To make these even easier to access, I needed a way to get these into the wiki. It would be easiest if I could just copy and paste the HTML contents onto a wiki page, but MediaWiki can’t quite process all the HTML in that file so I needed a way to clean it up a bit.

Take a look at a single line that contains a release note:

<tr class="spacer"><td class="reportid"><a href="feedback://showreport?report_id=52522">52522</a></td><td class="category">Framework » All</td><td class="desc">Added AntiAliasMode property on Graphics class. This property controls the level of interpolation/quality when drawing scaled Pictures. Valid modes are from the Graphics.AntiAliasModes enumeration: LowQuality, DefaultQuality, and HighQuality. The default is DefaultQuality.</td></tr>

This is a row in an HTML table. The problem is the <a href> part which MediaWiki does not use for displaying links. So I needed a way to clean that up. There are also some <tbody> tags that I needed to remove.

Normally I’d crack my knuckles and start on a Xojo project and use string find/replacing to massage the text. But this is messy and tedious.

So I thought, why not try Regular Expressions (RegEx)? I don’t really know much about RegEx and frankly they scare me a bit. After all, who can make heads or tails of something that looks like this:

\b(\d{1,2})([/-])(\d{1,2})\2(\d{4}|\d{2})\b

That gibberish above is called a RegEx pattern and apparently that one is for date validation.

Anyway, I figured I would give RegEx a try. My first challenge was to find a pattern that I could start with and maybe modify. Heading to the Xojo doc pages for the RegEx class, I found this pattern to remove HTML tags from text on the RegEx.Replace page’s Sample Code:

<[^<>]+>

Since I did need to remove HTML tags, I figured it was a good place to start.

To explain this pattern:

1. The “<” is the starting character
2. The “[” bracket starts a character class
3. The “^” means not, so “^<>” means all characters except “<” and “>”
4. The “]” ends the character class and the “+” means match 1 or more characters
5. End with a “>” character

So that means match all the text that starts with a “<” has one or more characters after that except “<” and “>” and ends with a “>”.

FYI: For all my RegEx testing I used RegExRX — an excellent tool for dealing with RegEx written in Xojo by long-time Xojo developer Kem Tekinay.

After starting RegExRX I put the above pattern in the Search Pattern field and pasted a small snippet of the HTML into the Source Text field. I could immediately see that all the HTML tags were matched.

But I did not need all the HTML tags to be matched, I only needed <a> and <tbody>. A slight modification to the above pattern can tell it to only find text that starts with “<a” or “</a”. Looking at the RegEx reference on the RegEx page I saw that the “?” character means 0 or 1 matches. So I can change the pattern to this:

</?a[^<>]*>

Which means find HTML tags that start with “<a” or “</a”. With this pattern I can see that only the <a> and </a> tags are highlighted:

To actually remove the tags I use a Replacement and replace an empty string for the matched tag. Switching to the Replace tab in RegExRX, this is what that looks like:

To do the same things for <tbody> I changed the pattern to this:

</?tbody[^<>]*>

Now it was time to use this in a Xojo project. You use the RegEx class to work with regular expressions. This function takes text (loaded from a release notes file) and applies the search pattern and an empty replace pattern to remove the tags and stick the result into the Clipboard so I can paste it into the wiki:

Private Function CleanHTML(html As String) as String
  Dim re As New RegEx
  re.SearchPattern = "</?a[^<>]*>" // Find <a href></a> tags for removal
  re.ReplacementPattern = ""
  re.Options.ReplaceAllMatches = True
  Dim plainHTML As String = re.Replace(html)

  // Remove any <tbody> tags as MediaWiki doesn't process them
  re.SearchPattern = "</?tbody[^<>]*>"
  re.ReplacementPattern = ""
  re.Options.ReplaceAllMatches = True
  plainHTML = re.Replace(plainHTML)

  Dim c As New Clipboard
  c.Text = plainHTML

  Return plainHTML
End Function

This worked wonderfully, but after trying this out I decided I wanted to keep the links to the Feedback case around so that you can click on a case number to open the case in the Feedback app and read all its history. This means that instead of removing the <a> tags I needed to change them from this:

<a href="feedback://showreport?report_id=52522">52522</a>

to this:

[http://feedback.xojo.com/case/52522 52522]

To do this I now needed to use a subgroup to save the case ID so I could use it as part of the replacement string. To create subgroups you group parts of the RegEx pattern using parentheses. Essentially I wanted to have clear parts of the pattern for the <a> start tag, the value (case ID) and the </a> closing tag.

I started by simplifying the <a> tag search to this to just match the opening tag:

<a[^<>]*>

I then added a part to match the case ID which is just a series of 1 or more numbers. RegEx has a command to match only digits, which is “\d”, to which we can add the “+” to to repeat it one or more times as needed. I then wrapped that in parentheses to get a group resulting in this:

<a[^<>]*>(\d+)

Lastly I added the part to match the closing tag </a> to get the final pattern:

<a[^<>]*>(\d+)</a>

In the end you can see the <a>, </a> tags and the case ID value are all matched:

This changed pattern means I now have a group I can use for the replacement. Switching to the Replace tab in RegExRX you’ll notice that the entire matched text is removed because the Replace Pattern is blank. Typing “$1” (this contains the group with the Case ID value) in the Replace Pattern showed the Case ID in the replaced text:

And now I put the rest of the text I wanted for the Replace Pattern:

[feedback://showreport?report_id=$1 $1]

The replaced text now looks like what I wanted:

Here is the updated Xojo function:

Private Function CleanHTML(html As String) as String
  Dim re As New RegEx
  re.SearchPattern = "<a[^<>]*>(\d+)</a>" // Find <a href> tags and save case # as a group
  re.ReplacementPattern = "[http://feedback.xojo.com/case/$1 $1]" // Swap in wiki link format with correct URL
  re.Options.ReplaceAllMatches = True
  Dim plainHTML As String = re.Replace(html)

  // Remove any <tbody> tags as MediaWiki doesn't process them
  re.SearchPattern = "</?tbody[^<>]*>"
  re.ReplacementPattern = ""
  re.Options.ReplaceAllMatches = True
  plainHTML = re.Replace(plainHTML)

  Dim c As New Clipboard
  c.Text = plainHTML

  Return plainHTML
End Function

I hope this little adventure in RegEx has helped you appreciate how wonderful they are for string searching and replacement. I’m still no expert, but I found this to be much, much better than messy string searching and parsing using InStr and friends.

To learn more about Regular Expressions, check out the RegExOne site.