AltME groups: search

so I'll go back to the batcave and continue working on remark v2, 
and some other stuff... I want to release since a long time

The real advantage to the TO/THRU enhancement comes when it lets 
you avoid creating charsets, which are a lot less useful with Unicode. 
It should be pretty easy to implement.

I think that the proposals are more than Carl was thinking they would 
be - apparently he had forgotten the previous proposal lists. I don't 
think that it will be too much of a problem though, as there are 
not really that many proposals that are likely to be accepted. Some 
are competing proposals, of which only one would be chosen. Also, 
there aren't that many proposals overall - they are just thoroughly 
specified.

Lets see how things evolves. Proposal are very interesting as they 
would easy a lot of work on building parse rules. Everything is silent 
apart some blog messages where I have read for the first time the 
word "Beta" connected with REBOL3.

My main concern is that Carl's main requirements of the proposal 
process have been ignored in some cases:

- That the proposals be concisely specified. The Purpose and Importance 
statements should be one sentence each.
- That there be no discussion of theory.
- That there be no specification of equivalent rules.
- That all discussions happen outside of the wiki.
- That this is a proposals page, not documentation.

As it is, I hope Carl will read a paper that long when he gets to 
the point of taking on PARSE.

aha, I remeber I learned a lot from that green page too.. thanks 
for links so far , I will read the pages and hopefully I will find 
something related to the problems I have

or in terms of Brett's examples: == true

>> a: copy "dog cat" parse a [ ANY [ thru "dog" (print 1) | thru 
"cat" (print 2) ] ]
1
2
== true

>> a: copy "cat dog" parse a [ ANY [ thru "dog" (print 1) | thru 
"cat" (print 2) ] ]
1
== true

basically similar problem that last time as I see now.. so by looking 
at that mailing list answers I have 2 solutions ... I use parse 3 
times on a string.. or maybe I use Ladislav's parseen which he said 
solves this.. but I don't yet know how :)

S WORKS IF IN THIS ORDER

=heading=

{comment some comment}

- line 1
- line 2

------------->

<h1>heading</h1>

<p>comment some comment</p>

<li>line 1<li>
<li>line 2</li>

THIS DOESN'T WORK

=heading=

{comment some comment}

=heading=

- line 1
- line 2

=heading=

{comment some comment}

ADDITIONAL (SIMILAR) PROBLEM

- line 1
+ line 2
+ line 3
- line 4
+ line 5

----------------->

<li class="a">line 1</li>
<li class="a">line 2</li>
...

>> ? complement
USAGE:
    COMPLEMENT value

DESCRIPTION:
     Returns the one's complement value.
     COMPLEMENT is an action value.

ARGUMENTS:

     value -- (Type: logic number char tuple binary string bitset image)
>> ? union
USAGE:
    UNION set1 set2 /case /skip size

DESCRIPTION:
     Creates a new set that is the union of the two arguments.
     UNION is a native value.

ARGUMENTS:
     set1 -- first set (Type: series bitset)
     set2 -- second set (Type: series bitset)

REFINEMENTS:
     /case -- Use case-sensitive comparison
     /skip -- Treat the series as records of fixed size
         size -- (Type: integer)
>>

uh, that is some advanced parse :) .. I will need a couple of days 
to think it through

I'll try to explain complement.  I like to think of a charset being 
a list of valid chars that can be tested for.  However, say you need 
all characters of the alphabet minus a few.  Instead of defining 
multiple ranges of characters as in charset "A-FH-K N-T V-Wa-z0-9" 
which effectively skips the chars G L & U, you could simply state 
complement[GLU], which would exclude these three characters from 
the charset but include all others.

If there's something more specific or a technically better way to 
state the above please ad your infput

Very interesting, both versions (Oldes and Steeve) , thanks a lot.. 
I think I understood most of it now

Is there any better way how to change the main parse rules during 
parse like this one? (just a simple example..in real life the lexers 
would be more complicated :)  
d: charset "0123456789"

lexer1: [copy x 1 skip (probe x if x = "." [lexer: lexer2]) | end 
skip]
lexer2: [copy x some d (probe x lexer: lexer1) | end skip] 
lexer: lexer1
parse "abcd.123efgh" [ some [() lexer]]

Not really Oldes... but what is your purpose ? isn't that a little 
obfuscated again 

You said it's just an example, but why can't you use the normal way 
? I would like to know...
 
parse "..." [
   some [
	#"." lexer2
            | lexer1
  ]
]

This weekend I got an interesting idea: algebraic (and recursive) 
data types are well known for their ability to implement parsers. 
And they are a great data modeling tool.

