Cannot compile sketch containing invisible characters

[kittaakos]

Describe the problem

Sketch compilation fails if the sketch contains invisible characters. I do not know if it's a builder, preprocessor, or pure C++ issue.

To reproduce

void setup() {
  char s = 'x' ;
}
void loop() {}

Note that after the 'x' there is an invisible character.

Compilation:

a.kitta@Akoss-MacBook-Pro sketch_jul18b % ~/dev/git/arduino-ide/arduino-ide-extension/build/arduino-cli version
arduino-cli  Version: 0.25.0-rc1 Commit: 63b53c0f Date: 2022-07-11T16:09:52Z
a.kitta@Akoss-MacBook-Pro sketch_jul18b % cat ~/Documents/Arduino/sketch_jul18b/sketch_jul18b.ino 
void setup() {
  char s = 'x' ;
}
void loop() {}
a.kitta@Akoss-MacBook-Pro sketch_jul18b % ~/dev/git/arduino-ide/arduino-ide-extension/build/arduino-cli compile -b arduino:samd:mkr1000 ~/Documents/Arduino/sketch_jul18b 
/Users/a.kitta/Documents/Arduino/sketch_jul18b/sketch_jul18b.ino:2:15: error: stray '\302' in program
   char s = 'x'��;
               ^
/Users/a.kitta/Documents/Arduino/sketch_jul18b/sketch_jul18b.ino:2:16: error: stray '\240' in program
   char s = 'x'��;
                ^



Used platform Version Path                                                                  
arduino:samd  1.8.13  /Users/a.kitta/Library/Arduino15/packages/arduino/hardware/samd/1.8.13

Error during build: exit status 1

Expected behavior

If it's possible to compile a cpp file that contains the <0xa0> invisible character, I would expect to work with the CLI too.

Arduino CLI version

arduino-cli Version: 0.25.0-rc1 Commit: 63b53c0 Date: 2022-07-11T16:09:52Z

Operating system

macOS

Operating system version

12.3.1

Additional context

Based initially on arduino/arduino-ide#1210.
Possibly related (based on the compiler error message): #1785

Issue checklist

• I searched for previous reports in the issue tracker
• I verified the problem still occurs when using the nightly build
• My report contains all necessary details

[matthijskooijman]

Looks like an error thrown by gcc. I am pretty (but not 100%) sure that including such characters (at least outside of a string constant) produces an invalid C++ file, so gcc is correct to complain. I'm going to go ahead and closes this for not being an actual bug.

One could argue that there is a bug where these invisible characters should be more prominently displayed somehow, but that would need to reported in another repository (and I think this is unlikely to be fixed in 1.x, and I suspect your last screenshot is from 2.x, so it would already be reasonably handled there?).