用正则表达式或替代方法替换唯一的嵌套语句

Question

From the countless questions posted I know it's not possible/advisable to use regex to replace nested statements.

I'm wondering if it makes any difference in a case where statements are unique:

[if @test]TEST[if @second]SECOND[/if][/if]

I've gotten it work when the end blocks are also unique, which I know is clumsy workaround:

[if @test]TEST[if @second]SECOND[/if @second][/if @test]
$pattern = '%\[if @'.$dynamic.'.*?\](.*?)\[/if @'.$dynamic.'\]%s'; //Works with above

Is it possible to use regex without the end block being unique? Are there alternatives to regex that would accomplish this?

I would like to parse something like: [if @test]TEST[if @second]SECOND[/if][/if] with arbitrary nesting levels. If regex is not practical, can anyone suggest viable alternative in PHP?

Answer 1

In a proper solution you should tokenize the string in to its basic components such as tags, comments, text and whatever else you have there. This step can be done with regex, and produces a flat list of tokens. Next you go trough the tokens building a parse tree with all the structure and details needed. (Both steps can be combined and done in one pass as well.)

That way everything is under your control and you don't need to reparse any part of the code.

On the other hand it can be done with regex, but then you are more limited, and you need to reparse the nested parts of the code for every added depth.

Since you asked for a regex, here is one to match such nested ifs:

~
\[if\ @(\w++)]
(
    (?>
        (?: (?!\[if\ @\w++]|\[/if]) . )++
        |
        (?R)
    )*+
)
\[/if]
~xs