E.g: 

data Bill = Name BankAccount | 
                   Company CreditCard

data CreditCard = CVC2 CCNumber CCExpiryDate 


However, the opposite also holds, i.e you can model data domain using 
named parse rules without actions just as easy. Now, what if you 
would combine two dialects: one to define data structures and a separate 
one to attach actions. 

E.g.

Post: [ message [string!] author [string!] timestamp [date!] ]
Comments: [ some posts]
blog [ 1 post comments]


action 'JSON 'Post [  .... the action to convert the Post to JSON 
here ...]

action 'XHTML 'POST [ ..... the action to convert Post to XHTML here...]

process some-data 'JSON

-> this gives back the data processed as for the JSON actions. It 
is a bit SAX like, with the difference that this models classes of 
action and separates them from the data in stead of scattering some 
lose actions. And, the data modeling still holds.

Then make actions for data to go to JSON, XML, XHTML, back and forth 
to a database,....

Chris:

1) Yes, actually, that would be yhe idea

2) I think the data dialect would be a strict subset of parse, forcing 
you to use set-word/parse-rule pairs Hence, the set-words are available 
in the action.

They call it a big Mac not a big Mc ... odd

hi, it's me again with parse problems...  I need this concretely 
to parse out web-page meta tags.. but I distilled the problem out 
of it to a minimal example..

doc1: "start A 1 end start B 2 end"  how can you get value of  2 
out

It works with a because it's first , but becasuse it enters the "parse" 
with it and then doesn't match it doesn't again test the B 

>> parse doc1 [ "start" "A" copy R to "end" (print R) to end ]
 1
== true
>> parse doc1 [ "start" "B" copy R to "end" (print R) to end ]
== false

>> doc1: "start A 1 end xyz B 2 end" ;; in this case it must not 
take 2
== "start A 1 end xyz B 2 end"

>> parse doc1 [ "start" thru "B" copy R to "end" (print R) to end 
] ;; but it will that's why I can't u
se to\thru
 2
== true

some ["start" ["A" | "B"] copy R to "end" "end"]

ups ... my example above is wrong .. just a sec

(this is the right example .. I forgot to use thru above so second 
wouldn't pass anyway... but result is the same)

>> doc1: "start A 1 end start B 2 end"
== "start A 1 end start B 2 end"

>> parse doc1 [ thru "start" "A" copy R to "end" (print R) to end 
]
 1
== true

>> parse doc1 [ thru "start" "B" copy R to "end" (print R) to end 
]
== false

>> parse doc1 [ SOME [ thru "start" "B" copy R to "end" (print R) 
to end ] ]
== false

Is there anything expected between "start" and "A", for instance 
?

parse doc1 [some [thru "start" ["A" | "B"] copy R to "end" (?? R) 
"end"]]

hm.. just a sec so I try few things

Your solution, I thought it won't work if I reverse order of A and 
B in the string but it seems it does.  I would need to know which 
one is A and B but I think this can be solved by setting some word 
( ) inside [ A | B] ...  so basically it seems to work... I think 
I can apply this way also to my concrete problem which is this

( I need to parse meta tags description and keywords and abstract 
if they exist -- they can come in any order, there can be one or 
multiple spaces/newlines/tabs between tag arguments, there can be 
" or '  used as argument="asdasd" )
>> doc2: {<head>

{    <title>Dragonicum.com - making the right business connections 
!</title>

{    <meta name="keywords" content="Company Directory, Join Us, Advanced 
Search, Trade Leads, Forum, Trade S

{    hows, Advertising, Translation, fair trade, trade portal, business 
to business, trade leads, trade even
{    ts, china export, china manufacturer" />

{    <meta name="description" content="New international trade portal 
and company directory for Asia, Europe

{     and North America. Our priority No.1 is to create and maintain 
a safe, well lit business-to-business m

{    arketplace, by assisting our members in identifying new trustworthy 
business partners!" />

{    <link rel="stylesheet" href="style/blue_main.css" type="text/css" 
/>}
== {<head>

<title>Dragonicum.com - making the right business connections !</title>
<meta name="keywords" content="Company Directory...


