when SERIAL_RX_BUFFER_SIZE is defined as greater than 256, the current HardwareSerial.cpp code almost works correctly. but not quite. the function serial.available may return an incorrect number of characters waiting (though when treated purely as a boolean it seems not to fault).serial.peek has a similar issue detecting if a character is waiting, while serial.read may also leave _rx_buffer_tail corrupt to an interrupt occurring mid-update (this corruption may cause the ISR to be mistaken about how much free is available in the Rx buffer).
the proposed changes to these three functions add cli/sti pairings around the critical pieces of code. i have tested and verified the change made to serial.available as correcting the problem, while the change to serial.peek follows the exact same pattern. the changes to serial.read have been confirmed as not breaking the function, but i have not been in a position to test the failure of the original (non-atomic) version to make comparisons.
multiple testing was conducted at 115,200 baud and 500,000 baud using data streams of 1,000,000 characters sent over a 2 minute interval. any error causing loss of character would have resulted in a catastrophic (ie, very obvious) failure.
cheers,
rob :-)
-> Squash and rebase of https://github.com/arduino/Arduino/pull/3848
These functions were changed from private to protected in 99f2a27553 but the comments were not updated at that time.
In conjunction with equivalent pull requests to Arduino SAM Boards and Arduino SAMD Boards, solves https://github.com/arduino/Arduino/issues/6146.
This method originally flushed pending input bytes, which makes sense in
Stream. At some point it was changed to flush output bytes instead, but
it was never moved to Print to reflect this.
Since Stream inherits from Print, this should not really affect any
users of the Stream or Print classes. However to prevent problems with
existing implementations of the Print class that do not provide a
flush() implementation, a default implementation is provided. We should
probably remove this at some point in the future, though.
- "." is a string literal, and so is treated as the char '.' plus the null char '\0'.
- Single quotes reduces the necessary memory for this literal to only one char instead.
A string literal as the one actually present may require the use of the
method "write(const char *str)", so there could be also a performance overhead.
- Another reason to change quotes style is for consistency with line 235.
Signed-off-by: Patrick Roncagliolo <ronca.pat@gmail.com>
On Linux, setting autosuspend_delay_ms to N and control to auto allows the host pc to suspend the peripheral. Some Linux distro (Ubuntu, Mint) apply this behaviour by default.
If the sketch's prints where less frequent than N milliseconds the sketch prints would never arrive.
This patch allows sending a remote wakeup event to unsuspend the peripheral and allow the serial prints to be received.
to allow WebUSB development, provide a way to change the USB_VERSION reported
using an additional core.
The additional (webUSB) core will survive IDE and AVR core updates
In an excess of confidence, these defines were added to Leonardo's variant.h
3rd party boards sometimes avoid inheriting this variant but they still define USBCON, thus breaking the build
As I was not able to base the return types of `begin()` & `end()` off the c_str() function, I have changed the source so the features can be used by C++98 code, while still allowing ranged loops in C++11.
Replaces #4280, only checks for the bootloader once
Tested with Hoodloader2, should work with every LUFA-derived bootloader released after 2014 (.apitable_signatures section must be placed at end of the flash)
BootloaderAPITable.S :
.global BootloaderAPI_Signatures
BootloaderAPI_Signatures:
.long BOOT_START_ADDR ; Start address of the bootloader
.word 0xDF00 ; Signature for the CDC class bootloader
.word 0xDCFB ; Signature for a LUFA class bootloader
makefile:
BOOT_API_LD_FLAGS += $(call BOOT_SECTION_LD_FLAG, .apitable_signatures, BootloaderAPI_Signatures, 8)
Avoid using the overload of print() for signed integer since a negative value is not allowed here.
This results in a smaller (unless print(int) is used somewhere else in the program) and faster code because the overload for unsigned integer is simpler.
The current code is very verbose and a painful to maintain (Change ISR implementation in 20 different places? No Thanks!).
(This was originally part of #4519, but we all agreeded it deserved it's own PR)
Stream::parseInt & Stream::parseFloat previously had protected
overloads which allowed skipping a custom character. This commit
brings this feature to the public interface.
To keep the public API simpler, the single paramter overload remains
protected. However its functionality is available in the public
interface using the two parameter overload.
robert rozee
Torben Nielsen
per1234
Cristian Maglie
Matthijs Kooijman
Eric Wieser
Cristian Maglie
liu nick
Sandeep Mistry
Martino Facchin
Patrick Roncagliolo
Iván Pérez
Niklas Mischkulnig
Kyle Chisholm
Arturo Guadalupi
Christopher Andrews
mattb5906
Benoît Blanchon
tico-tico
Paulo Costa
vbextreme
NicoHood
Chris--A