By having a pair of buffers, the capture hardware can fill one buffer while
Python code (including displayio, etc) operates on the other buffer. This
increases the responsiveness of camera-using code.
On the Kaluga it makes the following improvements:
* 320x240 viewfinder at 30fps instead of 15fps using directio
* 240x240 animated gif capture at 10fps instead of 7.5fps
As discussed at length on Discord, the "usual end user" code will look like
this:
camera = ...
with camera.continuous_capture(buffer1, buffer2) as capture:
for frame in capture:
# Do something with frame
However, rather than presenting a context manager, the core code consists of
three new functions to start & stop continuous capture, and to get the next
frame. The reason is twofold. First, it's simply easier to implement the
context manager object in pure Python. Second, for more advanced usage, the
context manager may be too limiting, and it's easier to iterate on the right
design in Python code. In particular, I noticed that adapting the
JPEG-capturing programs to use continuous capture mode needed a change in
program structure.
The camera app was structured as
```python
while True:
if shutter button was just pressed:
capture a jpeg frame
else:
update the viewfinder
```
However, "capture a jpeg frame" needs to (A) switch the camera settings and (B)
capture into a different, larger buffer then (C) return to the earlier
settings. This can't be done during continuous capture mode. So just
restructuring it as follows isn't going to work:
```python
with camera.continuous_capture(buffer1, buffer2) as capture:
for frame in capture:
if shutter button was just pressed:
capture a jpeg frame, without disturbing continuous capture mode
else:
update the viewfinder
```
The continuous mode is only implemented in the espressif port; others
will throw an exception if the associated methods are invoked. It's not
impossible to implement there, just not a priority, since these micros don't
have enough RAM for two framebuffer copies at any resonable sizes.
The capture code, including single-shot capture, now take mp_obj_t in the
common-hal layer, instead of a buffer & length. This was done for the
continuous capture mode because it has to identify & return to the user the
proper Python object representing the original buffer. In the Espressif port,
it was convenient to implement single capture in terms of a multi-capture,
which is why I changed the singleshot routine's signature too.
New design:
* capture output to a vstr
* compare the complete vstr to boot_out.txt
* rewrite if not a complete match
This is resilient against future changes to the automatic
text written to boot_out.txt.
This also fixes rewriting boot_out.txt in the case where
boot.py prints something.
Perhaps it also saves a bit of code space. Some tricks:
* no need to close a file in read mode
* no need to switch on/off USB write access, going down to the
oofatfs layer doesn't check it anyway
This makes it possible for cooperative multitasking systems to keep running
event loops during garbage collector operations.
For example, this can be used to ensure that a motor control loop runs
approximately each 5 ms. Without this hook, the loop time can jump to
about 15 ms.
Addresses #3475.
Signed-off-by: Laurens Valk <laurens@pybricks.com>
This is an stm32-specific feature that's accessed via the pyb module, so
not something that will be widely enabled.
Signed-off-by: Damien George <damien@micropython.org>
Some of these will later be moved to CORE or BASIC, but EXTRA is a good
starting point based on what stm32 uses.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This involves:
* Adding a new "L8" colorspace for colorconverters
* factoring out displayio_colorconverter_convert_pixel
* Making a minimal "colorspace only" version of displayio for the
unix port (testing purposes)
* fixing an error message
I only tested writing B&W animated images, with the following script:
```python
import displayio
import gifio
with gifio.GifWriter("foo.gif", 64, 64, displayio.Colorspace.L8) as g:
for i in range(0, 256, 14):
data = bytes([i, 255-i] * 32 + [255-i, i] * 32) * 32
print("add_frame")
g.add_frame(data)
# expected to raise an error, buffer is not big enough
with gifio.GifWriter("/dev/null", 64, 64, displayio.Colorspace.L8) as g:
g.add_frame(bytes([3,3,3]))
```
Before this, "%c" % 3.14 would raise the error `integer required`
even though python3 uses `%c requires int or char`, just like both
already did for "%c" % "ab".
This gets rid of a string to translate, and makes behavior match
python3 better.
F-strings, new machine.I2S class, ESP32-C3 support and LEGO_HUB_NO6 board
This release of MicroPython adds support for f-strings (PEP-498), with a
few limitations compared to normal Python. F-strings are essentially
syntactic sugar for "".format() and make formatting strings a lot more
convenient. Other improvements to the core runtime include pretty printing
OSError when it has two arguments (an errno code and a string), scheduling
of KeyboardInterrupt on the main thread, and support for a single argument
to the optimised form of StopIteration.
In the machine module a new I2S class has been added, with support for
esp32 and stm32 ports. This provides a consistent API for transmit and
receive of audio data in blocking, non-blocking and asyncio-based
operation. Also, the json module has support for the "separators" argument
in the dump and dumps functions, and framebuf now includes a way to blit
between frame buffers of different formats using a palette. A new,
portable machine.bitstream function is also added which can output a stream
of bits with configurable timing, and is used as the basis for driving
WS2812 LEDs in a common way across ports.
There has been some restructuring of the repository directory layout, with
all third-party code now in the lib/ directory. And a new top-level
directory shared/ has been added with first-party code that was previously
in lib/ moved there.
The docs have seen further improvement with enhancements and additions to
the rp2 parts, as well as a new quick reference for the zephyr port.
The terms master/slave have been replaced with controller/peripheral,
mainly relating to I2C and SPI usage. And u-module references have been
replaced with just the module name without the u-prefix to help clear up
the intended usage of modules in MicroPython.
