#887 HTML parsing error with Р (ER) character

[Alexander Vasilyev]

Hi,
I encountered issue on parsing HTML code with writeHTML.

Example how to reproduce:

require_once 'tcpdf/tcpdf.php';

$pdf = new TCPDF(PDF_PAGE_ORIENTATION, PDF_UNIT, PDF_PAGE_FORMAT, true, 'UTF-8', false);

$pdf->AddPage();
$pdf->SetFont('freeserif', '', 12);

$content = '<div>Test: Р 123</div>';

$pdf->writeHTML($content);
$pdf->Write(0, $content);

$pdf->Output('test.pdf', 'F');

It works fine and outputs 2 lines (HTML and plain) to PDF.

But if change $content:

$content = '<div>Test: Р</div>';

then TCPDF outputs only second (plain text) line.

Looks like the problem is the last character "Р" (Unicode: U+0420, UTF-8: D0A0, CYRILLIC CAPITAL LETTER ER) before closing </div>.

[Nicola Asuni]

The following code seems working fine with the freesans font:

$content = '

Test: Ƥ 123

';
$content = '

Test: Ƥ

';

However the freeserif font seems to have some problems to render on Firefox when the subsetting is turned on, so you can turn it off for now:

$pdf->SetFont('freeserif', '', 12, '', false);

I'll investigate if there is a font issue or not.

[Alexander Vasilyev]

Unfortunately it does not help.

Check the code below:

require_once 'tcpdf/tcpdf.php';

$pdf = new TCPDF(PDF_PAGE_ORIENTATION, PDF_UNIT, PDF_PAGE_FORMAT, true, 'UTF-8', false);
$pdf->setFontSubsetting(false);

$pdf->AddPage();

// Default font
$pdf->Write(0, 'Test 1: Р');
$pdf->writeHTML('Test 2: Р');
$pdf->Write(0, '<div>Test 3: Р</div>');
$pdf->writeHTML('<div>Test 4: Р</div>'); // No output

$pdf->SetFont('FreeSerif', '', 12, '', false);
$pdf->Write(0, 'Test 5: Р');
$pdf->writeHTML('Test 6: Р');
$pdf->Write(0, '<div>Test 7: Р</div>');
$pdf->writeHTML('<div>Test 8: Р</div>'); // No output

$pdf->SetFont('DejaVuSans', '', 12, '', false);
$pdf->Write(0, '<div>Test 9: Р</div>');
$pdf->writeHTML('<div>Test 10: Р</div>'); // No output
$pdf->writeHTML('<div>Test 11: М</div>'); // Outputs "Test 11: М"

$pdf->Output('test.pdf', 'F');

Sreenshot from Adobe Reader 11 (TCPDF v.6.0.058) attached. No Firefox used.

[Nicola Asuni]

I am unable to reproduce your problem.
Maybe your letter is not properly encoded?
Try to use the html entity instead for the Ƥ letter:
&#x01A4;

[Alexander Vasilyev]

According http://htmlentities.net "Р" letter is &#1056; not Ƥ. In origin it is a normal capitalized cyrillic letter "Р" (pronunciation: ER). Other cyrillic letters placed before </div> outputs as normal. But "Р" fails.

$pdf->writeHTML('<div>Test A: &#1056;</div>'); // Outputs: Test A: Р
$pdf->writeHTML('<div>Test B: Р</div>'); // No output

Very sad.

[Nicola Asuni]

Seems that you are trying to use the wrong character:
http://www.fileformat.info/info/unicode/char/1056/index.htm
instead of:
http://www.fileformat.info/info/unicode/char/420/index.htm

420 DEC = 1A4 HEX

[Alexander Vasilyev]

Try to use the html entity instead for the Ƥ letter:

Did you try it yourself?

$pdf->writeHTML('<div>Test A: &#x01A4;</div>');

Outputs: Test A: &#x01A4;

HTML entity is not Unicode.

See http://www.fileformat.info/info/unicode/char/420/index.htm — Section HTML Entity (decimal) = &#1056;

Then:

$pdf->writeHTML('<div>Test B: &#1056;</div>');

Outputs: Test A: Р (as needed result)

But:

$pdf->writeHTML('<div>Test B: Р</div>');

Outputs nothing.

Look at the bug hack below:

class MyTCPDF extends TCPDF
{
    public function writeHTML ($html, $ln = true, $fill = false, $reseth = false, $cell = false, $align = '')
    {
        // Bug hack
        // @see: https://sourceforge.net/p/tcpdf/bugs/887/
        $html = str_replace('Р</div>', '&#1056;</div>', $html); // Replace "Р" (CYRILLIC CAPITAL LETTER ER) for HTML Entity (decimal)
        parent::writeHTML($html, $ln, $fill, $reseth, $cell, $align);
    }
}

Maybe it helps somebody to save time.

[Alexander Vasilyev]

Just corrected replace code to support of any closing tag:

$html = str_replace('Р</', '&#1056;</', $html); // Replace "Р" (CYRILLIC CAPITAL LETTER ER) for HTML Entity (decimal)

[Nicola Asuni]

Please do NOT reply on a closed ticket, instead use the Help forum.
Your document seems to be bad encoded since the html entities are accepted.
&#1056; and &#x01A4; are equivalent since one is in decimal and the other in hexadecimal representation.