Need Help with a regular expression not replacing all instances of an expression

Question

Been banging my head against the wall on this one all day and am getting close to my wits end on this. Looking for some fresh perspective.

Sample Input Text:
(line breaks added for clarity, not in actual data )

</div>#My Novel<br />  
##Chapter1<br />  
It was a dark and stormy night<br />
##Chapter 2<br />
The End

Desired Output

</div><h1>My Novel</h1><br />
<h1>Chapter1</h1><br />  
It was a dark and stormy night<br />  
<h1>Chapter 2</h1><br />  
The End

Actual Output

</div><h1>My Novel</h1><br />
##Chapter1<br />  
It was a dark and stormy night<br />  
<h1>Chapter 2</h1><br />  
The End

Here is the match expression
(formatted for easy reading, comments/linebreaks are not in expression)

(?<preamble>
    (                             
     ([<]\/\w+\d*[>])|([<]\w+\d*\s*\/[>])   #</tag> or <tag />
    )
    \s*  #optional whitespace              
)

(?<hashmarks>
    \#{1,6}      #1-6 hash marks
)    

(?<content>
    .+?          #header content
 )   

(?<closing>
    ([<](br|\/\s*br|br\s*\/)[>])   #<br>,</br>, or <br />
)

Here is the replace Expression

${preamble}<h1>${content}</h1>${closing}

If it matters I am using the following C# regex.replace overload:

Regex.Replace(Source,SrchExp,ReplExpr,RegexOptions.IgnoreCase)

The question (finally)
Can anyone see why it is replacing #My Novel and ##Chapter 2, but not ##Chapter 1?

Sorry for the long post, and hopefully I didn't munge anything trying to format it to make it readible for SO.

Update:

One more thing that might help. Adding an extra break tag right after "Novel" makes the provided code start working perfectly. No idea why yet.

Sample Input Text (modified):

</div>#My Novel<br /><br />
##Chapter1<br />  
It was a dark and stormy night<br />
##Chapter 2<br />
The End

Answer 1

Here's one that was actually tested and appears to work.

The issue is that once a match is found, the search continues exactly where the first one left off. As a result, the closing <br /> of #My Novel will not be captured again, and so #Chapter1 is missed.

To capture #Chapter1-like constructs anyway, we can use a lookbehind assertion. Lookbehinds enforce the presence of the prefix, even if it extends before the current position. This also prevents the need to drop it in the replacement string:

• Replace (?<preamble> with (?<=

• Then in the replacement string, remove the ${preamble} portion.

The overall search expression now looks like:

(?<=             # removed the preamble capture and replaced with a lookbehind
    (                             
        ([<]\/\w+\d*[>])|([<]\w+\d*\s*\/[>])   #</tag> or <tag />
    )
    \s*  #optional whitespace                               
)

(?<hashmarks>
    \#{1,6}      #1-6 hash marks
)    

(?<content>
    .+?          #header content
 )      

(?<closing>
    ([<](br|\/\s*br|br\s*\/)[>])   #<br>,</br>, or <br />
)

And the replacement string looks like:

<h1>${content}</h1>${closing}

Our output is now faithfully:

</div><h1>My Novel</h1><br />
<h1>Chapter1</h1><br />
It was a dark and stormy night<br />
<h1>Chapter 2</h1><br />
The End