Introducing DELIMIT (plus SPACED and UNSPACED)

Ren-C offers an alternative to historical AJOIN and REJOIN called DELIMIT:

>> delimit "," ["foo" 1 + 2 "bar"]
== "foo,3,bar"

>> delimit/head "," ["foo" 1 + 2 "bar"]
== ",foo,3,bar"

>> delimit/tail "," ["foo" 1 + 2 "bar"]
== "foo,3,bar,"

>> delimit/head/tail "," ["foo" 1 + 2 "bar"]
== ",foo,3,bar,"

If you don't want it to reduce, you can use an @ block:

>> delimit "," @["foo" 1 + 2 "bar"]
== "foo,1,+,2,bar"

SPACED and UNSPACED are specializations, running through the same code for stringifying any non-string contents.

>> spaced ["foo" 1 + 2 "bar"]
== "foo 3 bar"

>> unspaced ["foo" 1 + 2 "bar"]
== "foo3bar"

PRINT uses SPACED internally when you pass it a block. You can use print unspaced to pass it a string if you want unspaced intent.

>> print ["foo" 1 + 2 "bar"]
foo 3 bar

>> print unspaced ["foo" 1 + 2 "bar"]

If you use a character (like #a) instead of a string (like "a"), the delimiting won't be applied. For instance, a newline won't leave spaces on either end of the line break:

>> spaced ["foo" 1 + 2 newline "bar"]
== "foo 3^/bar"

NULLs will trigger errors, VOIDs will vanish.

>> spaced ["foo" if false [1 + 2] "bar"]
== "foo bar"

>> spaced ["foo" pick [a b] 3 "bar"]
** Script Error: non-NULL value required (see MAYBE, TRY, REIFY)

>> spaced ["foo" maybe pick [a b] 3 "bar"]
== "foo bar"

Historically Rebol has had problems with this, that you don't really notice casually, unless you are doing something like writing to a file. If you take the program:

Rebol []
print ["Hello" newline "World"]

And then you figure out how to run it and redirect it to a file (which I don't know how to do in Red or Rebol2, it just brings up a I did R3-Alpha) you get:

>> read/string %output.txt
== "Hello ^/World^/"

Because it lacks shades of distinction in non-valued intents, the #[none] value basically just always prints in Redbol.

>> ajoin ["foo " if false [1 + 2] " bar"]
== "foo none bar"

>> ajoin ["foo " pick [a b] 3 " bar"]
== "foo none bar"

By and large, it gives the worst of all possible worlds... because wanting to know about an error or having it vanish are both more common intents than wanting to see the word "none".


A post was split to a new topic: Naming for Line and Space Adders and Removers