forgot to adjust the debug condition in my previous commit.
NB: the condition only shows a debug message when the max wait time was exceeded, which can only happen when line 1692 has waited for the maximum allowed time. ->Is this intended?
the timing logic did not work in case that millis()+100 + frametime rolls over; in this case millis() > maxWait, and waiting would be skipped which might lead to crashes.
-> logic slightly adjusted to be robust against rollover.
* improvements to Aurora FX
- converted to integer math, increasing speed on all ESPs, also shrinks code size
- caching values to avoid repeated calculations
- CRGBW instead or CRGB, adds white channel support when not using palette
- fix for new brightness/gamma handling
* overflow & unsigned fix
- add better support for 1D gifs: use the full gif, row by row, scale if needed
- add blur slider to image FX
- improved safety checks to avoid crashes
- add "fast path" if image size matches virtual segment size
* catch some error that would lead to undefined behavior
* additional debug messages in case of errors
* robustness: handle OOM exception from decoder.alloc() gracefully
- no mendatory external JS dependency, works in offline mode
- optional external dependency is used for highlighting JSON, plain text edit is used if not available
- WLED styling (dark mode only)
- JSON files are displayed "prettyfied" and saved "minified"
- JSON color highlighting (if available)
- JSON verification during edit and on saving both in online and offline mode
- special treatment for ledmap files: displayed in aligned columns (2D) or as lines (1D), saved as minified json: no more white-space problems
- displays file size and total flash usage
Add a field to the OTA metadata structure indicating the oldest base
version it's safe to install this update /from/. This provides a clear
path forward in case there are incompatibilities, eg. some case
(bootloader compatibility) where 0.16.0 cannot be installed safely from
0.15.2, but a transitional 0.15.3 can arrange the groundwork.
- Enabling DDP over WebSocket: this allows for UI or html tools to stream data to the LEDs much faster than through the JSON API.
- first byte of data array is used to determine protocol for future use
- Moved the duplicate function to establish a WS connection from the live-view htm files to common.js
- add better safety check for DDP: prevent OOB reads of buffer
during testing at low brightness I noticed that gradients can be "jumping" in colors quite wildly, turning a smooth gradient into a flickering mess. This is due to the color hue preservation being inaccurate and a bit too aggressive. This can be seen for example using a gradient palette and "Running" FX.
Removing the hue preservation completely fixes it but leaves color artefacts for example visible in PS Fire at very low brightness: the bright part of the flames gets a pink hue. This change is a compromise to fix both problems to a "good enough" state
- since we draw on a segment, we need to use virtual segment dimensions or scaling will be off when using any virtualisation like grouping/spacing/mirror etc.
- Enabling DDP over WebSocket: this allows for UI or html tools to stream data to the LEDs much faster than through the JSON API.
- first byte of data array is used to determine protocol for future use
- Moved the duplicate function to establish a WS connection from the live-view htm files to common.js
- add better safety check for DDP: prevent OOB reads of buffer
during testing at low brightness I noticed that gradients can be "jumping" in colors quite wildly, turning a smooth gradient into a flickering mess. This is due to the color hue preservation being inaccurate and a bit too aggressive. This can be seen for example using a gradient palette and "Running" FX.
Removing the hue preservation completely fixes it but leaves color artefacts for example visible in PS Fire at very low brightness: the bright part of the flames gets a pink hue. This change is a compromise to fix both problems to a "good enough" state
Improves cache utilization as fewer things are passed via CFLAGS to
all files. In the event that no metadata is available, let the cpp
file handle warning about default usage.
Implement a comprehensive solution for validating a firmware before an
OTA updated is committed. WLED metadata such as version and release
is moved to a data structure located at near the start of the firmware
binary, where it can be identified and validated.
Co-authored-by: netmindz <442066+netmindz@users.noreply.github.com>
- if a segment is destroyed or turned inactive, disable the gif player: only one gif player instance can run at a time, if a inactive or destroyed segment uses it, the effect is broken for other segments.
- copy FX ironically copied the source segment on each call, should use reference not a copy!
Sends a shimmer across the strip at defined (or random) intervals
Optional brightness modulators: sine or perlin noise
Can be used as an overlay to other effects.
- speed optimization in color_add, PS fast_color_add and blur functions
- applying more bit and shift manipulation tricks to squeeze out just a bit more speed on color manipulation functions.
- Optimization on blur is based on work by @blazoncek
- Renamed PS fast_color_add() to fast_color_scaleAdd()
- allow more than 10 custom palettes
- move palettes into CPP file
- Fix for minimizing cpal.htm (saves 2k of flash)
- shortened names in cpal, saves about 400 bytes of Flash after packing
- removed async from common.js loading to prevent errors on page loads if the file is not cached
- restricted nubmer of user palettes on ESP8266 to 10
- unrestricted number of user palettes on all other platforms (total max palettes: 256)
- added a warning when adding more than 10 palettes to let the user decide to risk it
- Bugfixes in palette enumeration, fixed AR palette adding
- AR palettes are now also added if there are more than 10 custom palettes
Co-authored-by: Blaž Kristan <blaz@kristan-sp.si>
* adding center bin selection to 2D GEQ: this makes it possible to use subsets of the GEQ on distributed strips
setting custom3 to 0 gives the "old" behaviour, this is the default. existing presets will have the custom3 slider at the center, changing presets that do not use the full width so this is a breaking change for those but I assume theser are rare.
