Problem str_pad function in PHP

The has to do with coding and does not affect the str_pad problem. . You can correct it by setting the utf-8 encoding that would be most appropriate or even using the special encoding for that letter é .

For str_pad remember that in html multiple followed whitespace are ignored, that's why you do not see str_pad working correctly.

If you change the letter of str_pad to any other letter, for example a ! , now see how it is:

Tosolvethespaceproblemyoucanapplyareplacementto immediatelyafterdoingstr_pad:

<?php$nome="Island Avenue";
    $novo = str_pad($nome,30," ",STR_PAD_BOTH);
    $novo = str_replace(" ", "&nbsp;",$novo); //aqui substitui os espaços por nbsp's

    print("Minha nova casa, localizada na $novo &eacute; linda!");
?>

And it already works as expected:

Youcanevencombinethetwofunctionsdirectly:

$nome="Island Avenue";
$novo = str_replace(" ", " ",str_pad($nome,30," ",STR_PAD_BOTH));

str_pad does not work with multi-byte characters, since it does not add all the characters at once.

The original code of str_pad is exactly:

switch (pad_type_val) {
    case STR_PAD_RIGHT:
        left_pad = 0;
        right_pad = num_pad_chars;
        break;

    case STR_PAD_LEFT:
        left_pad = num_pad_chars;
        right_pad = 0;
        break;

    case STR_PAD_BOTH:
        left_pad = num_pad_chars / 2;
        right_pad = num_pad_chars - left_pad;
        break;
}

/* First we pad on the left. */
for (i = 0; i < left_pad; i++)
    ZSTR_VAL(result)[ZSTR_LEN(result)++] = pad_str[i % pad_str_len];

/* Then we copy the input string. */
memcpy(ZSTR_VAL(result) + ZSTR_LEN(result), ZSTR_VAL(input), ZSTR_LEN(input));
ZSTR_LEN(result) += ZSTR_LEN(input);

/* Finally, we pad on the right. */
for (i = 0; i < right_pad; i++)
    ZSTR_VAL(result)[ZSTR_LEN(result)++] = pad_str[i % pad_str_len];

ZSTR_VAL(result)[ZSTR_LEN(result)] = '
const STR_PAD_INSERT_ALL = 4;

function mb_str_pad(string $input, int $pad_length, string $pad_string, int $pad_type, string $pad_encoding = 'utf8') : string {

    $result = '';
    $pad_insert_all = 0;
    $pad_inset_limit = 1;
    $pad_str_len = mb_strlen($pad_string, $pad_encoding);
    $input_len = mb_strlen($input, $pad_encoding);

    if ($pad_length < 0 || $pad_length <= $input_len) {
        return $input;
    }

    if(($pad_type & STR_PAD_INSERT_ALL) === STR_PAD_INSERT_ALL){
        $pad_insert_all = PHP_INT_MAX;
        $pad_inset_limit = null;
        $pad_type -= STR_PAD_INSERT_ALL;
    }

    if ($pad_str_len === 0) {
        trigger_error ( "Padding string cannot be empty", E_WARNING);
        return $input;
    }

    if ($pad_type < STR_PAD_LEFT || $pad_type > STR_PAD_BOTH) {
        trigger_error ("Padding type has to be STR_PAD_LEFT, STR_PAD_RIGHT, or STR_PAD_BOTH", E_WARNING);
        return $input;
    }

    $num_pad_chars = $pad_length - $input_len;
    if ($num_pad_chars >= PHP_INT_MAX) {
        trigger_error ("Padding length is too long", E_WARNING);
        return $input;
    }

    switch ($pad_type) {
        case STR_PAD_RIGHT:
            $left_pad = 0;
            $right_pad = $num_pad_chars;
            break;

        case STR_PAD_LEFT:
            $left_pad = $num_pad_chars;
            $right_pad = 0;
            break;

        case STR_PAD_BOTH:
            $left_pad = floor($num_pad_chars / 2);
            $right_pad = $num_pad_chars - $left_pad;
            break;
    }

    for ($i = 0; $i < $left_pad; $i++){
        $result .= mb_substr($pad_string, ($i % $pad_str_len) &~$pad_insert_all, $pad_inset_limit, $pad_encoding);
    }

    $result .= $input;

    for ($i = 0; $i < $right_pad; $i++){
        $result .= mb_substr($pad_string, ($i % $pad_str_len) &~$pad_insert_all, $pad_inset_limit, $pad_encoding);
    }

    return $result;
}
';

RETURN_NEW_STR(result);

Source.

Note the presence of i % pad_str_len , that is, it only adds a single byte, which can cause an unknown byte to remain. For example, if you are using chr(160) , this is for Latin1 and not for UTF8.