For the esp8266 and esp32 ports, hidden networks are now included in WLAN
scan results. On the esp32 the RMT class is enhanced with idle_level and
write_pulses modes. There is initial support for ESP32-C3 chips with
GENERIC_C3 and GENERIC_C3_USB boards.
The javascript port has had its Makefile and garbage collector
implementation reworked so it compiles and runs with latest the Emscripten
using asyncify.
The mimxrt port sees the addition of hardware I2C and SPI support, as well
as some additional methods to the machine module. There is also support
for Hyperflash chips.
The nrf port now has full VFS storage support, enables source-line on
traceback, and has .mpy features consistent with other ports.
For the rp2 port there is now more configurability for boards, and more
boards added.
The stm32 port has a new LEGO_HUB_NO6 board definition with detailed
information how to get this LEGO Hub running stock MicroPython. There is
also now support to change the CPU frequency on STM32WB MCUs. And USBD_xxx
descriptor options have been renamed to MICROPY_HW_USB_xxx.
Thanks to everyone who contributed to this release: Amir Gonnen, Andrew
Scheller, Bryan Tong Minh, Chris Wilson, Damien George, Daniel Mizyrycki,
David Lechner, David P, Fernando, finefoot, Frank Pilhofer, Glenn Ruben
Bakke, iabdalkader, Jeff Epler, Jim Mussared, Jonathan Hogg, Josh Klar,
Josh Lloyd, Julia Hathaway, Krzysztof Adamski, Matúš Olekšák, Michael
Weiss, Michel Bouwmans, Mike Causer, Mike Teachman, Ned Konz, NitiKaur,
oclyke, Patrick Van Oosterwijck, Peter Hinch, Peter Züger, Philipp
Ebensberger, robert-hh, Roberto Colistete Jr, Sashkoiv, Seon Rozenblum,
Tobias Thyrrestrup, Tom McDermott, Will Sowerbutts, Yonatan Goldschmidt.
What follows is a detailed list of changes, generated from the git commit
history, and organised into sections.
Main components
===============
all:
- fix signed shifts and NULL access errors from -fsanitize=undefined
- update to point to files in new shared/ directory
py core:
- mpstate: make exceptions thread-local
- mpstate: schedule KeyboardInterrupt on main thread
- mperrno: add MP_ECANCELED error code
- makeqstrdefs.py: don't include .h files explicitly in preprocessing
- mark unused arguments from bytecode decoding macros
- objexcept: pretty print OSError also when it has 2 arguments
- makeversionhdr: add --tags arg to git describe
- vm: simplify handling of MP_OBJ_STOP_ITERATION in yield-from opcode
- objexcept: make mp_obj_exception_get_value support subclassed excs
- support single argument to optimised MP_OBJ_STOP_ITERATION
- introduce and use mp_raise_type_arg helper
- modsys: optimise sys.exit for code size by using exception helpers
- objexcept: make mp_obj_new_exception_arg1 inline
- obj: fix formatting of comment for mp_obj_is_integer
- emitnative: reuse need_reg_all func in need_stack_settled
- emitnative: ensure stack settling is safe mid-branch
- runtime: fix bool unary op for subclasses of native types
- builtinimport: fix condition for including do_execute_raw_code()
- mkrules: automatically build mpy-cross if it doesn't exist
- implement partial PEP-498 (f-string) support
- lexer: clear fstring_args vstr on lexer free
- mkrules.mk: do submodule sync in "make submodules"
extmod:
- btstack: add missing call to mp_bluetooth_hci_uart_deinit
- btstack: check that BLE is active before performing operations
- uasyncio: get addr and bind server socket before creating task
- axtls-include: add axtls_os_port.h to customise axTLS
- update for move of crypto-algorithms, re1.5, uzlib to lib
- moduselect: conditionally compile select()
- nimble: fix leak in l2cap_send if send-while-stalled
- btstack/btstack.mk: use -Wno-implicit-fallthrough, not =0
- utime: always invoke mp_hal_delay_ms when >= to 0ms
- modbluetooth: clamp MTU values to 32->UINT16_MAX
- nimble: allow modbluetooth binding to hook "sent HCI packet"
- nimble: add "memory stalling" mechanism for l2cap_send
- uasyncio: in open_connection use address info in socket creation
- modujson: add support for dump/dumps separators keyword-argument
- modlwip: fix close and clean up of UDP and raw sockets
- modbluetooth: add send_update arg to gatts_write
- add machine.bitstream
- modframebuf: enable blit between different formats via a palette
lib:
- tinyusb: update to version 0.10.1
- pico-sdk: update to version 1.2.0
- utils/stdout_helpers: make mp_hal_stdout_tx_strn_cooked efficient
- axtls: switch to repo at micropython/axtls
- axtls: update to latest axtls 2.1.5 wih additional commits
- re1.5: move re1.5 code from extmod to lib
- uzlib: move uzlib code from extmod to lib
- crypto-algorithms: move crypto-algorithms code from extmod to lib
- update README's based on contents of these dirs
drivers:
- neopixel: add common machine.bitstream-based neopixel module
- neopixel: optimize fill() for speed
- neopixel: reduce code size of driver
- cyw43: fix cyw43_deinit so it can be called many times in a row
- cyw43: make wifi join fail if interface is not active
mpy-cross:
- disable stack check when building with Emscripten
Support components
==================
docs:
- library: document new esp32.RMT features and fix wait_done
- library: warn that ustruct doesn't handle spaces in format strings
