* big buffer
* mostly hooked up big buffer
* bad merge
* s
* sneak preview of trigger scope
* s
* constness
* channel limits
* s
* unnecessary
* efilib
* TcpServerSandbox
* includes fix
* binary format
* firmware error in case of low stack only on HW CI
* send date
* I don't java good
* I still don't java good
* send the right command
* I guess we can check it on every single test?
* javaaaaaa
* atoi breaks on later non-digit
* avoid sscanf use (for atoi)
to avoid bloating the stack
* less churn
* less churn
Co-authored-by: Nathan Schulte <nmschulte@desmas.net>
* logingcentral: check final string for prohibited chars
Invalid char can come from arguments, not only format string so
scan final buffer.
* logingcentral: comment say '\r' is also not allowed
* logingcentral: just replace all non-printable chars with space
Do not assert
* loging: remove strlen()
* Do lightweight checks first
* Dead getTimeIgnitionSeconds()
Dead since 81fffe87b7
* Get rid of efitime_t.
Purpose are not clear. Use efitick_t where ticks are used,
use efitimeus_t for uS, use efitimems_t or efitimems64_t for mS.
* Fix unit tests
* Fix misc: stm32f1_test_project
* Extract time helpers from engine_controller_misc to efitime.cpp
* Rename currentTimeMillis() to getTimeNowMs()
We have getTimeNowNt(), getTimeNowUs(), currentTimeMillis() and getTimeNowSeconds()
Align a bit.
find . -type f -name '*.c*' -exec sed -i 's/currentTimeMillis/getTimeNowMs/g' {} \;
* Rename getTimeNowSeconds() to getTimeNowS()
To align with Nt, Us, Ms versions.
* Some comments about getTimeNowLowerNt()
* allow logging multiple warnings at once
* comment
* .Code
* Revert ".Code"
This reverts commit 2b986bd50035aeec051d11aafd83fdbc1694351c.
* add a better comparison operator instead of conversion
* dot code
* Revert "dot code"
This reverts commit 35f6ae1007c41e3b30ea129a324f33ab4205036d.
* force gcc-10 maybe?
* Revert "force gcc-10 maybe?"
This reverts commit 4ddf8bcefddd354ccdc1da682c5fff1e68d44273.
* vq
* dot code
* dead
* dot code
* return may be null
* static_vector; maybe cyclic_buffer is broken or wrong tool for the job?
* move static vector
* put cyclic buffer back how it was since we don't use it now
* make things c++ friendly
* install various canaries in coal mines
* using is much clearer than typedef
* fixing builds
* lps25
* trigger includes were a mess
* c -> c++ cleanup
* os util
* why did so many include os_util.h?
* fwd declare
* cypress/kinetis
* s
* eg33 happy
* happy sim
* cypress usb
* pch
* cypress
* doesn't need to include all that
* h7
* shuffle around output channels
* postState should take a ref
* s
* s
* patch input file
* limits
* accel bad bounds check
* units
* units
* Revert "accel bad bounds check"
This reverts commit b2bb493be331d5d2cf40b9ed01363a53182c883e.
* missed an autoscale
* cli registry: rework this hell, add FFF and FFFFF
* kinetis: no nodefaultlib
* revert back token helpers
* cli_registery: argument parser: do not eat qoutes
* Unit test for FFF
* cli_registry: remove debug
* less dependence on that logger
* comma
* refactor a bit, make it clear what this thing actually *does*
* static
* shrink the buffer a bit and make names useful
* dead util that's just subtraction lol
* consumers and api
* dead test
* baro uses sensor model
* remove old (copy of?) map lookup logic
* I guess we don't need FastInterpolation any more?
* don't double init analog input
* ui & config
* bump flash version
* compaction
* s
* make things happy
* fsio too
* OK, well those defaults were insane
* b
* why are these defaults insane
* adjust cylinder offset ui
* s
* cleanup
This exposed a buffer overrun, so double the size of the buffer (to account for 720 degree engine
cycle vs 360 degree crank events).
Also use proper numeric limits when computing min/max.
Finally, add a lock around the call to cyclic_buffer that actually caused the contention.
rusefillc
Matthew Kennedy
rusefillc
Nathan Schulte
Andrey G
rusefillc
Andreika
GitHub build-firmware Action
Scott Smith