to-string and form omit the "decoration" for any-word! values , but not for any-path! values

[meijeru]

Expected behavior

to-string and form should treat any-word! and any-path! values similarly

Actual behavior

form quote :a => a ;; the : is left out
form quote :a/b/c => :a/b/c ;; the : is kept

to-string does exactly the same; both do this for get-, lit- and set- words and paths.

I plead for consistency, as usual

Steps to reproduce the problem

Execute in any interpreter

Red and platform version

all

[meijeru]

A full overview of the behaviour of to-string form mold and mold/all is found here View as HTML.

[greggirwin]

This is interesting. It occurred to me that we should look at how values are round-tripped in conversions. e.g.

>> to file! "a"
== %a
>> to string! to file! "a"
== "a"
>> to file! mold to file! "a"
== %%25a
>> to file! form to file! "a"
== %a

Which came from looking at the code and seeing that the current behavior is by design. *-word! types inherit form/mold from the base word! type, but *-path! types do not inherit them from path!. They explicitly have their own implementation.

This makes them compatible with R2, but not R3. In R3 you have this behavior:

>> form quote :a
== ":a"

>> form quote :a/b/c
== ":a/b/c"

>> to get-word! form quote :a
** Syntax error: invalid character in: ":a"
** Where: to
** Near: to get-word! form quote :a

>> to get-path! form quote :a/b/c
== ::a/b/c

Seems we need to think about big picture goals and uses, decide, and document those choices.

[hiiamboris]

@meijeru how about the 1st line in your list: mold/all () returns "unset" which is loaded as a function
Is there an issue about that?

[meijeru]

@hiiamboris Not that I know of.