- esp8266/tutorial: change flash mode from dio to dout
- replace master/slave with controller/peripheral in I2C and SPI
- rp2: enhance quickref and change image to Pico pinout
- rp2: update general section to give a brief technical overview
- library/utime.rst: clarify behaviour and precision of sleep ms/us
- library/uasyncio.rst: document stream readexactly() method
- library/machine.I2S.rst: fix use of sd pin in examples
- zephyr: add quick reference for the Zephyr port
- library/zephyr: add libraries specific to the Zephyr port
- templates: add unix and zephyr quickref links to top-index
- rename ufoo.rst to foo.rst
- replace ufoo with foo in all docs
- library/index.rst: clarify module naming and purpose
- library/builtins.rst: add module title
- library/network.rst: simplify socket import
- add docs for machine.bitstream and neopixel module
- library: fix usage of :term: for frozen module reference
- esp8266: use monospace for software tools
- reference: mention that slicing a memoryview causes allocation
examples: no changes specific to this component/port
tests:
- extmod: make uasyncio_heaplock test more deterministic
- cpydiff/modules_struct_whitespace_in_format: run black
- extmod/ujson: add tests for dump/dumps separators argument
- run-multitests.py: add broadcast and wait facility
- multi_bluetooth/ble_subscribe.py: add test for subscription
- extmod/vfs_fat_finaliser.py: ensure alloc at never-used GC blocks
- basics: split f-string debug printing to separate file with .exp
- pybnative: make while.py test run on boards without pyb.delay
tools:
- autobuild: add scripts to build release firmware
- remove obsolete build-stm-latest.sh script
- ci.sh: run apt-get update in ci_powerpc_setup
- makemanifest.py: allow passing flags to mpy-tool.py
- autobuild: add mimxrt port to build scripts for nightly builds
- pyboard.py: add cmd-line option to make soft reset configurable
- mpremote: swap order of PID and VID in connect-list output
- ci.sh: build unix dev variant as part of macOS CI
- ci.sh: build GENERIC_C3 board as part of esp32 CI
- autobuild: use separate IDF version to build newer esp32 SoCs
- autobuild: add FeatherS2 and TinyS2 to esp32 auto builds
- mpremote: add seek whence for mounted files
- mpremote: raise OSError on unsupported RemoteFile.seek
- autobuild: add the MIMXRT1050_EVKB board to the daily builds
- ci.sh: add mpy-cross build to nrf port
- codeformat.py: include ports/nrf/modules/nrf in code formatting
- gen-cpydiff.py: don't rename foo to ufoo in diff output
- autobuild: add auto build for Silicognition wESP32
- mpremote: fix connect-list in case VID/PID are None
- mpremote: add "devs" shortcut for "connect list"
- mpremote: remove support for pyb.USB_VCP in/out specialisation
- autobuild: don't use "-B" for make, it's already a fresh build
- pyboard.py: move --no-exclusive/--soft-reset out of mutex group
- pyboard.py: make --no-follow use same variable as --follow
- pyboard.py: add --exclusive to match --no-exclusive
- pyboard.py: make --no-soft-reset consistent with other args
- uncrustify: force 1 newline at end of file
- mpremote: bump version to 0.0.6
CI:
- workflows: add workflow to build and test javascript port
- workflows: switch from Coveralls to Codecov
- workflows: switch from lcov to gcov
- workflows: add workflow to build and test unix dev variant
The ports
=========
all ports:
- use common mp_hal_stdout_tx_strn_cooked instead of custom one
- update for move of crypto-algorithms, uzlib to lib
- rename USBD_VID/PID config macros to MICROPY_HW_USB_VID/PID
bare-arm port: no changes specific to this component/port
cc3200 port: no changes specific to this component/port
esp8266 port:
- add __len__ to NeoPixel driver to support iterating
- Makefile: add more libm files to build
- include hidden networks in WLAN.scan results
- replace esp.neopixel with machine.bitstream
- remove dead code for end_ticks in machine_bitstream
esp32 port:
- boards/sdkconfig.base: disable MEMPROT_FEATURE to alloc from IRAM
- add __len__ to NeoPixel driver to support iterating
- main: allow MICROPY_DIR to be overridden
- esp32_rmt: fix RMT looping in newer IDF versions
- esp32_rmt: enhance RMT with idle_level and write_pulses modes
- add new machine.I2S class for I2S protocol support
- machine_spi: calculate actual attained baudrate
- machine_hw_spi: use a 2 item SPI queue for long transfers
- machine_dac: add MICROPY_PY_MACHINE_DAC option, enable by default
- machine_i2s: add MICROPY_PY_MACHINE_I2S option, enable by default
- fix use of mp_int_t, size_t and uintptr_t
- add initial support for ESP32C3 SoCs
- boards/GENERIC_C3: add generic C3-based board
- modmachine: release the GIL in machine.idle()
- mphalport: always yield at least once in delay_ms
- machine_uart: add flow kw-arg to enable hardware flow control
- boards: add Silicognition wESP32 board configuration
- mpconfigport.h: enable reverse and inplace special methods
- include hidden networks in WLAN.scan results
- makeimg.py: get bootloader and partition offset from sdkconfig
- enable MICROPY_PY_FSTRINGS by default
- machine_hw_spi: release GIL during transfers
- machine_pin: make check for non-output pins respect chip variant
