What happens in 1...1 in PHP?

Question

On the web I came across the following piece of code:

echo 1...1;

Resulting in 10.1. I did the var_dump of the same and the return is:

string(4) "10.1"

Works with variable assignment:

$x = 1...1;
echo $x; // 10.1

And for other values:

echo 2...2; // 20.2
echo 1...1; // 10.2
echo 3...2; // 30.2

I know that in PHP 5.6+ there is the operator ... (splat Operator), that extracts the values of a array - similar to the * Python - but it doesn’t seem to be the case, because, according to the site 3V4L, the result is the same since version 4.3.0, that there was no such operator.

Apparently it only works with integers.

Would be some implicit value concatenation that PHP does?

Answer 1

In fact, what occurs is the concatenation of values of the type float. PHP analyzes 1. as being float(1.0) and .1 as being float(0.1), in this way, the concatenation of float(1.0) . float(0.1), however, how 1.0 is whole, PHP keeps only 1.

1. . .1
|  |  |
|  |  +--- Analisado como float 0.1
|  +--- Operador de concatenação de string, faz o cast dos operandos para string
+--- Analisado como float 1.0, feito o cast para int

That is, to do:

echo 1...1;

Is the same as:

echo (1.).(.1);

So much so that it is possible to add the decimal part of the first value:

echo 5.32..7; // 5.320.7

But it is not possible to add an entire part to the second value:

echo 3..5.67; // Syntax error

Only if parentheses are properly inserted:

echo 3..(5.67); // 35.67

Or spaces between operators:

echo 3. . 5.67; // 35.67

Using space it is still possible to define both the decimal part of the first value and the entire part of the second:

echo 3.14 . 1.41; // 3.141.41

Summary of the opera

I ran a series of testing and the result is below. I did not include situations that would generate syntax error.

1...1        = 10.1
2...2        = 20.2
10...10      = 100.1
(1.).(.1)    = 10.1
1. . .1      = 10.1
5.32..7      = 5.320.7
3..(5.67)    = 35.67
3. . 5.67    = 35.67
3.14 . 1.41  = 3.141.41

According to the tests in 3V4L, the output is the same since version 4.3.0: