Rebol / Red Parse html rules returns true but nothing is inserted

Question

I have a parse rules that returns true but it doesn't insert my text as expected : the html is unchanged whereas it should have inserted at the end of the main closing div. I tried to use a counter like How to parse inside HTML tags with REBOL?

Update: I also don't know how to break out of the parsing as soon as counter = 0 so as not to insert text before last closing div after main.

    content: {<div class="main">
      <h1>
        Big TITLE
      </h1>
      <div>
        <section>
          <p>a paragraph</p>
        </section>
         <section>
          <p>a paragraph</p>
        </section>
          <section>
          <p>a paragraph</p>
        </section>
       </div>
       <div>
          <p>Blah Blah</p>
       </div>

    </div>
    <div>
      Another Div
    </div>
    }

    rules: [
      thru <div class="main">
      (div-count: 1)
      some [
        to "<div" (++ div-count) thru "<div" thru ">"
        |
        to </div> mark: (-- div-count if div-count = 0 [insert mark "closing main div"]) thru </div>
      ]
      to end 
    ]
    parse content rules

The answer you cite turns the input string into a structured block. If you are trying string parsing with HTML on your own...it's the *danger zooone*! PARSE is clearer than RegEx and [strictly more powerful](https://github.com/gchiu/rebol.net/blob/master/wikipedia/Parse%20Project.wiki#Theory_of_PARSE), yet most every [warning about trying to process HTML directly with RegEx](https://stackoverflow.com/a/1732454/211160) applies to PARSE as well. See @rgchris's [markup work](https://github.com/rgchris/Scripts/blob/master/experimental/markup.reb) (he is on chat here and on Red gitter) — HostileFork says dont trust SE, Oct 28 '17 at 09:02
@HostileFork I don't need and I don't want a full blown html parser, I just need to parse very specific html so rules are normally very simple. In the case above that seems trivial idea except it returns true but still doesn't insert, I failed to see why. — user310291, Oct 28 '17 at 09:09
I understand the desire to keep it simple, but one also needs to realize (as the linked post regarding RegEx points out) that there are limits. If it's just quick and dirty, ok, but that's for one-off scripts... — HostileFork says dont trust SE, Oct 28 '17 at 10:55

sqlab · Accepted Answer · 2017-10-28T09:34:05.427

Here a solution with probe to debug

rules: [
     thru <div class="main">
     (div-count: 1)
      some [
        "<div" (probe ++ div-count) skip
      |
        "</div>" mark:  ( probe -- div-count   if div-count = 0 [insert mark "closing main div"]) skip 
      |  skip
     ]
  ]
parse/all content rules

The problems with your rules are, that the div-count is never or seldom subtracted. The parse pointer goes straight to the next opening div as to is always the first fulfilled condition.

You can break out or better return from parse if you add a to end after a successful condition. If you are unsure use brackets for grouping [ sucessful sub-rules ... to end ]

An example with an end-rule

end-rule: [] ; or none
rules: [
    thru <div class="main">
    (div-count: 1)
    some [
        ["<div" (++ div-count) skip]
    |
        ["</div>"mark:  (-- div-count   if div-count = 0 [insert mark "closing main div"  end-rule: [to end]]) end-rule ]
    |  skip
]

]

hi works better than mine :) but I failed to understand why skip works and not thru: skip only skips one character. Also I forgot to ask how to insert only at first div count = 0 ? since I want to insert only for div main class. — user310291, Oct 28 '17 at 09:14

Rebol / Red Parse html rules returns true but nothing is inserted

1 Answers1

Linked