mb_convert_encoding vs utf8_encode()

Question

I did an update on php on the server and identified that they were not being coded in the standard utf-8, the first thing I checked was the connection class I use, which in this case is adodb.

In my connection I perform a process of converting the columns that are in another format to utf-8 in this way:

 $dados[$i]  = mb_convert_encoding($dados[$i],"UTF-8");

In the php 5.2 was working normally, after updating to the php 5.6.9 he "stopped working," I checked his documentation is the same is not depreciated.

To solve the problem I used the utf8_encode, thus:

 $dados[$i]  = utf8_encode($dados[$i]);

Detail, I use adodb + firebird.

Doubts:

1° What is the difference of these functions, and why mb_convert_encoding not working anymore.

2° The return of my database is on ASCII it is possible to simplify this issue of codification utf-8 forASCII and ASCII for utf-8?

Man php.ini comes set by default default_charset = "UTF-8" // linha 680

Testing:

$f[$i] = mb_detect_encoding($f[$i]); //  ASCII

$f[$i] = mb_convert_encoding($f[$i], "HTML-ENTITIES", "UTF-8");
SUÉLLEM // Entrada
SU�LLEM // Saida

$f[$i] = mb_convert_encoding($f[$i], "ISO-8859-1", "UTF-8");
SUÉLLEM // Entrada
SU?LLEM // Saida

$f[$i] = mb_convert_encoding($f[$i], 'UTF-8', 'ISO-8859-1');
SUÉLLEM // Entrada
SUÉLLEM // Saida

In the third test it works, how strange!

Example of the method _fetch of adodb:

   function _fetch() {
     $f = @ibase_fetch_row($this - > _queryID);
     if ($f === false) {
       $this - > fields = false;
       return false;
     }
     // OPN stuff start - optimized
     // fix missing nulls and decode blobs automatically

     global $ADODB_ANSI_PADDING_OFF;
     //$ADODB_ANSI_PADDING_OFF=1;
     $rtrim = !empty($ADODB_ANSI_PADDING_OFF);
     for ($i = 0, $max = $this - > _numOfFields; $i < $max; $i++) {
       if ($this - > _cacheType[$i] == "BLOB") {
         if (isset($f[$i])) {
           $f[$i] = $this - > connection - > _BlobDecode($f[$i]);
         } else {
           $f[$i] = null;
         }
       } else {
         if (!isset($f[$i])) {
           $f[$i] = null;
         } else if ($rtrim && is_string($f[$i])) {
           $f[$i] = rtrim($f[$i]);
         }
       }

       $f[$i] = utf8_encode($f[$i]);

     }
     // OPN stuff end

     $this - > fields = $f;
     if ($this - > fetchMode == ADODB_FETCH_ASSOC) {
       $this - > fields = $this - > GetRowAssoc(ADODB_ASSOC_CASE);
     } else if ($this - > fetchMode == ADODB_FETCH_BOTH) {
       $this - > fields = array_merge($this - > fields, $this - > GetRowAssoc(ADODB_ASSOC_CASE));
     }
     return true;
   }

Answer 1

Simply and quickly:

mb_convert_encoding converts an X encoding to a Y encoding.

utf8_encode encodes the string ISO-8859-1 for UTF-8.

Note then that the difference between the two is very wide.

5.5.9-1ubuntu4.11 A small example of encoding conversion UTF-8 for HTML-ENTITIES with mb_convert_encoding:

$str = 'É assim que você faz, na programação';

$converted = mb_convert_encoding($str, "HTML-ENTITIES", "UTF-8");

var_dump($converted, $str);

The exit is:

string(63) "É assim que você faz, na programação"

string(40) "É assim que você faz, na programação"

Settings for default_charset

I don’t know if that’s pertinent, but the configuration of my default_charset php.ini influenced when displaying the previous test result.

Look what happened to the default_charset defined as ISO-8859-1:

ini_set('default_charset', 'ISO-8859-1');

$str = 'É assim que você faz, na programação';

$converted = mb_convert_encoding($str, "HTML-ENTITIES", "UTF-8");


var_dump($converted, $str);

The exit was:

string(63) "É assim que você faz, na programação"
string(40) "É assim que você faz, na programação"

So, for you to check if this is the problem presented in the first question, then try to define your default_charset thus:

ini_set('default_charset', 'UTF-8');

Note: This is the version of my PHP, where I did the tests: 5.5.9-1ubuntu4.11