automatically
* sorry jyoung
* civic22_stop_faster
limit stopAccel to current safety model
* relax calib temporarily (12 deg)
relax calib temporarily (10 deg)
* compute accel (car overshoots requests at low speed)
* revert brake pressed caused controls mismatches
* fix crash
* cause a cruise fault
* Revert "cause a cruise fault"
This reverts commit 09a6469ac6fc38ebd38dd1c5d464712a15d0fde1.
* no brake error
* use alt brake
* Update selfdrive/car/honda/carcontroller.py
* change HONDA_BOSCH_ALT_BRAKE_SIGNAL to CP.flags
* Revert "change HONDA_BOSCH_ALT_BRAKE_SIGNAL to CP.flags"
This reverts commit e0f8123b33db3218f27c4b64a7b2a0b1f740860f.
* remove FPv2 from brazilian HR-V
* bump panda
* fix a bug
* revert calib threshold
* new test route
* separate pr
* update
* update that
* thought i removed this
* revert steer ratio
---------
Co-authored-by: Shane Smiskol <shane@smiskol.com>
old-commit-hash: 93c1810522
* add timing unit test
* formatting
* not clean
* Revert "not clean"
This reverts commit dab35d582c274efc32d693fc0a2773d8727be2bd.
* Revert "formatting"
This reverts commit 56a1ae4bb721594cc68952e5747a3ebc652f62d9.
* need to consider subaddrs (0.3 to 0.7s for toyota)
* clean up
* move to test_fw_fingerprint
* use a thread to get real time
* lower threshold
* revert that
* ?
* some sleep
* at top
* use the config
* function to measure query time, similar to opendbc's convention
* revert that
* avg
* Update selfdrive/car/tests/test_fw_fingerprint.py
* draft
* starting a thread takes a long time!
* refs
* uncomment
old-commit-hash: f222d434b8
* add ES 2018
* standard
* add FW versions
* need to add a route
* verified
* add to DSU messages
* add to torque params
* older lexus do not have all-speed dynamic radar cruise control
* update docs
* Update selfdrive/car/toyota/values.py
* add CI route
* update docs
* add to releases
* assume no stop and go (can't know without dsu unplugged)
* Update selfdrive/car/toyota/values.py
* update docs
* update docs
old-commit-hash: 57f3b483e3
* Add trailblazer support
* Update routes.py
* Update routes.py
* Update CARS.md
* Fix incorrect torque name
* Update values.py and CARS.md
* Change to update another route
* update docs
* only 2021 https://en.wikipedia.org/wiki/Chevrolet_Trailblazer_(crossover)
only 2021
* new FP with missing addrs
* Update selfdrive/car/gm/interface.py
* tune the torque params
* update docs
* remove as a new platform
* Revert "remove as a new platform"
This reverts commit c9c65f29dbb1e1a5926540bb75433928325d6c0e.
* add comment and comment out FP
* need it to be in all_known_cars to test_model
* add to releases
---------
Co-authored-by: Shane Smiskol <shane@smiskol.com>
old-commit-hash: e1ad8df519
* better ordering and comments
* remove
* mark done if timed out
* response_timeouts should be only for extending timeouts, not marking done
* so the print still works
* more clear comment
* that can be removed
old-commit-hash: f5519fdca8
Erich Moraga