- replace esp.neopixel with machine.bitstream
- remove dead code for end_ticks in machine_bitstream
- boards: add GENERIC_C3_USB board with USB serial/JTAG support
javascript port:
- rework Makefile and GC so it works with latest Emscripten
- Makefile: suppress compiler errors from array bounds
- Makefile: change variable to EXPORTED_RUNTIME_METHODS
mimxrt port:
- move calc_weekday helper function to timeutils
- machine_spi: add the SPI class to the machine module
- moduos: seed the PRNG on boot using the TRNG
- boards: set vfs partition start to 1 MBbyte
- main: skip running main.py if boot.py failed
- main: extend the information returned by help()
- mimxrt_flash: remove commented-out code
- modmachine: add a few minor methods to the machine module
- machine_led: use mp_raise_msg_varg helper
- machine_i2c: add hardware-based machine.I2C to machine module
- add support for Hyperflash chips
- boards: add support for the MIMXRT1050_EVKB board
- machine_pin: implement ioctl for Pin
minimal port:
- Makefile: add support for building with user C modules
nrf port:
- modules: replace master/slave with controller/peripheral in SPI
- boards/common.ld: calculate unused flash region
- modules/nrf: add new nrf module with flash block device
- drivers: add support for using flash block device with SoftDevice
- mpconfigport.h: expose nrf module when MICROPY_PY_NRF is set
- README: update README.md to reflect internal file systems
- mpconfigport.h: tune FAT FS configuration
- Makefile: add _fs_size linker script override from make
- modules/uos: allow a board to configure MICROPY_VFS_FAT/LFS1/LFS2
- mpconfigport.h: enable MICROPY_PY_IO_FILEIO when an FS is enabled
- qstrdefsport.h: add entries for in-built FS mount points
- main: add auto mount and auto format hook for internal flash FS
- boards: enable needed features for FAT/LFS1/LFS2
- facilitate use of freeze manifest
- boards: set FROZEN_MANIFEST blank when SD present on nrf51 targets
- modules/scripts: add file system formatting script
- Makefile: set default manifest file for all targets
- mphalport: add dummy function for mp_hal_time_ns()
- boards: enable MICROPY_VFS_LFS2 for all target boards
- modules/uos: add ilistdir to uos module
- modules/nrf: add function to enable/disable DCDC
- enable source line on tracebacks
- set .mpy features consistent with documentation and other ports
pic16bit port: no changes specific to this component/port
powerpc port: no changes specific to this component/port
qemu-arm port: no changes specific to this component/port
rp2 port:
- use 0=Monday datetime convention in RTC
- machine_rtc: in RTC.datetime, compute weekday automatically
- CMakeLists.txt: suppress compiler errors for pico-sdk and tinyusb
- tusb_config.h: set CFG_TUD_CDC_EP_BUFSIZE to 256
- machine_uart: add hardware flow control support
- machine_uart: allow overriding default machine UART pins
- machine_i2c: allow boards to configure I2C pins using new macros
- machine_spi: allow boards to configure SPI pins using new macros
- machine_uart: fix poll ioctl to also check hardware FIFO
- machine_uart: fix read when FIFO has chars but ringbuf doesn't
- tusb_port: allow boards to configure USB VID and PID
- boards/ADAFRUIT_FEATHER_RP2040: configure custom VID/PID
- boards/ADAFRUIT_FEATHER_RP2040: configure I2C/SPI default pins
- boards/SPARKFUN_PROMICRO: configure UART/I2C/SPI default pins
- boards/SPARKFUN_THINGPLUS: configure I2C/SPI default pins
- boards: add Adafruit ItsyBitsy RP2040
- boards: add Adafruit QT Py RP2040
- boards: add Pimoroni Pico LiPo 4MB
- boards: add Pimoroni Pico LiPo 16MB
- boards: add Pimoroni Tiny 2040
- CMakeLists.txt: allow a board's cmake to set the manifest path
- enable MICROPY_PY_FSTRINGS by default
- Makefile: add "submodules" target, to match other ports
- rp2_flash: disable IRQs while calling flash_erase/program
- CMakeLists.txt: add option to enable double tap reset to bootrom
- mpconfigport.h: allow boards to add root pointers
samd port:
- add support for building with user C modules
stm32 port:
- softtimer: add soft_timer_reinsert() helper function
- mpbthciport: change from systick to soft-timer for BT scheduling
- provide a custom BTstack runloop that integrates with soft timer
- usb: make irq's default trigger enable all events
- boardctrl: skip running main.py if boot.py had an error
- sdio: fix undefined reference to DMA stream on H7
- dma: add DMAMUX configuration for H7 to fix dma_nohal_init
- main: call mp_deinit() at end of main
- adc: allow using ADC12 and ADC3 for H7
- adc: define the ADC instance used for internal channels
- adc: simplify and generalise how pin_adcX table is defined
- add new machine.I2S class for I2S protocol support
- boards/NUCLEO_F446RE: fix I2C1 pin assignment to match datasheet
- replace master/slave with controller/peripheral in I2C and SPI
- systick: always POLL_HOOK when delaying for milliseconds
- sdram: make SDRAM test cache aware, and optional failure with msg
- boards/NUCLEO_F446RE: enable CAN bus support
- boards: add support for SparkFun STM32 MicroMod Processor board
- uart: fix LPUART1 baudrate set/get
- uart: support low baudrates on LPUART1
