Skip to content

Changelog

What changed in each release of the driver. Entries under Breaking need source changes in code that used the affected API.

[Unreleased]

  • Review round 1: DisplayOff() is never refused; vertical scroll needs 64 MUX in SetMultiplexRatio() and ApplyConfig(); per-variant supply and pin documentation; the ESP32 example uses SPI DMA; the provenance of the initialisation values is stated as inherited and derived.
  • Review round 4: every wait from a configuration time is clamped to its limit (the stop path runs on an unchecked configuration too); SetGrayScaleTable() no longer clears linear_gray_scale, and SetPrechargeSpeed() sets send_precharge_speeds, so a restart is documented and consistent; the debouncer is reset only when the switch settings change; a raw 2Fh/2Eh keeps IsScrolling() in step and a raw ADh takes only 8Eh; the ESP32 example restarts only after a controller transfer failed; the compile-fail tests share a resource lock.
  • Review round 3: the stored configuration stays valid in every setter order (SetDimSettings() also checks the contrast in force; a failed ApplyConfig() leaves the dim limit); ApplyConfig() returns ScrollActive while scrolling (Breaking: it was sent); Poll() delivers a press on a locked controller without waking it; a raw FDh takes only 12h or 16h; the three second pre-charge speeds are pinned to one session by a test.
  • NKK application notes pass: the start-up values are checked against Tables 1 to 3 (steps 4.1 to 4.18) by tests/test_nkk_tables.cpp; the ISC15ANP4 and ISC01P presets no longer send the second pre-charge speeds (not in Tables 1 and 2: Config::send_precharge_speeds); the ISF15ACP4 sends BBh 1Dh as Table 3 does (Config::precharge_bit0); a dim contrast above the start-up contrast is refused by Validate() and SetDimSettings(), and SetContrast() refuses a contrast below the dim contrast (Breaking for a caller that set a contrast below the dim values); the controller is identified from the notes; the supply, charge-pump enable, scrolling and CS# rules of the notes are in the docs.
  • Round 2: CopyWindow() takes its destination as a Point (six parameters); the start-up pin failures, the stop path and the idle policy errors are pinned by tests; the model requires the whole start-up sequence before VCC.

Added

  • ISF15ACP4 driver, first version: a 96 x 64 colour OLED on write-only SPI and a debounced switch.
  • Power sequence: the OLED supply is never enabled before the controller is reset and configured; the power-down always runs.
  • The SSD1331 command set: display on, dim and off, remap and colour depth, contrast, master current, pre-charge, VCOMH, gray scale, pixel windows in three depths, the drawing engine, scrolling, the command lock.
  • Switch: debounce, press, release, long press and repeat; idle dim and sleep with wake on a press.
  • Presets for the ISC15ANP4 and ISC01P variants.
  • Tests against an independent model of the controller with fault injection, mutation checks of the safety logic, a fuzz target, and compile-fail tests of the port contract.
  • Not verified on hardware; see docs/limitations.md.