>> T: "" parse doc [ thru "<meta" "name=" skip "keywords" skip "content=" 
m: skip (m1: first m ) copy T to m1  to end ] print T

Company Directory, Join Us, Advanced Search, Trade Leads, Forum, 
Trade Shows, Advertising, Translation, fair trade, trade portal, 
business to business, trade
leads, trade events, china export, china manufacturer

>> T: "" parse doc [ thru "<meta" "name=" skip "description" skip 
"content=" m: skip (m1: first m ) copy T to m1  to end ] print T

>>

( as you see because keywords are first it works for them , but doesn't 
for description , they can be in different order in other document 
etc)

maybe your solution for A | B would work.. I will try

yes :) thanks a lot!

>> T: K: D: "" parse doc [ SOME [ thru "<meta" "name=" skip [ "description" 
(V: 'D) | "keywords" (V: 'K)] skip "content=" m: skip (m1: first 
m ) copy T to m1
(set V T) ]  to end ] ?? K ?? D

K: {Company Directory, Join Us, Advanced Search, Trade Leads, Forum, 
Trade Shows, Advertising, Translation, fair trade, trade portal, 
business to business, tr
ade leads, trade events, china export, china manufacturer}

D: {New international trade portal and company directory for Asia, 
Europe and North America. Our priority No.1 is to create and maintain 
a safe, well lit busi

ness-to-business marketplace, by assisting our members in identifying 
new trustworthy business partners!}

== {New international trade portal and company directory for Asia, 
Europe and North America. Our priority No.1 is to create and mai...
>>

I intended to make a blogpost .. "REBOL parse challenge" and present 
this problem and ask if people can provide solutions in other languages 
that would be more elgant ... (in similar note as the "arc challenge" 
... now that it seems even more hard nut to crack I should probably 
really do it .. does anyone think this would be easy to solve using 
the conventional language? (I think not)

hm.. would this be nicely solvable with a regex? .. I think it would 
be quite a pain by using regular string functions like strpos substr 
etc... having the same requirenments (one or more spaces/tabs/newlines 
" or ' , undefined order)

I don't know - I only learn regex when I have to .. then a short 
time later I forget.

What would you build a state machine with, which would generate so 
much code ?

You say "state machines ... require more code". What code ? Obviously, 
you can build a state machine in any language, but I guess I'm wondering 
what ... ohh... I'm so tired after all those cheese sandwiches....

Anyway, I think I understand what you're saying. A state machine 
is big and clunky, expressing everything you don't want to hear about, 
while parse allows you to express your target more directly, cutting 
through anything you don't want without having to specify it.

I don't know the exact term for this but I build many parsers for 
things like xml, wiki text and some other custom things in various 
lower level langauges using simple state machine (at least that's 
how I called it)... To my understanding you can parse anything with 
something like that, also structured nested data with it but it of 
course takes some more coding than this rebol solution... what I 
mean as a state machine is a loop that accepts characters or words 
and has a predefined number of states and code for what to do at 
each state and when to switch to another state etc..

The first one without any quotes causes a little bit of a problem 
(solvable).

You have to use a variable to store which one was used, then parse 
until that character is encountered again.

Is this a surprise ?
>> parse "abc" [some ["b" | "c" | "a"]]
== true

Yes, it takes a little while to become familiar with parse.

this does surprise me a little , but I am not sure if this was the 
problem or something else, because I hrought I tried with some and 
all things

It means, basically:
	SOME: Do this 1 or more times, until fail or end is reached:

  [Try "b", if that fails, try "c". If that fails, try "a"]     <--- 
  Given "a" "b" "c", this rule always succeeds.

( the problem is at things where things repeat adn I don't know in 
which order they will appear .. I had this problem with parsing something 
like simplified wiki text )
>> a: "start1 1 end start2 2 end start1 3 end"
== "start1 1 end start2 2 end start1 3 end"

>> parse a [ SOME [ [ thru "start2" | thru "start1" ] copy T to "end" 
(print T) ] to end ]
 2
 3
== true

>> parse a [ SOME [ [ thru "start1" | thru "start2" ] copy T to "end" 
(print T) ] to end ]
 1
 3
== true

( to not give impression I have only problems with parse, I used 
parse to solve many things that would be headhurting any other way... 
these and problem upthere are just cases where I got into trouble)

apiece: [copy T to "end" (?? T)]

parse a [some [thru "start2" apiece | thru "start1" apiece]  to end]

This is basically not a problem , as I solve these things wiht multiple 
passes and it works more than fast enought for me that way also ... 
I think this problem would not exist if in case of [ .. | .. | .. 
] parse would check all options and take the one stat is least characters 
away from current position (that comes true the first) .. but this 
would most probably slow down the parse and you would loose the feature 
that you define "priority" with [ .. | ..  | .. ] now .. so maybe 
if there would be a different | for this

( I have to go to eat... will be back .. thanks a lot for before)

hm.. really thanks for this example.. I took it as unsolvable, but 
this is totaly elegant way to solve it .. I will need to think on 
this a little and do some more examples to difest it :) thanks

If you need to parse complex structures, like the marup language, 
you should use charsets and not 'to or 'thru commands... for example 
you cannot say that tag starts with < and ends with > because such 
a tag is valid as well:
<input value="<>">

The 'to and 'thru commands are useful, if you, for example, do datamining 
and don't care to parse all page structure to get just a bit of information 
from it.

