* jotpluggler!
* demo, executable, fontfile
* calc max and min, numpy, cloudlog
* mypy things
* simplified data.py
* multiprocessed data ingest
* allow verrryyy long search results
* stream in multiprocessed segments
* bug fixes
* simplify and speed up timeseries
* small fixes
* rewrite layout
* resizable layouts
* cleanup
* downsampling
* deque for consistency
* use item_visible_handler
* only build visible UI
* don't delete item handlers, add locks, don't expand large lists
* delete item handlers after a frame
* small data tree improvements
* seperate datatree into its own file
* reset when loading new segments
* fix plot window resizing and recursive split resizing logic
* stasj
* remove one of many classes
* clean up and fix
* clean up
* stash/draft: oh this is sick
* so epic
* some clean up
* what the fuck, it doesn't even use these
* more epic initializers + make it kind of work
* so simple, wonder if we should further 2x reduce line count
* i've never ever seen this pattern b4, rm
* remove bs add niceness
* minor organization
* set security type and support listing and rming conns
* forget and connect
* jeepney is actually pretty good, it's 2x faster to get wifi device (0.005s to 0.002s)
* temp
* do blocking add in worker thread
* add jeepney
* lets finish with python-dbus first then evaluate - revert jeepney
This reverts commit 7de04b11c2285c298bb1ec907782026c795ab207.
and
* safe wrap
* missing
* saved connections
* set rest of callbacks
* skip hidden APs, simplify _running
* add state management
* either wrong password or disconnected for now
* i can't believe we didn't check this...
* disable button if unsupported!!!
* hide/show event no lag hopefully yayay
* fix hide event
* remove old wifi manager
* cache wifi device path + some clean up
* more clean up
* more clean up
* temp disable blocking prime thread
* hackily get device path once
* ok
* debug
* fix open networks
* debug
* clean up
* all threads wait for device path, and user functions dont ever attempt to get, just skip
* same place
* helper
* Revert "helper"
This reverts commit e237d9a720915fb6bd67c539123d3e2d9d582ce1.
* organize?
* Revert "organize?"
This reverts commit 3aca3e5d629c947711ade88799febeb3f61eda87.
* c word is a bad word
* rk monitor debug for now
* nothing crazy
* improve checkmark responsiveness
* when forgetting: this is correct, but feels unresponsive
* this feels good
* need these two to keep forgetting and activating responsive
* sort by connected, strength, then name
* handle non-critical race condition
* log more
* unused
* oh jubilee is sick you can block on signals!!
* proof of concept to see if works on device
whoiops
* so sucking fick
* ah this is not generic, it's a filter on the return vals
* flip around to not drop
* oh thank god
* fix
* stash
* atomic replace
* clean up
* add old to keep track of what's moved over
* these are already moved
* so much junk
* so much junk
* more
* tethering wasn't used so we can ignore that for now
* no params now
* rm duplicate imports
* not used anymore
* move get wifi device over to jeepney! ~no additional lines
* request scan w/ keepney
* get_conns
* _connection_by_ssid_jeepney is 2x faster (0.01 vs 0.02s)
* do forget and activate
* _update_networks matches!
* rm old update_networks
* replace connect_to_network, about same time (yes i removed thread call)
* no more python-dbus!k
* doesn't hurt
* AP.from_dbus: actually handle incorrect paths w/ jeep + more efficient single call
* properly handle errors
* it's jeepney now
* less state
* using the thread safe router passes a race condition test that conn failed!
* bad to copy from old wifimanager
* fix conn usage
* clean up
* curious if locks are lagging
* not for now
* Revert "curious if locks are lagging"
This reverts commit 085dd185b083f5905a4e71ba3e8c0565175e04aa.
* clean up _monitor_state
* remove tests
* clean up dataclasses
* sort
* lint: okay fine it can be non by virtue of exiting right at the perfect time
* some network clean up
* some wifi manager clean up
* this is handled
* stop can be called manually, from deleting wifimanager, or exiting python. some protection
* its not mutable anymore
* scan on enter
* clean up
* back
* lint
* catch dbus fail to connect
catch dbus fail to connect
* ban it!
* some legitimate blunders
* even more!
* this is from commaai/one
* here too
* this too
* ignore rest
* one liner
* Revert "one liner"
This reverts commit d96a1624032aaad829a005b523e579d953cfe690.
* draft
* here too
* fixes
* fix
* ugh more fix, wifiManager is craze
* more
* be optimistic and let refresh happen naturally, the immediate refresh causes some paths/active connect to temporarily be unavailable
selfdrive/ui/qt/network/wifi_manager.cc: DBus call error: "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the re
ply timeout expired, or the network connection was broken."
selfdrive/ui/qt/network/wifi_manager.cc: DBus call error: "Object path cannot be empty"
* nm is slow -- it takes 2s to commit to disk, and dbus errors are raised if you try again while previous is running (this is an ubuntu 24.04 bug)
* nice experience
* rm
* minor
* clean up
* self-explanatory
* rename
* rm this too
* revert
* draft
* Revert "draft"
This reverts commit 15283d977880fc60b8f9732772256e8337d6ac8e.
* Reapply "draft"
This reverts commit 8629921b0086ca71b88746d77ec7b8d3af3cd289.
* rm colors
* trinary, bit more code
* choose default when disabled
* only if enabling, wait for disable as normal
* remove original binary toggle
* clean up
* collapse
* clean up wifimanager
* update comment
* lite is a word
* Revert "make test_onroad faster (#34704)"
This reverts commit 470ec46830.
* try this
* add there
* cleanup
* try this
* lil more
* list
* classic
* don't skip test
* squash
* bump tg
* bump tg
* debump tinygrad
* bump tinygrad
* bump tg
* Skip init iteration
* fixes
* cleanups
* skip first test sample
* typos
* linter unhappy
* update cpu usage
* OPENCL just zeros for now
* imports
* Try printing
* Runs again, but slower
* unused import
* Allow more buffer with tg and all on gpu
* bump tinygrad
* seems ok
* stricter timings for driving looser for dm
* try llvm
* check nvidia
* More timeout for now
* make test pass
* Revert "try llvm"
This reverts commit ef136e478320101fea262bae3579e558da991902.
* small fixes
* whitespace
* revert test timeout
* No model runners
* Always CPU always fast
* No onnx runtime GPU
* more cores
* cleanup
* Is this faster
* Is this faster
* at least runs
* FP32 is faster than 16
* fix deps
* whitespace
* comment
---------
Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com>
* test selfdrived!
* exit() is for interactive sessions
* fix
* comments
* more
* test all of selfdrive/
* ignore what we used to
* fix test_alerts
* fix test_alertmanager.py
Jason Young