Question

I’m writing a regular expression that should do the following:

== Text ==
Other text
==     Text==

Becomes

<h2>Text</h2>
<p>Other text</p>
<h2>Text</h2>

I’m almost there, the problem is that this is what I currently get:

<h2>Text</h2>
<p>Other text</p>
<h2>Text</h2>
<p></p>

Even though it’s unlikely the heading will not be followed by text, I want to fix it at least for learning purposes.

Here is my function:

preg_replace('/== *(.*?) *==([^=]*)/m', 
             '<h2>$1</h2>
              <p>$2</p>
             ', '== Text ==
                 Other text
                 ==     Text==');

So basically, I want to ignore the <p></p> part if $2 is empty.

Any other tips / improvements are welcomed, I want to learn 🙂

Answer 1

You need one simple conditional to prevent the empty <p> tag from appearing. While I would not recommend this usually, the easiest way to insert this simple if is by using the /e regex modifier to preg_replace:

preg_replace('/== *(.*?) *==([^=]*)/me', 
             '"<h2>$1</h2>".(trim("$2") == ""?"":"<p>$2</p>")',
             '== Text ==
                 Other text
              ==     Text==');

This modifier makes the replacement string be evaluated as PHP code before making the replacement, so you can fit a small conditional in there easily.

See it in action.

Another option would be to use preg_replace_callback, which is effectively the same idea only that you now write the code as a separate function. This is better IMHO because it makes for clearer code.

As a final note, if you intend to add more formatting options you might want to consider breaking your parsing down into multiple steps and possibly processing one line at a time because regular expressions are not designed to handle this kind of processing. You can force it up to a point, but then it starts to become very unmaintainable very quickly.