Álvaro G. Vicario - 2008-12-02

I've read the docs and searched the forums but I'm still not fully sure of how charsets work.

My scripts use ISO-8859-15 (single byte) so I've set $unicode=FALSE and $charset='ISO-8859-15' in the TCPDF constructor. I've written this test code:

$pdf->writeHTML('<p>Character 151: ' . chr(151) . ' (dash in cp1252, control char in ISO-8859-1 and ISO-8859-15)</p>');
$pdf->writeHTML('<p>Character 128: ' . chr(128) . ' (euro in cp1252, control char in ISO-8859-1 and ISO-8859-15)</p>');
$pdf->writeHTML('<p>Character 164: ' . chr(164) . ' (diaeresis in cp1252, currency in ISO-8859-1 and euro in ISO-8859-15)</p>');
$pdf->writeHTML('<p>Entity &amp;euro;: &euro;</p>');
$pdf->writeHTML('<p>Entity &amp;#8364;: &#8364;</p>');
$pdf->writeHTML('<p>Entity &amp;#x20AC; &#x20AC;</p>');

I've run it with times, helvetica and courier and I get the same (inconsistent) result:

Character 151: — (dash in cp1252, control char in ISO-8859-1 and ISO-8859-15)
Character 128: € (euro in cp1252, control char in ISO-8859-1 and ISO-8859-15)
Character 164: ¤ (diaeresis in cp1252, currency in ISO-8859-1 and euro in ISO-8859-15)
Entity &euro;: ¤
Entity &#8364;: &#8364;
Entity &#x20AC; &#x20AC;

That is, the output sometimes uses the cp1252 encoding (dash and plain euro) and sometimes uses ISO-8859-15 (currency).

The code in example_006.php sets $unicode=TRUE and no $charset and works fine in my system but it's using 7-bit ASCII data. If I do the same in my code and use iconv() to get UTF-8 out of my ISO-8859-15 I still get inconsistent results.

My questions:

  1. Is it a bug?
  2. Is the issue font related? Do I need to rebuild the included fonts?
  3. Will these chars look the same in all computers?

Please apologize me if I look too picky but I'll be generating PDF files from user input.