Merge pull request #54 from bxparks/develop

merge 1.5.1 into master

Author: bxparks

CHANGELOG.md

@@ -1,6 +1,15 @@
 # Changelog
 
 * Unreleased
+* 1.5.1 (2022-09-20)
+    * Add Adafruit nRF52 boards to "Tier 2" after validation by community
+      member.
+    * Upgrade tool chain
+        * Arduino CLI from 0.19.2 to 0.27.1
+        * Arduino AVR Core from 1.8.4 to 1.8.5
+        * STM32duino from 2.2.0 to 2.3.0
+        * ESP32 Core from 2.0.2 to 2.0.5
+        * Teensyduino from 1.56 to 1.57
 * 1.5.0 (2022-03-19)
     * (Re)add support for human-readable coroutine names.
         * See [Coroutine Names](USER_GUIDE.md#CoroutineNames) in the

README.md

@@ -120,7 +120,7 @@ AceRoutine is a self-contained library that works on any platform supporting the
 Arduino API (AVR, Teensy, ESP8266, ESP32, etc), and it provides a handful of
 additional macros that can reduce boilerplate code.
 
-**Version**: 1.5.0 (2022-03-19)
+**Version**: 1.5.1 (2022-09-20)
 
 **Changelog**: [CHANGELOG.md](CHANGELOG.md)
 
@@ -510,7 +510,7 @@ The development version can be installed by cloning the following git repos:
 
 You can copy these directories to the `./libraries` directory used by the
 Arduino IDE. (You should see 2 directories,  named `./libraries/AceRoutine` and
-`./libraries/AceCommon). Or you can create symlinks from `/.libraries` to these
+`./libraries/AceCommon`). Or you can create symlinks from `/.libraries` to these
 directories.
 
 The `develop` branch contains the latest working version.
@@ -834,13 +834,13 @@ Arduino Nano:
 +---------------------------------+--------+-------------+--------+
 | Functionality                   |  iters | micros/iter |   diff |
 |---------------------------------+--------+-------------+--------|
-| EmptyLoop                       |  10000 |       1.900 |  0.000 |
+| EmptyLoop                       |  10000 |       1.700 |  0.000 |
 |---------------------------------+--------+-------------+--------|
-| DirectScheduling                |  10000 |       2.800 |  0.900 |
-| DirectSchedulingWithProfiler    |  10000 |       5.800 |  3.900 |
+| DirectScheduling                |  10000 |       2.900 |  1.200 |
+| DirectSchedulingWithProfiler    |  10000 |       5.700 |  4.000 |
 |---------------------------------+--------+-------------+--------|
-| CoroutineScheduling             |  10000 |       7.000 |  5.100 |
-| CoroutineSchedulingWithProfiler |  10000 |       9.300 |  7.400 |
+| CoroutineScheduling             |  10000 |       7.100 |  5.400 |
+| CoroutineSchedulingWithProfiler |  10000 |       9.300 |  7.600 |
 +---------------------------------+--------+-------------+--------+
 ```
 
@@ -850,13 +850,13 @@ ESP8266:
 +---------------------------------+--------+-------------+--------+
 | Functionality                   |  iters | micros/iter |   diff |
 |---------------------------------+--------+-------------+--------|
-| EmptyLoop                       |  10000 |       0.100 |  0.000 |
+| EmptyLoop                       |  10000 |       0.200 |  0.000 |
 |---------------------------------+--------+-------------+--------|
-| DirectScheduling                |  10000 |       0.500 |  0.400 |
-| DirectSchedulingWithProfiler    |  10000 |       0.800 |  0.700 |
+| DirectScheduling                |  10000 |       0.500 |  0.300 |
+| DirectSchedulingWithProfiler    |  10000 |       0.800 |  0.600 |
 |---------------------------------+--------+-------------+--------|
-| CoroutineScheduling             |  10000 |       0.900 |  0.800 |
-| CoroutineSchedulingWithProfiler |  10000 |       1.100 |  1.000 |
+| CoroutineScheduling             |  10000 |       0.900 |  0.700 |
+| CoroutineSchedulingWithProfiler |  10000 |       1.100 |  0.900 |
 +---------------------------------+--------+-------------+--------+
 ```
 
@@ -880,12 +880,16 @@ These boards are tested on each release:
 
 **Tier 2: Should work**
 
-These boards should work but I don't test them as often:
+These boards should work, but they are not tested frequently by me, or I don't
+own the specific hardware so they were tested by a community member:
 
 * ATtiny85 (8 MHz ATtiny85)
 * Arduino Pro Mini (16 MHz ATmega328P)
 * Mini Mega 2560 (Arduino Mega 2560 compatible, 16 MHz ATmega2560)
 * Teensy LC (48 MHz ARM Cortex-M0+)
+* [Adafruit nRF52 Boards](https://github.com/adafruit/Adafruit_nRF52_Arduino)
+    * [Circuit Playground Bluefruit](https://www.adafruit.com/product/4333)
+      tested by a community member
 
 **Tier 3: May work, but not supported**
 
@@ -924,6 +928,7 @@ This library was developed and tested using:
 * [ESP8266 Arduino 3.0.2](https://github.com/esp8266/Arduino)
 * [ESP32 Arduino 2.0.2](https://github.com/espressif/arduino-esp32)
 * [Teensyduino 1.56](https://www.pjrc.com/teensy/td_download.html)
+* [Adafruit nRF52 1.3.0](https://github.com/adafruit/Adafruit_nRF52_Arduino)
 
 This library is *not* compatible with:
 

USER_GUIDE.md

@@ -4,7 +4,7 @@ See the [README.md](README.md) for installation instructions and other
 background information. This document describes how to use the library once it
 is installed.
 