- boards/STM32F429DISC: set correct UART2 pins and add UART3/6
- boards/NUCLEO_F439ZI: add board definition for NUCLEO_F439ZI
- boards/LEGO_HUB_NO6: add board definition for LEGO_HUB_NO6
- Makefile: update to only pull in used Bluetooth library
- README.md: update supported MCUs, and submodule and mboot use
- usbd_desc: rename USBD_xxx descriptor opts to MICROPY_HW_USB_xxx
- usbd_cdc_interface: rename USBD_CDC_xx opts to MICROPY_HW_USB_xx
- powerctrl: support changing frequency on WB MCUs
- boards/NUCLEO_H743ZI2: add modified version of NUCLEO_H743ZI
- mbedtls: fix compile warning about uninitialized val
- enable MICROPY_PY_FSTRINGS by default
- add implementation of machine.bitstream
- Makefile: allow GIT_SUBMODULES and LIBS to be extended
- stm32_it: support TIM17 IRQs on WB MCUs
- disable computed goto on constrained boards
- storage: make extended-block-device more configurable
- boards/LEGO_HUB_NO6: change SPI flash storage to use hardware SPI
- boards/LEGO_HUB_NO6: skip first 1MiB of SPI flash for storage
- boards/LEGO_HUB_NO6: add make commands to backup/restore firmware
teensy port: no changes specific to this component/port
unix port:
- modffi: add option to lock GC in callback, and cfun access
- Makefile: add back LIB_SRC_C to list of object files
- variants: enable help and help("modules") on standard and dev
- Makefile: disable error compression on arm-linux-gnueabi-gcc
windows port:
- Makefile: add .exe extension to executables name
- appveyor: update to VS 2017 and use Python 3.8 for build/test
zephyr port:
- machine_spi: add support for hardware SPI
This commit removes all parts of code associated with the existing
MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE optimisation option, including the
-mcache-lookup-bc option to mpy-cross.
This feature originally provided a significant performance boost for Unix,
but wasn't able to be enabled for MCU targets (due to frozen bytecode), and
added significant extra complexity to generating and distributing .mpy
files.
The equivalent performance gain is now provided by the combination of
MICROPY_OPT_LOAD_ATTR_FAST_PATH and MICROPY_OPT_MAP_LOOKUP_CACHE (which has
been enabled on the unix port in the previous commit).
It's hard to provide precise performance numbers, but tests have been run
on a wide variety of architectures (x86-64, ARM Cortex, Aarch64, RISC-V,
xtensa) and they all generally agree on the qualitative improvements seen
by the combination of MICROPY_OPT_LOAD_ATTR_FAST_PATH and
MICROPY_OPT_MAP_LOOKUP_CACHE.
For example, on a "quiet" Linux x64 environment (i3-5010U @ 2.10GHz) the
change from CACHE_MAP_LOOKUP_IN_BYTECODE, to LOAD_ATTR_FAST_PATH combined
with MAP_LOOKUP_CACHE is:
diff of scores (higher is better)
N=2000 M=2000 bccache -> attrmapcache diff diff% (error%)
bm_chaos.py 13742.56 -> 13905.67 : +163.11 = +1.187% (+/-3.75%)
bm_fannkuch.py 60.13 -> 61.34 : +1.21 = +2.012% (+/-2.11%)
bm_fft.py 113083.20 -> 114793.68 : +1710.48 = +1.513% (+/-1.57%)
bm_float.py 256552.80 -> 243908.29 : -12644.51 = -4.929% (+/-1.90%)
bm_hexiom.py 521.93 -> 625.41 : +103.48 = +19.826% (+/-0.40%)
bm_nqueens.py 197544.25 -> 217713.12 : +20168.87 = +10.210% (+/-3.01%)
bm_pidigits.py 8072.98 -> 8198.75 : +125.77 = +1.558% (+/-3.22%)
misc_aes.py 17283.45 -> 16480.52 : -802.93 = -4.646% (+/-0.82%)
misc_mandel.py 99083.99 -> 128939.84 : +29855.85 = +30.132% (+/-5.88%)
misc_pystone.py 83860.10 -> 82592.56 : -1267.54 = -1.511% (+/-2.27%)
misc_raytrace.py 21490.40 -> 22227.23 : +736.83 = +3.429% (+/-1.88%)
This shows that the new optimisations are at least as good as the existing
inline-bytecode-caching, and are sometimes much better (because the new
ones apply caching to a wider variety of map lookups).
The new optimisations can also benefit code generated by the native
emitter, because they apply to the runtime rather than the generated code.
The improvement for the native emitter when LOAD_ATTR_FAST_PATH and
MAP_LOOKUP_CACHE are enabled is (same Linux environment as above):
diff of scores (higher is better)
N=2000 M=2000 native -> nat-attrmapcache diff diff% (error%)
bm_chaos.py 14130.62 -> 15464.68 : +1334.06 = +9.441% (+/-7.11%)
bm_fannkuch.py 74.96 -> 76.16 : +1.20 = +1.601% (+/-1.80%)
bm_fft.py 166682.99 -> 168221.86 : +1538.87 = +0.923% (+/-4.20%)
bm_float.py 233415.23 -> 265524.90 : +32109.67 = +13.756% (+/-2.57%)
bm_hexiom.py 628.59 -> 734.17 : +105.58 = +16.796% (+/-1.39%)
bm_nqueens.py 225418.44 -> 232926.45 : +7508.01 = +3.331% (+/-3.10%)
bm_pidigits.py 6322.00 -> 6379.52 : +57.52 = +0.910% (+/-5.62%)
misc_aes.py 20670.10 -> 27223.18 : +6553.08 = +31.703% (+/-1.56%)
misc_mandel.py 138221.11 -> 152014.01 : +13792.90 = +9.979% (+/-2.46%)
misc_pystone.py 85032.14 -> 105681.44 : +20649.30 = +24.284% (+/-2.25%)
misc_raytrace.py 19800.01 -> 23350.73 : +3550.72 = +17.933% (+/-2.79%)
In summary, compared to MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE, the new
MICROPY_OPT_LOAD_ATTR_FAST_PATH and MICROPY_OPT_MAP_LOOKUP_CACHE options:
- are simpler;
- take less code size;
- are faster (generally);
- work with code generated by the native emitter;
- can be used on embedded targets with a small and constant RAM overhead;
- allow the same .mpy bytecode to run on all targets.
