REBOL3 tracker
  0.9.12 beta
Ticket #0000733 User: anonymous


Short URL:
TypeIssue Statustested Date29-Apr-2009 10:09
Versionalpha 50 Categoryn/a Submitted bymeijeru
PlatformWindows Severityminor Prioritynormal

Summary numerical refinements have type any-word! => inconsistency
Description typeset any-word! includes refinement!

not all refinements are of the form /word, though

this leads to inconsistency
Example code
>> to-word /1
== 1
>> type? to-word /1
== word!
>> type? load mold to-word /1
== integer!

Assigned ton/a Fixed inalpha 58 Last Update22-Jun-2009 15:59

29-Apr-2009 16:26

All refinements are words, syntax differences aside. They can be bound - the important part of being in any-word! - and they even have single instance storage.

What your example code shows is that TO-WORD can generate words that when molded, can't be loaded as words. This is a separate issue that has been with us since the beginning of REBOL.
5-May-2009 05:17

I've reopened this ticket because it is problematic to validate such function refinements during function definition (we'd have to validate every refinement word of every function!). It's better we discover them earlier, throwing them as invalid words during lexical translation.
5-May-2009 05:21

See the comments to ticket #743.
21-Jun-2009 00:09

Such words are now more strictly validated and will throw a syntax error.

Date User Field Action Change
22-Jun-2009 15:59 BrianH Status Modified built => tested
21-Jun-2009 00:09 carl Comment : 0001002 Added -
21-Jun-2009 00:08 carl Fixedin Modified => alpha 58
21-Jun-2009 00:08 carl Status Modified waiting => built
5-May-2009 05:21 BrianH Comment : 0000626 Added -
5-May-2009 05:17 carl Comment : 0000625 Added -
5-May-2009 05:13 carl Status Modified dismissed => waiting
29-Apr-2009 16:26 BrianH Comment : 0000592 Added -
29-Apr-2009 16:19 BrianH Description Modified -
29-Apr-2009 16:19 BrianH Code Modified -
29-Apr-2009 16:19 BrianH Status Modified submitted => dismissed
29-Apr-2009 10:09 meijeru Ticket Added -