-**Version**: 1.5.0 (2022-03-19)
+**Version**: 1.5.1 (2022-09-20)
 
 ## Table of Contents
 
@@ -1153,22 +1153,25 @@ about 14 bytes of flash per invocation.
 
 Version 1.5 added the ability to profile the execution time of
 `Coroutine::runCoroutine()` and render the information as a formatted table, or
-as a JSON object. If the profiling feature is not used, no additional flash
-memory is consumed. The static RAM usage does increase by 2 bytes (8-bits) and 4
-bytes (32-bits) per coroutine.
+as a JSON object. (Thanks to peufeu2@ who proposed the idea and provided the
+initial proof of concept in
+[Discussion#50](https://github.com/bxparks/AceRoutine/discussions/50)).
+
+If the profiling feature is not used, no additional flash memory is consumed.
+The static RAM usage does increase by 2 bytes (8-bits) and 4 bytes (32-bits) per
+coroutine even if this feature is not used. The feature seemed useful enough to
+accept this small increase in static memory size, because most applications will
+not use more than 5-10 coroutines, and that translates into only 10-20 bytes of
+additional static RAM usage on 8-bit processors.
 
 The following classes and API methods were added to support the profiling
-feature. The `CoroutineProfiler` class is an interface that allows an object to
-receive information about the execution time of the `Coroutine::runCoroutine()`
-method:
+feature. The `CoroutineProfiler` class is an interface whose
+`updateElapsedMicros()` should be called with the execution time of the
+`Coroutine::runCoroutine()` method:
 
 ```C++
 class CoroutineProfiler {
   public:
-    /**
-     * Process the completion of the runCoroutine() method which took
-     * `micros` microseconds.
-     */
     virtual void updateElapsedMicros(uint32_t micros) = 0;
 };
 ```
@@ -1188,14 +1191,17 @@ class Coroutine {
 };
 ```
 
+The `runCoroutineWithProfiler()` method calls `runCoroutine()`, measures the
+elapsed time in microseconds, then calls the `profiler->updateElapsedMicros()`.
+
 **Note**: When creating Coroutines with profiling enabled, it will probably be
 necessary to assign human-readable names to each coroutine for identification
 purposes. See [Coroutine Names](#CoroutineNames) for information on the
 `setName()`, `getCName()`, `getFName()`, `getNameType()`, and `printNameTo()`
 methods. Each coroutine name will consume additional flash memory.
 
 Currently only a single implementation of `CoroutineProfiler` is provided, the
-`LogBinProfiler`. It contains 32 bins of `uint16_t` which tracks the number of
+`LogBinProfiler`. It contains 32 bins of `uint16_t` which track the number of
 times a `micros` was seen. The bins are logarithmically scaled, so that Bin 0
 collects all events `<2us`, Bin 1 collects events `<4us`, Bin 2 collects events
 `<8us`, ..., Bin 30 collects events `<2147s`, and the last Bin 31 collects
@@ -1242,7 +1248,7 @@ COROUTINE(coroutine1) {
   ...
 }
 
-COROUTINE(coroutine1) {
+COROUTINE(coroutine2) {
   ...
 }
 
@@ -1275,10 +1281,12 @@ COROUTINE(coroutine1) {
   ...
 }
 
-COROUTINE(coroutine1) {
+COROUTINE(coroutine2) {
   ...
 }
 
+...
+
 void setup() {
   ...
   LogBinProfiler::createProfilers();
@@ -1391,7 +1399,7 @@ class LogBinJsonRenderer{
 
 * The `printer` is usually the `Serial` object, but can be changed to something
   else if needed.
-* The `startBin` (0-31) and `endBin` (0-32) identify the bins which should be
+* The `startBin` [0-31] and `endBin` [0-32] identify the bins which should be
   printed.
     * A range of something like [2, 13) is useful to keep the width of the table
       reasonable.
@@ -1423,6 +1431,9 @@ this:
 }
 ```
 
+The `LogBinProfiler` uses a `uint16_t` counter, so the maximum value is
+saturated to `65535`.
+
 <a name="ProfilerResourceConsumption"></a>
 ### Profiler Resource Consumption
 
@@ -1459,13 +1470,13 @@ If the profiling feature is enabled, the
 
 The [AutoBenchmark](examples/AutoBenchmark) program shows that calling the
 profiler-enabled methods, `Coroutine::runCoroutineWithProfiler()` and
-`CoroutineScheduler::loopWithProfiler(), increases latency by:
+`CoroutineScheduler::loopWithProfiler()`, increases latency by:
 
-* 3 - 3.2 micros on AVR
+* 2.2 - 3.0 micros on AVR
 * 0.4 micros on STM32F1
 * 0.2 - 0.3 micros on ESP8266
 * 0.1 micros on ESP32
-* 0.033 - 0.166 micros on Teensy 3.2
+* 0.03 - 0.17 micros on Teensy 3.2
 
 On 32-bit processors, the overhead seems neglegible. On 8-bit processors, the 3
 microsecond of overhead might be an issue with sensitive applications.