See #7680 for further discussion. And see also #7653 for a discussion
about simplifying mpy-cross options.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
The existing inline bytecode caching optimisation, selected by
MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE, reserves an extra byte in the
bytecode after certain opcodes, which at runtime stores a map index of the
likely location of this field when looking up the qstr. This scheme is
incompatible with bytecode-in-ROM, and doesn't work with native generated
code. It also stores bytecode in .mpy files which is of a different format
to when the feature is disabled, making generation of .mpy files more
complex.
This commit provides an alternative optimisation via an approach that adds
a global cache for map offsets, then all mp_map_lookup operations use it.
It's less precise than bytecode caching, but allows the cache to be
independent and external to the bytecode that is executing. It also works
for the native emitter and adds a similar performance boost on top of the
gain already provided by the native emitter.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
When the LOAD_ATTR opcode is executed there are quite a few different cases
that have to be handled, but the common case is accessing a member on an
instance type. Typically, built-in types provide methods which is why this
is common.
Fortunately, for this specific case, if the member is found in the member
map then there's no further processing.
This optimisation does a relatively cheap check (type is instance) and then
forwards directly to the member map lookup, falling back to the regular
path if necessary.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This is the beginning of a set of changes to simplify enabling/disabling
features. The goals are:
- Remove redundancy from mpconfigport.h (never set a value to the default
-- make it clear exactly what's being enabled).
- Improve consistency between ports. All "similar" ports (i.e. approx same
flash size) should get the same features.
- Simplify mpconfigport.h -- just get default/sensible options for the size
of the port.
- Make it easy for defining constrained boards (e.g. STM32F0/L0), they can
just set a lower level.
This commit makes a step towards this and defines the "core" level as the
current default feature set, and a "minimal" level to turn off everything.
And a few placeholder levels are added for where the other ports will
roughly land.
This is a no-op change for all ports.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This commit simplifies and optimises the parse tree in-memory
representation of lists of expressions, for tuples and lists, and when
tuples are used on the left-hand-side of assignments and within del
statements. This reduces memory usage of the parse tree when such code is
compiled, and also reduces the size of the compiler.
For example, (1,) was previously the following parse tree:
expr_stmt(5) (n=2)
atom_paren(45) (n=1)
testlist_comp(146) (n=2)
int(1)
testlist_comp_3b(149) (n=1)
NULL
NULL
and with this commit is now:
expr_stmt(5) (n=2)
atom_paren(45) (n=1)
testlist_comp(146) (n=1)
int(1)
NULL
Similarly, (1, 2, 3) was previously:
expr_stmt(5) (n=2)
atom_paren(45) (n=1)
testlist_comp(146) (n=2)
int(1)
testlist_comp_3c(150) (n=2)
int(2)
int(3)
NULL
and is now:
expr_stmt(5) (n=2)
atom_paren(45) (n=1)
testlist_comp(146) (n=3)
int(1)
int(2)
int(3)
NULL
Signed-off-by: Damien George <damien@micropython.org>
The boot counter is a uint8_t single-byte counter stored in the first NVM byte position (`micrcontroller.nvm[0]`). The counter increments by 1 each time the board boots, regardless if it's a hard or soft reset.
Enable the boot counter by adding `#define CIRCUITPY_BOOT_COUNTER 1` to your board's mpconfigboard.h file. Note that an error will be thrown during the build if `CIRCUITPY_INTERNAL_NVM_SIZE` is not also set within mpconfigboard.h.
This commit refactors machine.PWM and creates extmod/machine_pwm.c. The
esp8266, esp32 and rp2 ports all use this and provide implementations of
the required PWM functionality. This helps to reduce code duplication and
keep the same Python API across ports.
This commit does not make any functional changes.
Signed-off-by: Damien George <damien@micropython.org>
The zephyr port doesn't support SoftI2C so it's not enabled, and the legacy
I2C constructor check can be removed.
Signed-off-by: Damien George <damien@micropython.org>
* Reduce the number of supported HID reports of IDs per descriptor.
This saves ~200 bytes in the default HID objects.
* (Not enabled) Compute QSTR attrs on init. This trades 1k RAM for
flash. Flash is the default (1).
Messages like 'command-line option is [not valid] for C++' can result
from the way the preprocessor is invoked by `genlast`. Instead, cause
the files to be preprocessed as though their content is "C". This
should generally be OK, as they'll eventually be _compiled_ as C++.
When preprocessed as C, the file simply needs to generate all the same
QSTRS and TRANSLATEs.
When you wish to use the valgrind memory analysis tool on micropython,
you can arrange to define MICROPY_DEBUG_VALGRIND to enable use of
special valgrind macros. For now, this only fixes `gc_get_ptr`
so that it never emits the diagnostic "Conditional jump or move depends
on uninitialised value(s)".