In Latin1, the byte A0 represents "non-breaking space". But the same thing in UTF8 requires two bytes, being C2 A0 . If you cut one of them, for example, by isolating C2 , you will have a ? .

If you want a "new version" of str_pad we could create a mb_str_pad() :

mb_str_pad($nome, 30, "\xc2\xa0", STR_PAD_BOTH, 'utf8');

  

This requires PHP 7 +

This is a version extremely based on the original version of PHP, indicated above, with some changes:

It supports multi-bytes, so you can do:

switch (pad_type_val) {
    case STR_PAD_RIGHT:
        left_pad = 0;
        right_pad = num_pad_chars;
        break;

    case STR_PAD_LEFT:
        left_pad = num_pad_chars;
        right_pad = 0;
        break;

    case STR_PAD_BOTH:
        left_pad = num_pad_chars / 2;
        right_pad = num_pad_chars - left_pad;
        break;
}

/* First we pad on the left. */
for (i = 0; i < left_pad; i++)
    ZSTR_VAL(result)[ZSTR_LEN(result)++] = pad_str[i % pad_str_len];

/* Then we copy the input string. */
memcpy(ZSTR_VAL(result) + ZSTR_LEN(result), ZSTR_VAL(input), ZSTR_LEN(input));
ZSTR_LEN(result) += ZSTR_LEN(input);

/* Finally, we pad on the right. */
for (i = 0; i < right_pad; i++)
    ZSTR_VAL(result)[ZSTR_LEN(result)++] = pad_str[i % pad_str_len];

ZSTR_VAL(result)[ZSTR_LEN(result)] = '
const STR_PAD_INSERT_ALL = 4;

function mb_str_pad(string $input, int $pad_length, string $pad_string, int $pad_type, string $pad_encoding = 'utf8') : string {

    $result = '';
    $pad_insert_all = 0;
    $pad_inset_limit = 1;
    $pad_str_len = mb_strlen($pad_string, $pad_encoding);
    $input_len = mb_strlen($input, $pad_encoding);

    if ($pad_length < 0 || $pad_length <= $input_len) {
        return $input;
    }

    if(($pad_type & STR_PAD_INSERT_ALL) === STR_PAD_INSERT_ALL){
        $pad_insert_all = PHP_INT_MAX;
        $pad_inset_limit = null;
        $pad_type -= STR_PAD_INSERT_ALL;
    }

    if ($pad_str_len === 0) {
        trigger_error ( "Padding string cannot be empty", E_WARNING);
        return $input;
    }

    if ($pad_type < STR_PAD_LEFT || $pad_type > STR_PAD_BOTH) {
        trigger_error ("Padding type has to be STR_PAD_LEFT, STR_PAD_RIGHT, or STR_PAD_BOTH", E_WARNING);
        return $input;
    }

    $num_pad_chars = $pad_length - $input_len;
    if ($num_pad_chars >= PHP_INT_MAX) {
        trigger_error ("Padding length is too long", E_WARNING);
        return $input;
    }

    switch ($pad_type) {
        case STR_PAD_RIGHT:
            $left_pad = 0;
            $right_pad = $num_pad_chars;
            break;

        case STR_PAD_LEFT:
            $left_pad = $num_pad_chars;
            $right_pad = 0;
            break;

        case STR_PAD_BOTH:
            $left_pad = floor($num_pad_chars / 2);
            $right_pad = $num_pad_chars - $left_pad;
            break;
    }

    for ($i = 0; $i < $left_pad; $i++){
        $result .= mb_substr($pad_string, ($i % $pad_str_len) &~$pad_insert_all, $pad_inset_limit, $pad_encoding);
    }

    $result .= $input;

    for ($i = 0; $i < $right_pad; $i++){
        $result .= mb_substr($pad_string, ($i % $pad_str_len) &~$pad_insert_all, $pad_inset_limit, $pad_encoding);
    }

    return $result;
}
';

RETURN_NEW_STR(result);

Differences from the original version:

PS: Assuming I did not insert a bug.

• Multi-byte support:

  It supports characters that require multiple bytes. You can specify the encoding type used, including UTF8, which is the default.

• A new "STR_PAD_INSERT_ALL":

  You can insert the whole string, instead of "switching to one another", if you have a string with more than one character (eg "abc"), you can specify to always insert "abc" this has a side effect since the number of characters entered is not measured. To use, just use STR_PAD_BOTH | STR_PAD_INSERT_ALL , but this is not necessary in YOUR CASE.

• Return on error:

  Even in case where a WARNING is issued it will return the original string, which is not the behavior of the original function.