SPLIT does not mimic R2 behaviour with STRING! delimiter

[rebolek]

Expected behavior

This is R2 version of parse:

>> line: "a,s.d#f;g"
== "a,s.d#f;g"
>> parse line ",.#;"
== ["a" "s" "d" "f" "g"]

Actual behavior

However, Red's split can’t reproduce the behaviour:

>> line: "a,s.d#f;g"
== "a,s.d#f;g"
>> split line ",.#;"
== ["a,s.d#f;g"]

Steps to reproduce the problem

See above 👆.

Red version and build date, operating system with version.

latest version on macOS

[xqlab]

It can

>> line: "a,s.d#f;g"
== "a,s.d#f;g"
>> split line charset ",.#;"
== ["a" "s" "d" "f" "g"]
>>

[rebolek]

Right, it can, but the question is, if it's bug or feature. I can live with both I guess 😃

[xqlab]

feature by design

[dockimbel]

It's by design.

Red's split is not meant to be limited to what R2's splitting with Parse could accomplish. Splitting with a sequence of characters instead of just a set of characters is an example of such extended feature. For splitting with a set of characters (like R2's parse), you have to provide a bitset! as shown above.