Hi Patrick,

The name of the method is misleading. The intention was not to change the 
encoding of the receiver nor to filter out out-of-range bytes but to 
create a ByteString from a WideString when it only contains byte 
characters. So, the method will return a string equal to the receiver. The 
returned string will be a ByteString if an only if #isOctetString returns 
true.
I don't think the conversion in your example would make much sense, 
because it's not reversible: there's no way to recreate the string from a 
ByteArray.

Levente

On Fri, 14 Jun 2019, patrick.rein at hpi.uni-potsdam.de wrote:

> Hi everyone,
>
> while working on parsing network data I tried to use asOctetString and noticed that it did not yield a ByteString when called on a WideString. Is this the intendended behavior? I would have expected a behavior similar to the one I documented in the test case below. If not I think that the problem simply originates from using #at: which is overridden by WideString. 
>
> Bests,
> Patrick
>
> testAsOctetStringFromWideString
>
> 	| rawStringOctet wideStringAsOctet wideString |
> 	rawStringOctet := #[103 114 252 223 101 "The character 16r1fA02 starts here" 0 1 250 2].
> 	wideString := 'grüße' , (String value: 16r1FA02).
> 	wideStringAsOctet := wideString asOctetString asByteArray.
> 	self assert: rawStringOctet equals: wideStringAsOctet.