Strange looking limit checks in Teensy3 Version of Stream.cpp

[Frank2]

Hello,

when looking around how Serial is implemented, I stopped at this piece of code in Stream.cpp

String Stream::readString(size_t max)
{
	String str;
	size_t length = str.length();
	while (length < max) {
		int c = timedRead();
		if (c < 0) {
			setReadError();
			break;	// timeout
		}
		if (c == 0) break;
		str += (char)c;
	}
	return str;
}

If I understand the Arduino reference right then str.length() returns the actual size of the String, which is 0 after initialisation. After that the value of length is not more modified in the loop. So the condition is always while(0 < max).

Do I miss something that is happening in background or is this a bug ?

The same is happening in readStringUntil, too.

[MichaelMeissner]

If I understand the Arduino reference right then str.length() returns the actual size of the String, which is 0 after initialisation. After that the value of length is not more modified in the loop. So the condition is always while(0 < max).

Do I miss something that is happening in background or is this a bug ?

The same is happening in readStringUntil, too.

I haven't looked at the code. However, I would suspect that this line, which adds a single line to the end of the string bumps up the length:

		str += (char)c;

[Frank2]

Thanks for your answer,

yes that's exactly what I mean. The String object and its internal length ist growing in this line. But the variable with name "length" isn't, because it's nowhere changed in the loop.

I think

while (str.length() < max) {

is what is realy meant there.

But anyway tho code is Arduino compatible as it is, because the original Arduino libraries don't do any length checks there. So introducing a limit may break some people's code. The only problem is, that not having a limit may cause other trouble.

[PaulStoffregen]

Yes, I believe you're right. The length check is probably ineffective. A low priority bug which should get tested and fixed if the tests confirm the problem...