Beautiful HTML Ticket Posts

We have run across this ourselves, as well as a few other bugs. I am not actively maintaining this project, so I will just list the bunch of fixes (for this particular issue as well as other various issues we have discovered for ourselves).

Line numbers are for the original code I posted in this forge project. I apologize for the weird formatting of the code below - I am not sure why the code tag is not working properly.

Around line 815:

@

// process data between tags
                if (!preg_match('#(.*)<#U', $_contents, $_matches, 0, $_offset))
{
                    $_offset = $_length;
                    return;
                }
@

Becomes:

@

// process data between tags
                if (!preg_match('#(.*)<#U', $_contents, $_matches, 0, $_offset))
{
                    $_currentTag->_data[] = substr($_contents, $_offset);
                    $_offset = $_length;
                    return;
                }
@

Around line 1084:

@

// marvin: nuke trailing whitespace tags
    private static function TrimTrailingWhitespace(HtmlTag $_currentTag)
{
        $_lastTag = self::FindLastWhitespaceTag($_currentTag);

while (($_lastTag instanceof HtmlTag) && ($_lastTag->_name != 'img'))
{
            self::RemoveTag($_lastTag);

$_lastTag = self::FindLastWhitespaceTag($_lastTag->_parentTag);

if ($_lastTag === $_currentTag)
{
                break;
            }
        }
    }
@

Becomes:

@

// marvin: nuke trailing whitespace tags
    private static function TrimTrailingWhitespace(HtmlTag $_currentTag)
{
        if (!empty($_currentTag->_data))
    {
            $_lastTag = self::FindLastWhitespaceTag($_currentTag);

while (($_lastTag instanceof HtmlTag) && ($_lastTag->_name != 'img'))
{
                self::RemoveTag($_lastTag);

$_lastTag = self::FindLastWhitespaceTag($_lastTag->_parentTag);

if ($_lastTag === $_currentTag)
{
                    break;
                }
            }
        }
    }
@

Around line 1172 ADD:

@

// marvin: verify the content is in fact html
        if ($_isContentHTML)
{
            if (strpos($_contents, "<") === false) {
                $_isContentHTML = false;
            }
        }

// marvin: process the html so we have a nice clean valid unbroken html for display in the staff cp and for history in emails
@

Around line 1185:

@

$_contents = str_replace('<newline-was-here>', ' ', $_contents);

// parse the (potentially broken) original html into a tree
@

Becomes:

@

$_contents = str_replace('<newline-was-here>', ' ', $_contents);

// try to find the breakline (mail parser may not have been able to get it)
            if (strpos($_contents, '<!-- breaklines were processed -->') !== false)
{
                $_splitContents = preg_split('#=+\splease\sreply\sabove\sthis\sline\s=+#i', $_contents);

if (count($_splitContents) > 1)
{
                    $_contents = $_splitContents[0];
                    $_contents .= '<!-- breakline was here -->';
                }
            }

// parse the (potentially broken) original html into a tree
@

These should solve your problem as well as a few other problems I have discovered over the past year.