Is there a way to force parse to inclose results in {} instead of 
double quotes "" regardless of length?

Hello everybody!

I have a problem. I need to extract email addresses from a big text 
like

bla bla [me-:-demo-:-com] bla bla ...  <[you-:-example-:-org]>  etc. [he-:-italy-:-it]

There is possible to obtain a text with all the addresses withou 
the "<" and ">"?

Here's absolutly terrible parser - it does NOT follow RFC, allow 
any combination of alpha chars, dots, one @ char, and the same, once 
again to the next space char ...

space: #" "
mailchar: charset [#"0" - #"9" #"A" - #"Z" #"a" - #"z" ".-"]
at-char: #"@"

email: [

   space
   start:
   some mailchar
   at-char
   some mailchar
   end:
   space
   (print copy/part start end)

]


str: "afadfa adfa asdfasdfa fd [asdfas-:-adfadf-:-adfa-adfadfsda-:-com] adfafaf 
a af"

parse/all str [any [email | skip]]

It would be nice if REBOL could LOAD foreign! data.  :)  Hint hint 
wink wink.


And being here in a public REBOL forum I might get in trouble for 
suggesting this one.

$ grep -o -E '\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,4}\b' 
files...

Brian, you can probably do that grep with a few CHARSET and PARSE 
in REBOL.

And actually I think it's wrong anyway ... as it should be.  Posting 
regex in a REBOL forum ... shame on me.   ;)

I'm having trouble parsing the "none" datatype from within blocks.
The following example illustrates my problem (hopefully):
junk: [none [1 2 [3 4]]]

parse/all junk [none (print ["nothing"]) text: (print ["text:" mold 
text]) set b block! (print ["block:" mold b])]

This produces the following output:
nothing
text: [none [1 2 [3 4]]]
== false


Notice that the block doesn't get parsed. It seems that parse ignores 
"none" tokens rather than extracting them from the input stream. 
If I put a number in place of none and parse for "number!", then 
the block does indeed get parsed.

Is this a bug or an oversight? Or am I just confused?

'none isn't a datatype - none! is:

>> parse/all junk [none! (print ["nothing"]) text: (print ["text:" 
mold text]) set b block! (print ["block:" mold b])]              
                           
nothing
text: [[1 2 [3 4]]]
block: [1 2 [3 4]]
== true

Ah, forgot to copy that part - I'd done "junk/1: none" to make sure 
it was a none value

it's just a serialized version of none!, so you can load it as a 
real none value instead of a word.

Pavel, this also works with datatypes.  For example:

>> mold/all string!
== "#[datatype! string!]"


This is useful if your loading values from a file.  This way your 
sure to set a value to a string datatype! when desired.

Hi, I have one question .. can you somehow break out of some loop 
by rebol code .. for example


parse [ aa zzz cc ]  [ some [ set W word! ( ?? W if equal? W 'zzz 
[ break ] ) ] ]  ...  that break doesn't work that way, but is there 
some way to do this? I need to compare W with a runtime value

I solved it in a way that I can just return out of whole function 
(with return) at that point so it's ok .. first I had it thought 
out in a way that I would need to exit the some [ ] loop but continue 
parsing .. error probably wouldn't work that way either? This is 
now my code..match: 

match func [ data rules ] [
	parse rules [ 
		SOME 

  [ 	set L lit-word! ( either equal? L reduce first data [ data: next 
  data ] [ return false ] ) | 
			set W word! ( set :W first data  data: next data ) 
		] 
	]
]

; Here's one way to do it...

