Console interpretes '-<- as a lit-word ('-) and a word (<-)

[PeterWAWood]

Describe the bug
The test lexer-test.red crashes with a script error:

>> do %lexer-test.red
~~~started test~~~ lexer
*** Script Error: --assert does not allow word! for its assertion argument
*** Where: --assert
*** Near :
*** Stack: do-file

The issue is caused by the fourth assertion in test #4562, line 1571

1565    ===start-group=== "issues"
1566                                                                                                                                         
1567         --test-- "#4562"
1568                 --assert [<< '<< <<: :<<] == transcode "<< '<< <<: :<<"                                                                 1569                 --assert ['=<= :=<= =<=:] == transcode "'=<= :=<= =<=:"
1570                 --assert word? transcode/one "=<="                                                                                      1571                 --assert '-<- == transcode/one "-<-"

To reproduce

>> '-=-
== -=-
>> '->-
== ->-
>> '-<-
*** Script Error: <- has no value
*** Where: catch
*** Near : <-
*** Stack:

>> <-: 1
== 1
>> '-<-
== 1

Expected behavior
-<- should be interpreted as a single lit-word not a lit-word and a word

Platform version

about/debug
-----------RED & PLATFORM VERSION-----------
RED: [ branch: "master" tag: #v0.6.4 ahead: 5075 date: 22-Dec-2023/9:07:34 commit: #6cb7d502bcbfae8b00630adee6561909eee2b478 ]
PLATFORM: [ name: "Raspbian GNU/Linux 11 (bullseye)" OS: 'Linux arch: 'aarch64 version: 6.1.21 build: "#1642 SMP PREEMPT Mon Apr 3 17:24:16 BST 2023" ]

---

[dockimbel]

The lexical rules for allowing < inside a word (or at the beginning of a word) are not precisely defined currently. This is not an easy task as it can easily conflict with the "sticky tag" rule, allowing a word sticking to a tag to be separated by the lexer and producing two values:

>> load "<p>text</p>"
== [<p> text </p>]

[greggirwin]

Since all other bracketing chars break words lexically, < probably should too. That makes it non-orthogonal with > being allowed, but is otherwise consistent. Just need to shift < from one list to the other in https://github.com/red/docs/blob/master/en/datatypes/word.adoc#3-literal-syntax.

The argument for supporting it in words would be that you can make nifty arrow-like words that go in both directions, but you could still do that with Unicode chars, and is far less common than lexing data where they may be together.