Comments on: A New Datatype: Typeset!

Carl Sassenrath, CTO
REBOL Technologies
8-Apr-2006 20:21 GMT

Article #0002
Main page || Index || Prior Article [0001] || Next Article [0003] || 7 Comments || Send feedback

This article has been changed from its original posting. The word "typemap" has been changed to "typeset".

In REBOL 3.0, a typeset is essentially a set of datatypes.

For example, when you define a function, you specify the datatypes of arguments in this way:

f: func [arg1 [integer! number!] arg2 [string! url!]] ...

What follows each arg word is a set of accepted datatypes.

REBOL 3.0 adds a new datatype called a typeset! A typeset is simply a compact, high-performance method of storing the datatype sets as a new kind of datatype. Typesets are important because the interpreter uses them to quickly validate function arguments.

The typeset! datatype knows how to convert a block of datatype names such as [integer! decimal! money!] into an internal representation (similar to the bitset datatype). It can also convert the internal format back to a block for output or changes.

The addition of typesets eliminates the need for the special REBOL pseudo-datatypes like series! and number!. These are now implemented as typesets, but you can use them the same way as before. They can be used in function argument specifications:

f: func [arg1 [number!] arg2 [series!]] ...

and also for datatype tests:

if series? value [....]

Note, however, this difference. In REBOL 2.*:

type? series!

in REBOL 3.0 becomes:

type? series!

For advanced users, REBOL 3.0 will provide a way to create and manage your own custom typesets.

tmp-type: make typeset! [time! money! percent!]

f: func [arg1 [tmp-type]] [...]

I will post more soon about a few issues related to typesets.



9-Apr-2006 13:00
why call it a typemap when it really is a set rather than a map?
Carl Sassenrath
9-Apr-2006 13:11
I considered calling it a typeset, and that is still possible. It is a good question, and I will answer it in a blog.
9-Apr-2006 13:20
thanks for the quick response, carl! keep up the good work!
9-Apr-2006 13:20
I assume this will be of use not only for function generation, but also block parsing, and 'find operations amongst others.

Also, How would one set a typemap query, eg:

mytype: make typemap! [string! issue!] mytype? "Foo" mytype? #Foo

Carl Sassenrath
9-Apr-2006 14:40
Yes: on parsing and find.

On building of the query, it would require a constructor function such as:

mytype?: make-query mytype

(I would hope we could find a better name than make-query too.)

Volker Nitsch
9-Apr-2006 18:24
:) match? if not match? integer! 0.5 [..] or mytype?: matches mytype something like that?
Allen Kamp
10-Apr-2006 21:16
:) Can typemap also handle ranges [integer! 0-9] ? And further restrictions on strings such as the maxlength? [string! min 0 max 20]

Very simple way to build in data validation

Post a Comment:

You can post a comment here. Keep it on-topic.


Blog id:



 Note: HTML tags allowed for: b i u li ol ul font span div a p br pre tt blockquote

This is a technical blog related to the above topic. We reserve the right to remove comments that are off-topic, irrelevant links, advertisements, spams, personal attacks, politics, religion, etc.

Updated 27-May-2024 - Edit - Copyright REBOL Technologies -