Signed-off-by: Jeff Epler <jepler@gmail.com>
Convert bitbangio, bitmaptools, _bleio, board, busio, countio, digitalio, framebufferio, frequencyio, gamepadshift, getpass, keypad, math, microcontroller, and msgpack modules to use MP_REGISTER_MODULE.
Related to #5183.
Convert adafruit_bus_device, adafruit_pixelbuf, analogio, atexit, audiobusio, audiocore, audioio, audiomixer, and audiomp3 modules to use MP_REGISTER_MODULE.
Related to #5183.
Convert from using MICROPY_PORT_BUILTIN_MODULES_STRONG_LINKS to using MP_REGISTER_MODULE for displayio, terminalio, and fontio modules.
Related to #5183.
This is a generic API for synchronously bit-banging data on a pin.
Initially this adds a single supported encoding, which supports controlling
WS2812 LEDs.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This was missed in 692d36d779. It's not
strictly necessary as the GC will clean it anyway, but it's good to
pre-emptively gc_free() all the blocks used in lexing/parsing.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This also tweaks the repr for unicode strings to only escape a few
utf-8 code points. This makes emoji show in os.listdir() for
example.
Also, enable exfat support on full builds.
Fixes#5146
This implements (most of) the PEP-498 spec for f-strings and is based on
https://github.com/micropython/micropython/pull/4998 by @klardotsh.
It is implemented in the lexer as a syntax translation to `str.format`:
f"{a}" --> "{}".format(a)
It also supports:
f"{a=}" --> "a={}".format(a)
This is done by extracting the arguments into a temporary vstr buffer,
then after the string has been tokenized, the lexer input queue is saved
and the contents of the temporary vstr buffer are injected into the lexer
instead.
There are four main limitations:
- raw f-strings (`fr` or `rf` prefixes) are not supported and will raise
`SyntaxError: raw f-strings are not supported`.
- literal concatenation of f-strings with adjacent strings will fail
"{}" f"{a}" --> "{}{}".format(a) (str.format will incorrectly use
the braces from the non-f-string)
f"{a}" f"{a}" --> "{}".format(a) "{}".format(a) (cannot concatenate)
- PEP-498 requires the full parser to understand the interpolated
argument, however because this entirely runs in the lexer it cannot
resolve nested braces in expressions like
f"{'}'}"
- The !r, !s, and !a conversions are not supported.
Includes tests and cpydiffs.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This leaves much more space on SAMD21 builds that aren't "full builds".
These are new APIs that we don't need to add to old boards.
Also, tweak two Arduino boards to save space on them.
* The new nonstandard '%S' format takes a pointer to compressed_string_t
and prints it
* The new mp_cprintf and mp_vcprintf take a format string that is a
compressed_string_t
By storing "count of words by length", the long `wends` table can be
replaced with a short `wlencount` table. This saves flash storage space.
Extend the range of string lengths that can be in the dictionary.
Originally it was to 2 to 9; at one point it was changed to 3 to 9.
Putting the lower bound back at 2 has a positive impact on the French
translation (a bunch of them, such as "ch", "\r\n", "%q", are used).
Increasing the maximum length gets 'mpossible', ' doit être ',
and 'CircuitPyth' at the long end. This adds a bit of processing time
to makeqstrdata. The specific 2/11 values are again empirical based on
the French translation on the adafruit_proxlight_trinkey_m0.
Commit 4173950658 removed automatic building
of mpy-cross, which rebuilt it whenever any of its dependent source files
changed.
But needing to build mpy-cross, and not knowing how, is a frequent issue.
This commit aims to help by automatically building mpy-cross only if it
doesn't exist. For Makefiles it uses an order-only prerequisite, while
for CMake it uses a custom command.
If MICROPY_MPYCROSS (which is what makemanifest.py uses to locate the
mpy-cross executable) is defined in the environment then automatic build
will not be attempted, allowing a way to prevent this auto-build if needed.
Thanks to Trammell Hudson aka @osresearch for the original idea; see #5760.
Signed-off-by: Damien George <damien@micropython.org>
Optionally enabled via MICROPY_PY_UJSON_SEPARATORS. Enabled by default.
For dump, make sure mp_get_stream_raise is called after
mod_ujson_separators since CPython does it in this order (if both
separators and stream are invalid, separators will raise an exception
first).
Add separators argument in the docs as well.
Signed-off-by: Peter Züger <zueger.peter@icloud.com>
Signed-off-by: Damien George <damien@micropython.org>
Commit e33bc597 ("py: Remove calls to file reader functions when these
are disabled.") changed the condition for one caller of
do_execute_raw_code() from
MICROPY_PERSISTENT_CODE_LOAD
to
MICROPY_HAS_FILE_READER && MICROPY_PERSISTENT_CODE_LOAD
The condition that enables compiling the function itself needs to be
changed to match.
Signed-off-by: David Lechner <david@pybricks.com>
Previously a subclass of a type that didn't implement unary_op, or didn't
handle MP_UNARY_OP_BOOL, would raise TypeError on bool conversion.
Fixes#5677.
This adds #if MICROPY_PY_USELECT_SELECT around the uselect.select()
function. According to the docs, this function is only for CPython
compatibility and should not normally be used. So we can disable it
and save a few bytes of flash space where possible.
Signed-off-by: David Lechner <david@pybricks.com>
The MP_OBJ_STOP_ITERATION optimisation is a shortcut for creating a
StopIteration() exception object, and means that heap memory does not need
to be allocated for the exception (in cases where it can be used). This
commit allows this optimised object to take an optional argument (before,
it could only have no argument).