>> digit: charset "1234567890"
== make bitset! #{
000000000000FF03000000000000000000000000000000000000000000000000
}

>> rule: [s: some digit e: (print copy/part s e) | h: #"a" (h: tail 
h) :h | skip ]

== [s: some digit e: (print copy/part s e) | h: #"a" (h: tail h) 
:h | skip]
>> parse "12b34c56a78" [any rule]
12
34
56
== true

Another possible way is by setting at runtime a [break] rule :

branch-rule: [ ]

parse [ aa zzz cc ]  [ 
	some [ 
		set W word! ( 
			?? W
			if equal? W 'zzz [ branch-rule: [ break ] ]
		)
		branch-rule
	]
]

charset create a "mask" in bitset form to be compared to the curent 
item read from the string

some digit since digit is a bitset containing the binary image of 
 what you looking for (numbers char from 1 to

the equivalent lame would be someting like foreach a string [ either 
find?  "1234567890" a [ append e a ][probe e clear e ] ]

Essentially what I'm doing with the above code is simply skipping 
to the end of the parse input when a given rule is matched. This 
works because a get-word in the parse rules sets the current parse 
input.  The get-word can be any value of the same type as the original 
parse input.  You can't set the parse input to a string! if a block! 
was provided to parse to start with.

I'd like to take an english sentence and tidy it up.  I want to automatically 
apply english grammar to it ... so capitalize the first letter after 
a period, and remove extraneous spaces eg. a comma after a space. 
 Anyone done anything like this with 'parse?

Not yet but I've been thinking about it for quite a while now... 
 I think I have a pretty good idea what the parse rules should look 
like but I haven't written any code for it yet.

Good start...

letter: charset [#"a" - #"z" #"A" - #"Z"]
dirt: complement letter
word: [some letter]
clean: [here: dirt :here (remove here)]
space: [here: (insert here #" ") skip]
capital: [here: letter (uppercase/part here 1)]
sentence: [
	some [
		  capital opt word break
		| clean
	]
	any [
		  [#";" | #","] any clean space word
		| #"." any clean space capital opt word
		| #" " word
		| clean
	]
]

parse/all text: {test  test . test;; test ..test } sentence
probe text
>>"Test test. Test; test. Test"

for #"'" you should add a rule to remove spaces

with that you supress unwanted spaces.
it'  s a good day
 --> "it's a good day"

so don't add ""'" as a vali

Also, I think have to add ' to the letter charset because words ending 
in s can have a trailing ' for possession ...

but what if they have inserted a space after or before '

hum ok, but you could handle that specif case with a different rule

parse is just amazing for such simple grammar.
A simple add and it's doing all you want.

Have I found a parse bug?

1)

>> parse/all {zybc} [ some ["b" break | "y"  break | skip] copy result 
thru "c" (print result)]
bc
== true

2)

>> parse/all {zybc} [ some ["b" break| "y"  break | skip] copy result 
thru "c" (print result)]
** Script Error: break| has no value

** Near: parse/all "zybc" [some ["b" break| "y" break | skip] copy 
result thru "c" (print result)]

3)

>> parse/all {zybc} [ some ["b" break | "y"  break| skip] copy result 
thru "c" (print result)]
== false


Such stupid bugs are really making the testing process difficult. 
I wondered at least 5 minutes, why the result of case 3 was wrong, 
and then I tried to add space behind the second break, and the code 
was corrected. How is that second break| does not report error? ;-)

3) is like 2)  you put a | to close of the second break. I noticed 
on rebol 2 strange reactions with find multi-case too

yes, you might be right doc. But - it is really very difficult to 
track down for user. It almost looks like scanner bug, but it is 
not. What actually happens in the case 3) is, that "break|" is being 
considered a regular word, which just does not have value. Stating 
that, it also means that 'skip is not part of OR expression. So, 
'some block fails on not matching "y" ....

Here's a parse question for the experts.

If I have a document with headings eg. a: b: .. z: and text optionally 
under each heading ... would it be possible to use parse to collect 
all the text from each heading if the headings are in any order and 
some headings with no text are optionally missing?

now was that a question of the "can you give me the solution" kind?

It's a little complicated because the headers can have spaces in 
them.

now if you have a rule
copy text [ to "a:" | to "b:" .... ] 

but if b: occurs before a: in the text, then you will include a header 
in copied text

yes, headers start on a newline and terminate in ":"

No, there can be a ":" in the content

but you know what the headers are ... so that's not a big problem.

can you give the name of some the headers... or an example.... so 
far it looks like a really simple rule to me.

I can assume it starts at a header?

So, I am trying to create an object from a semi structured document 
where the object elements are in any order or missing.

I guess I can do it without using parse .. just replace all the headers 
with a mark, that allows me to split off all the sections, and then 
i can match the sections with all the section headers.

There is no reason, the content is enclosed in a string before being 
loaded.
If it fails, it's because the whole grammar has changed