\ from Newsgroups: comp.lang.forth
\ Date: Sun, 28 Dec 2014 14:12:44 -0800 (PST)
\ Message-ID: <fdfa0f77-57f9-4ea6-b5bc-32d5651aabef@googlegroups.com>
\ Subject: String literacy
\ From: Julian Fondren <julian....@gmail.com>
\ Adapted to recognizer RFDv3 by MT
\ License: probably public domain, contact author
\ Also, notice how I intend code by two spaces? That makes it very
\ clear what's code and what's me talking about code, without the
\ reader having to keep switching between two files, say. It also
\ allows me to tie documentation and source together in such a way
\ that one will seldom be supplied without the other. You could
\ even organize your code into 'blocks', with corresponding 'shadow
\ blocks' ... well, once again, let's just do it:
\ This is the entire contents of literate.fs:
' noop ' noop ' noop recognizer: r:noop
: literacy-recognizer ( c-addr u -- r:noop | r:fail )
2drop source s" " string-prefix? if r:fail else 0 parse 2drop r:noop then ;
\ place it at the top of the recognizer stack
forth-recognizer get-recognizers
1+ ' literacy-recognizer swap
forth-recognizer set-recognizers
And the entire rest of my post will be the contents of a file
named hello.fs, which makes use of it:
--- hello.fs begins next line ---
require literate.fs
Having required literate.fs, the rest of this file is commentary unless
intended by two spaces.
For example, Forth will pass over the following:
: hello ( -- ) cr ." Hello!" ;
But will compile and execute these indented statements:
: goodbye ( -- )
cr ." Good bye." ;
goodbye bye bye bye
(This behavior is also seen interactively.)