* Attempt at better bus memory calculation and estimation
* Remov double buffer count for ESP8266 (thanks @dedehai)
* improve UI calculation
* adding mendatory LED buffers to UI memory calculation
* adding buffer for transitions to memory calculation, change "error" to "warning"
* bugfixes in settings_leds.htm
* fix getDataSize() forESP8266, ESP8266 does not use double buffering.
* update led settings: fix parsing by @blazoncek
* new warnings for LED buffer use, deny adding a bus if it exceeds limits
* adds recommendations for users (reboot, disable transitions)
Co-authored-by: Blaž Kristan <blaz@kristan-sp.si>
* Improved heap and PSRAM handling
- Segment `allocateData()` uses more elaborate DRAM checking to reduce fragmentation and allow for larger setups to run on low heap
- Segment data allocation fails if minimum contiguous block size runs low to keep the UI working
- Increased `MAX_SEGMENT_DATA` to account for better segment data handling
- Memory allocation functions try to keep enough DRAM for segment data
- Added constant `PSRAM_THRESHOLD` to improve PSARM usage
- Increase MIN_HEAP_SIZE to reduce risk of breaking UI due to low memory for JSON response
- ESP32 makes use of IRAM (no 8bit access) for pixeluffers, freeing up to 50kB of RAM
- Fix to properly get available heap on all platforms: added function `getFreeHeapSize()`
- Bugfix for effects that divide by SEGLEN: don't run FX in service() if segment is not active
-Syntax fix in AR: calloc() uses (numelements, size) as arguments
* Added new functions for allocation and heap checking
- added `allocate_buffer()` function that can be used to allocate large buffers: takes parameters to set preferred ram location, including 32bit accessible RAM on ESP32. Returns null if heap runs low or switches to PSRAM
- getFreeHeapSize() and getContiguousFreeHeap() helper functions for all platforms to correctly report free useable heap
- updated some constants
- updated segment data allocation to free the data if it is large
- replaced "psramsafe" variable with it's #ifdef: BOARD_HAS_PSRAM and made accomodating changes
- added some compile-time checks to handle invalid env. definitions
- updated all allocation functions and some of the logic behind them
- added use of fast RTC-Memory where available
- increased MIN_HEAP_SIZE for all systems (improved stability in tests)
- updated memory calculation in web-UI to account for required segment buffer
- added UI alerts if buffer allocation fails
- made getUsedSegmentData() non-private (used in buffer alloc function)
- changed MAX_SEGMENT_DATA
- added more detailed memory log to DEBUG output
- added debug output to buffer alloc function
Pull the RMT High-priority Interrupt driver in to a vendored local
library, pending inclusion in upstream NeoPixelBus.
Driver is enabled only for XTensa chips; there's some unresolved
issue with nested interrupts on RISCV.
* updated color scaling to preserve hue at low brightness resulting in much better colors
* replace NPBlg with NPB, moved brightness scaling to bus manager
* improved gamma table calculation: fixed mismatch in inverting gamma table calculation: inversion should now be as good as it gets
* code cleanup, fixed gamma being applied in unnecessary places
Improvements to ABL handling:
- removed strip level handling, ist now all done on bus level
- limiter now respects pixel mapping
- proper handling of white channel
- improved current estimation
- current is now always correctly reported to UI
- minimal FPS impact if the ABL is not limiting but slighly higher impact for global ABL limit due to double-scaling
- moved brightness scaling to BusDigital
- created new header file colors.h to be able to access color functions in bus-manager.
- updated colo_fade() with better video scaling to preserve hue's at low brightness
- added IRAM_ATTR to color_fade (negligible speed impact when compared to inline and benefits other functions)
- added IRAM_ATTR to color_blend as it is used a lot throughout the code, did not test speed impact but adding it to color_fade made it almost on-par with an inlined function
Additional changes:
- fixes for properly handling `scaledBri()` (by @blazoncek)
- also use bit-shift instead of division in blending for ESP8266
- improvements for faster "softlight" calculation in blending
- changed some variables to uint8_t to maybe let the compiler optimize better, uint8_t can be faster if read, store and set are all done in uint8_t, which is the case in the ones I changed
- various minor code formatting changes
Any repeating crash that prevents a human from logging in and fixing
the config should be treated as a boot loop. Increase the detection
timeout, so anything that's fast enough to preclude a user fix will
trigger the recovery behaviour.
Don't treat consecutive but infrequent crashes as bootloops. The
bootloop recovery actions only make sense when there is no opportunity
for a user to reconfigure their system.
Suggested by @coderabbitai
this fixes a very long loop when an overflow was happening in palette blending.
- reset prevPaletteBlends to prevent overflow
- add safety check in case overflow should still happen in another combination (or in future changes)