The commit also adds some new tests to cover corner cases with
StopIteration and generators that previously did not work.
Signed-off-by: Damien George <damien@micropython.org>
I was puzzled by why the dictionary words were sorted by length.
It was because TextSplitter sorted its parameter, instead of a copy.
This doesn't affect encoding size, but does affect the encoding NUMBER
of the found words. We'll deliberately restore sorting by length next,
for other reasons, but not by spooky action.
Try to accurately measure the costs of including a word in the dictionary
vs the gains from using it in messages.
This saves about 160 bytes on trinket_m0 ja, the fullest translation
for that board. Other translations on the same board all have savings,
ranging from 24 to 228 bytes.
```
Translation Before After Savings
ja 1164 1324 160
de_DE 1260 1396 136
fr 1424 1652 228
zh_Latn_pinyin 1448 1520 72
pt_BR 1584 1736 152
pl 1592 1640 48
es 1724 1816 92
ko 1724 1816 92
fil 1764 1800 36
it_IT 1896 2040 144
nl 1956 2136 180
ID 2072 2180 108
cs 2124 2148 24
sv 2340 2448 108
en_x_pirate 2644 2740 96
en_GB 2652 2752 100
el 2656 2768 112
en_US 2656 2768 112
hi 2656 2768 112
```
By comparing the address of the initial 'name' field instead of the
addresses of the objects themselves, a small amount of type safety is
added back, vs just casting to void.
In the event that some other kind of object is passed in as 't',
which happens to have a 'name' field of the right type, the construct
would be (undesirably) accepted but it would almost certainly evaluate
to false at runtime.
This adds the --tags argument to the git describe command that is used
to define the MICROPY_GIT_TAG macro. This makes it match non-annotated
tags. This is useful for MicroPython derivatives that don't use
annotated tags.
Signed-off-by: David Lechner <david@pybricks.com>
Only include .c and .cpp files explicitly in the list of files passed to
the preprocessor for QSTR extraction. All relevant .h files will be
included in this process by "#include" from the .c(pp) files. In
particular for moduledefs.h, this is included by py/objmodule.c (and
doesn't actually contain any extractable MP_QSTR_xxx, but rather defines
macros with MP_QSTR_xxx's in them which are then part of py/objmodule.c).
The main reason for this change is to simplify the preprocessing step on
the javascript port, which tries to compile .h files as C++ precompiled
headers if they are passed with -E to clang.
Signed-off-by: Damien George <damien@micropython.org>
Fixes the following (the line numbers match commit 0e87459e2b):
../../extmod/crypto-algorithms/sha256.c:49:19: runtime error: left shif...
../../extmod/moduasyncio.c:106:35: runtime error: member access within ...
../../py/binary.c:210:13: runtime error: left shift of negative value -...
../../py/mpz.c:744:16: runtime error: negation of -9223372036854775808 ...
../../py/objint.c:109:22: runtime error: left shift of 1 by 31 places c...
../../py/objint_mpz.c:374:9: runtime error: left shift of 4611686018427...
../../py/objint_mpz.c:374:9: runtime error: left shift of negative valu...
../../py/parsenum.c:106:14: runtime error: left shift of 46116860184273...
../../py/runtime.c:395:33: runtime error: left shift of negative value ...
../../py/showbc.c:177:28: runtime error: left shift of negative value -...
../../py/vm.c:321:36: runtime error: left shift of negative value -1```
Testing was done on an amd64 Debian Buster system using gcc-8.3 and these
settings:
CFLAGS += -g3 -Og -fsanitize=undefined
LDFLAGS += -fsanitize=undefined
The introduced TASK_PAIRHEAP macro's conditional (x ? &x->i : NULL)
assembles (under amd64 gcc 8.3 -Os) to the same as &x->i, since i is the
initial field of the struct. However, for the purposes of undefined
behavior analysis the conditional is needed.
Signed-off-by: Jeff Epler <jepler@gmail.com>
This introduces a new macro to get the main thread and uses it to ensure
that asynchronous exceptions such as KeyboardInterrupt (CTRL+C) are only
scheduled on the main thread. This is more deterministic than being
scheduled on a random thread and is more in line with CPython that only
allow signal handlers to run on the main thread.
Fixes issue #7026.
Signed-off-by: David Lechner <david@pybricks.com>
This moves mp_pending_exception from mp_state_vm_t to mp_state_thread_t.
This allows exceptions to be scheduled on a specific thread.
Signed-off-by: David Lechner <david@pybricks.com>
Prior to this commit, cache flushing for ARM native code was done only in
the assembler code asm_thumb_end_pass()/asm_arm_end_pass(), at the last
pass of the assembler. But this misses flushing the cache when loading
native code from an .mpy file, ie in persistentcode.c.
The change here makes sure the cache is always flushed/cleaned/invalidated
when assigning native code on ARM architectures.
This problem was found running tests/micropython/import_mpy_native_gc.py on
the mimxrt port.
Signed-off-by: Damien George <damien@micropython.org>
Damien George
Dan Halbert
Jeff Epler
microDev
Scott Shawcroft
Dan Halbert
Laurens Valk
Jim Mussared
Kattni Rembor
Mike Wadsten
stijn
Jeff Epler
iabdalkader
Max Holliday
Scott Shawcroft
Rob Capellini
Jeff Epler
root
Peter Züger
David Lechner
Bryan Tong Minh
Yonatan Goldschmidt
George Waters