Recursive Tag Contents

Question

I'm looking to improve the below regex as I now would like to match nested tags as well:

'%{if:\s*"\'([^\']*)\' == \'([^\']*)\'"}((?:(?!{else}|{/endif}).)*){else}((?:(?!{/endif}).)*){/endif}%sei'

Which basically matches:

{if: "'x' == 'y'"}
    a
{else}
    b
{/endif}

or

{if: "'x' == 'y'"}
    c
{/endif}

However, I would like this to be recursive in some sort of way, so nested statements could also be matched without breaking anything (at the moment it breaks if a nested statement is added).

There would also be a similar expression with !=.

I've found this http://www.devnetwork.net/viewtopic.php?f=38&t=102670&sid=02b7c691a2be894336c694700f8f911a#p551340 which matches <div> tags, though a bit unsure how to adapt it to suit my regex...

Sorry, its for some logic that I'm allowing for some end users to create a template to appear with an area that they can customise. I don't want to allow PHP, so that all gets stripped out for security as are — MrJ, May 22 '12 at 10:27
why dont you use a readily available template engine then? Twig if you want PHP. Else XSLT. — Gordon, May 22 '12 at 10:29
A bit too big for my needs and do not require all that functionality - `if` and `else` is all that is needed — MrJ, May 22 '12 at 10:31

score 0 · Answer 1 · edited May 23 '17 at 12:27

0

If you limit the nesting to some predetermined depth (which might or might not be a bad idea), you could match it with a regex. Otherwise, you cannot. The link you provided matches HTML with regex, which is often used, but is known to be a bad idea in general. If you do not want to use other forms of parsing, consider matching innermost ifs, replacing with something and matching again.

edited May 23 '17 at 12:27

Community

1
1

answered May 22 '12 at 12:10

Eugene Ryabtsev

2,232
1
23
37

score 0 · Accepted Answer · answered May 22 '12 at 12:31

Using regex (PCRE) is not optimal in such cases as you would need to re-parse the inner content for each nested level (one of the reasons using a proper parser would be better).

That said, it can be done with a pattern like:

~
{if:\s*+
    (?<condition>
        [^{}]++
    )
}

(?<then>
    (?:
        (?:(?!{if:[^{}]++}|{else}|{/endif}).)*+
        (?R)*+
    )*+
)

(?:
    {else}
    (?<else>
        (?:
            (?:(?!{if:[^{}]++}|{else}|{/endif}).)*+
            (?R)*+
        )*+
    )
)?+

{/endif}
~six

Perl example @ ideone.

On this text

if: "'x' == 'y'"}
    a
{else}
    b
{/endif}

{if: "'x' == 'y'"}
    c
{/endif}

{if:minimal}{else}{/endif}

{if: "'nested' == 'things'"}
    {if: "'x' == 'y'"}x{if:minimal}{else}{/endif}x{/endif}
{else}
    b{if: "'x' == 'y'"}c{/endif}{if: "'x' == 'y'"}c{/endif}
{/endif}

{if:foo} unbalanced {if:bar}ignores first if{/endif}

it matches

*** matched if:
  * cond: "'x' == 'y'"
  * then:
    a

  * else:
    b

*** matched if:
  * cond: "'x' == 'y'"
  * then:
    c

*** matched if:
  * cond: minimal
  * then:
  * else:
*** matched if:
  * cond: "'nested' == 'things'"
  * then:
    {if: "'x' == 'y'"}x{if:minimal}{else}{/endif}x{/endif}

  * else:
    b{if: "'x' == 'y'"}c{/endif}{if: "'x' == 'y'"}c{/endif}

*** matched if:
  * cond: bar
  * then: ignores first if

Thanks - would this still function if I were to put {$tag_name} within the single quotation marks? — MrJ, May 22 '12 at 13:04
Sorry, for example `{if: "'{$tag_name}' == '123'"}true{else}false{/endif}` — MrJ, May 22 '12 at 15:36
@MrJ, not as it is right now. Change the condition expression (`[^{}]++`) to whatever you want it to accept. — Qtax, May 22 '12 at 16:43
@MrJ, haven't I answered the question? Or is something wrong with the answer? — Qtax, Apr 01 '13 at 22:41

Recursive Tag Contents

2 Answers2