World: r3wp

>> parse [3 + 2] [some [integer! (print "number") | ['+ | '- ] (print 
"op")]]
number
op
number
== true

>> parse [3 - 2] [some [integer! (print "number") | ['+ | '- | '* 
| '/ ] (print "op")]]
** Syntax Error: Invalid word-lit -- '

** Near: (line 1) parse [3 - 2] [some [integer! (print "number") 
| ['+ | '- | '* | '/
 ] (print "op")]]

The second one failed when I tried to extend the dialect with multiply 
(*) and divide (/). After further experimentation, it seems that 
you can't escape the "/". Google has not been helpful here... Does 
anybody have any ideas? I could parse for just a word! instead of 
the +, -, etc., but I wanted parse to do the work of deciding what 
was a valid operation or not. Sorry for the multiple messages, I'm 
still trying to figure this client out... Thanks for any advice!

JoshF: Rebol load does not parse the '/, but you can do:

as-lit-word: func ['word [any-word!]] [to lit-word! word]
lit-div: as-lit-word /

parse [3 - 2] [some [integer! (print "number") | ['+ | '- | '* | 
lit-div] (print "op")]]

Ha! Black magic! That works a champ Ladislav, thanks very much!  
I had tried 
>> tdiv: to-word "/"
== /

>> parse [3 / 2] [some [integer! (print "number") | ['+ | '- | '* 
| tdiv ] (print "op
)]]
But had gotten the same error. What makes yours work?

Both tdiv and lit-div type? to a word!...

My example works, since the LIT-DIV variable refers to a lit-word, 
while your tdiv refers to a word

check as follows:

type? :lit-div
type? :tdiv

If LOAD won't eat a block, PARSE won't either, so you can test your 
block with LOAD. Some words can't be typed directly in, hence ladislav's 
solution.

And also hence the expression "a block is or isn't loadable"

OK... Mechanically, I see what you're saying, but what's the difference 
between a lit-word and a word? The spirit eludes me...

just a different datatype

I thought there was only word!'s and then everything else were more 
concrete types. I guess what I am asking is what is the purpose of 
lit-words?

Or are they just used for the special case of dealing with a / in 
load? ;  - )

in Parse, lit-words are used for matching, while words are looked 
up for values, which then are used for matching, so totally different 
behaviour

Compare:
>> parse [a] [a]
** Script Error: a has no value
** Near: parse [a] [a]
>> parse [a] ['a]
== true

I think you can say, that a word can be an evaluated lit-word. When 
you are typing a word directly into the console, you evaluate the 
word into a value that it's bound to. When entering a lit-word, it's 
evaluated into a word.

OK... So, let me paraphrase... As far as REBOL is concerned, lit-words 
are used only by the parse dialect to represent a thing to match 
to, whereas words are evaluated to find the thing to match to. However, 
because of parsing constraints in REBOL as a whole (the significance 
of "/" when dealing with indexable variables), there's no way to 
"escape" the slash into an unevaluated (literal) word without the 
dodge you showed me.

right

OK... Thanks very much. That helps a lot. I was right down the road 
to writing an expression parser, then that whole slash thing stopped 
me dead in my tracks. Now I should be able to get into some _real_ 
trouble!

a trap that you might fall into:

type? first [none]
== word!

type? first reduce [none]
== none!

type? first reduce ['none]
== word!

...except for the fact, that lit-words are used in the Do dialect 
(= when Rebol is concerned, as you say), when you want to write an 
expression, which evaluates to a specific word, so, e.g. the expression:

'a

evaluates to the same value as the expression:

first [a]

, which happens to be the word A

http://www.rebol.com/docs/core23/rebolcore-15.html#section-6

Depending on the situation, it can be hard to tell whether you are 
dealing with a word or a specific value. that's the price for freely 
interchangable code/data. :-)

a: [none]

b: copy a

b: reduce b ; me doing this behind your back

a
== [none] ; word!

b
== [none] ; none!

it is a bit difficult to understand recursive rules, but :-)

The difference between what I'm doing and what you linked to is that 
it's working against a string, while I'm doing a dialect, no?

I understood that character stuff wouldn't work in a dialect -- but 
my understanding is imperfect.

right, what you are doing is a dialect

OK. Thanks again for the timely help! I have to run off to work (which 
is firewalled up the yang), so you'll be able to avoid more silly 
questions from me for at least the next ten hours! ; - )

Dialect is a dialect. The only difference in string vs block parsing, 
imo is, that with block parsing, you are using REBOL datatypes to 
identify/match your types, whereas with string you are more "free-form" 
:-)

I know I was stopped by parse in some occasions where. I think always 
every time the problem would be solvable if I had for example >> 
to [ "A" | "B" ] where parser would check where is A and where is 
B and go to the closest one.

from Advocacy --> Graham [ to "A" | to "B" ] won't work as I want 
.. I will try to find a concrete example

this is a current parse limitation.

parse "start 111 end start 222 finish" [ some [ thru "start" copy 
NUMS [ to "finish | to "end" ] ] ]   this wont work

change it

[ to "end" | to "finish" ]

ok .. but I meant that you have "start 111 end start 222 finish start 
333 end "  then it won't work :)

change the rule again

I was trying to show an example where you have two possible endings 
and you want to process both (and you can differently with parens) 
) but you don't know in what order they will come or anything

In this case I would use block parsing ... then I'm no expert in 
parsing

parse string [ some [ "start" digits "end" | "start" digitis "finish 
]]

your problem is because you are using 'thru which breaks the other 
rule

yes , then you have to do charset parsing (but I don't know that 
yet :) ) .. I was just trying to say if there would be the way to 
say something like "to any [ "A" | "B" ] and it would go to the closest 
one A LOT of problems with parse would be easily solvable

you can use to but it still won't work

[ some [ "start" digits [ "end" | "finish" ] ]
should work

to go to the closest one .. means it has to try all the rules??

and see which has the best fit ?

no wgih is  the closest .. look at this example (I hope this will 
be better)

whigh = which

I know what you mean .. so you have to order your rules knowing what 
the data looks like