Atlassian Confluence：自定义代码宏 - 画笔语法中的分号导致错误

Question

I tried to write a custom code macro syntax for a language that uses semicolon as a comment character.

This shows how to do this: http://alexgorbatchev.com/SyntaxHighlighter/manual/brushes/

My regex looks like this:

{ regex: new RegExp(';(.*)$', 'gm'), css: 'comments' },

But this not only makes the line with the semicolon green, it also does funny things to other lines:

exten => s,n,Hangup becomes exten =>; s,n,Hangup exten =>; s,n,Hangup ( Note: it added a semicolon after the right angle bracket! ) where everything behind the semicolon is green!

Why is this? How ist the right syntax for comments with semicolons?

Answer 1

I know this question is old, but I had the same problem for days now and finally found the solution to fix this, so I wanted to share it with everyone having the same issue.

The Code Block Macro from Confluence uses the <![CDATA[ content ]]> tag to store the code inside the macro in the database. The rendering process then tries to render the page and hands over the CDATA tag and its content to the macro. Neither Confluence nor the macro want < , > and & to be rendered as html tags, so they convert them to the html-code < , > and & .

And now we run in the problem with the RegExp

{ regex: new RegExp(';(.*)$', 'gm'), css: 'comments' },

The expression finds the ; from the html-code elements and wraps them into a code tag with the 'comments' class. Then the matches get put together and you will find the result you got. Because the assembling works with positions and offsets, the > won't get replaced instead the new code tag will just be added.

To fix this, you can just catch the < , > and & before your RegExp so these elements won't be matches twice.

this.regexList = [
{ regex: new RegExp('(\\&lt;|<)|(\\&gt;|>)|(\\&amp;|\\&)', 'gm'), css: 'plain' },
{ regex: new RegExp(';(.*)$', 'gm'), css: 'comments' }];

The same effect happens when you paste a non-breakable-space as the character and not as the html-code   inside the Code-Block-Macro. I couldn't find a way to fix this, because they get converted when used in a normal text editor to a "normal" space.