GitLab

I have kept the original APPS size on purpose as 40KB SPIFFS is useless IMHO

Some fixes to nghttp2 to provide basic functionality. Added the missing main header. Removed the asio headers which require Boost libraries. Moved http_parser into the expected location. (#2068)

wait for FSM to return idle

* Added ESP:: functions for sketch size

* Fixed free space name to match ESP8266

* Working example for HTTPS over Eduroam network

* Update WiFiClientSecureEnterprise.ino

Added Olimex board ESP32-PoE. pins_arduino header file based on the ESP32-EVB with changed SS pin and removed BOARD_HAS_1BIT_SDMMC macro.

* updated WifiClientEnterprise example

* Update WiFiClientEnterprise.ino

* Create pins_arduino.h

* Update boards.txt

Pycom LoPy + LoPy4 boards added

* Create pins_arduino.h

* Update pins_arduino.h

bugfix antenna switch GPIO port 16 -> 21

* Update pins_arduino.h

* Update pins_arduino.h

* Update pins_arduino.h

* Update pins_arduino.h

* Update pins_arduino.h

* Update boards.txt

* Updated table of tested locations

* updated

* Added HTTPUpdate class for downloading sketches from a server

* Added HTTPUpdate class for downloading sketches from a server

* Added HTTPUpdate to CMakeLists.txt

* Change ESP8266 class references to ESP32 for httpUpdate.ino example

* Change ESP8266 class references to ESP32 for httpUpdate.ino example. setLedPin() commented out because not all boards support LED_BUITLIN

* Added check to handle mixup of old and present api properly

* Correct HTTPClient::setTimeout() to convert milliseconds to seconds. Correct WiFiClient::setTimeout() to call Stream::setTimeout() with seconds converted back to milliseconds. Remove inproper checks for _insecure.

* Added small comment because it looked like the Travis build did not finish

* Pass client parameter into two new begin() functions. Set other begin() functions deprecated. Updated library version to 1.2

* Added working HTTPS example on a public url with a certificate

* Remove two unnecessary tests in ::disconnect()

* Add a scoping block to BasicHttpsClient.ino to assure HTTPClient is destroyed before WiFiClientSecure

* Added check to handle mixup of old and present api properly

* Correct HTTPClient::setTimeout() to convert milliseconds to seconds. Correct WiFiClient::setTimeout() to call Stream::setTimeout() with seconds converted back to milliseconds. Remove inproper checks for _insecure.

* Added small comment because it looked like the Travis build did not finish

* Added baudrate detection to esp32-hal-uart and HardwareSerial

* Solved compiler warning for uartResizeRxBuffer()

* Add unit to header variable name (timeout_ms)

* Reverting accidentally changed files to master

* Add small delay after baudrate detection

Pin definitions T8/T9 & DAC1/DAC2 Back To Front (Issue #1831)

Pin definitions T8/T9 & DAC1/DAC2 Back To Front

* make WiFi.softAP() more robust

* WiFi.softAP() revert fallback to WIFI_AUTH_OPEN

fixes #1875

Incorrect values used, use the proper defines from header-files instead.

* removed uint8_t Wire.endTransmission(uint8_t sendStop)

Having both endTransmission(bool) and endTransmission(uint8_t) creates problems.
There is no need for endTransmission(uint8_t)
endTransmission(1) and endTransmission(0) works with endTransmission(bool).
Removing endTransmission(uint8_t) allows the ESP32 code to be compatible with
all the other Arduino cores by allowing endTransmission(1) and endTranmission(0)
to work as it does on all the other cores.

* Wire library version bump for endTransmission() update

#1869 exposed a resource exhaustion issue. The current HAL layer for I2C support is designed to use a shared interrupt, But, during debugging to solve the interrupt overloading condition identified in #1588, and the generation of pr #1717, the interrupt allocation parameters were changed.  This change was unnecessary, the code will work successfully with shared interrupts.  So, there is no need to assign a private interrupt for each I2C peripheral.

* TX Flow Control and Code cleanup

* Use semaphore instead of delay

TX functionality is done.

* Use single buffer and empty queue on exit

* Fix compile issues because of LwIP code relocation

* Add temporary header to fix Azure not compiling

* Fix AsyncUDP early init

* AsyncUDP Multicast fixes

* Add source mac address and rework multicast

* Allow redefinition of default pins for Serials 1 and 2

* Update IDF to 3276a13

* Update esptool.py to 2.5.0

* Fix sketches

* Fix log level in BluetoothSetial

* rmt driver initial version

* supporting conti mode plus interrupts

* using conitnous mode for sending more data

* working continous mode

* rmt driver cleanup after conti mode

* initial version of rmt driver

* adding a simple example

* adding channel and block locks

* modified of rmt interface for simpler/easier usage

* adding header sentinels, split interface to common and additional settings

* Fixes per code review + support for rx callback mode

* renamed internal structures and enums, fixed formatting

* cmake support for rmt

* refactored tx-conti interrupts to function to make it more readable

* added Tx and Rx examples

* added license headers

* minor updates per review

* used struct access, renamed defines, corrected diagram

* Initial

* Implementation

* Add to CMakelist.txt

* Add example

* Add IRAM_ATTR

* First commit of FFat library

* Fixed reboot loops if no fat present. Added CMakeLists

* Functionalize the partition checks

* Cleanup, especially in format

* Dont format if mounted.  More wording cleanup

* 16M ffat should only be on 16M board

* Fix some casting issues that trip up the compiler when building as ESP-IDF component

* Add BPI-BIT boards connfig

* Add BPI-BIT v1.3 boards

* Add wESP32 support

* Add default Ethernet config

Few testers let me know, they were unable with connection to Eduroam network under PEAP+MsCHAPv2 methods.
They were trying many modifications. Solution is very absurd. 
Change board to STA mode manually. 
One tester from Italy was able to connect to Eduroam network under TTLS + MsCHAPv2 with that sketch too!

Sketch was tested and worked almost for all testers with that problem.