11 replies to this topic

[Janj]

Welp, I've encountered an error this time... Certain characters don't seemed to be displayed when I echo the data in PHP file. For example, the mdash (—) and the a with a tilde (which is necessary for certain places such as São Paulo). These show up as �. Is there any way to get them to show normally? I've tried surrounding them with `'s and putting \'s in front of them, but I can't seem to find out how to get them to work.

[Jessica]

You'll need to have your character encoding set to UTF-8, both in MySQL and your PHP generated HTML.

[Janj]

You'll need to have your character encoding set to UTF-8, both in MySQL and your PHP generated HTML.

Thanks!

[Janj]

Okay, I changed the collation to utf8_unicode_ci. This changes the encoding set to UTF-8, right? For some reason I'm still getting �'s. Is there another way to do it or something?

[Jessica]

If the table wasn't already UTF-8 when you inserted the data, I don't know that you can convert it back. 

[Janj]

I tried exporting and importing the table.  I also went in and redid a row, and it still shows up wrong.

[DavidAM]

You'll need to have your character encoding set to UTF-8, both in MySQL and your PHP generated HTML.

[Janj]

I know my HTML has to be UTF-8. Does that mean I need all my PHP in the body (since the HTML that makes the charset utf-8 is in the head)?

Thanks, I think this is the last question!

[Christian F.]

The functions you use in PHP to manipulate the data have to be UTF-8 compatible as well, yes.
Though this is only true for the functions that actually manipulate the string data that contains UTF-8 characters, it's better to treat all strings as if they were UTF-8. That way you are ensured that you will not encounter any future bugs, related to differing charsets.

That said, I suspect there are one or two places where you've might have forgotten to define the proper charset:

• On the database connection, with *_set_charset ().
• In the HTTP headers themselves, with the proper header () call.

The latter one is necessary because the meta-HTML headers are ignored by the browser, if they conflict with a HTTP header. So if your web-server is set to serve HTML as ISO-8859-1 by default (which a lot of them are), then your "meta http-equiv" HTML header is quite ineffective.

I also recommend reading the article "The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)". As it says: It's the absolute minimum, and no excuses!

[Janj]

The functions you use in PHP to manipulate the data have to be UTF-8 compatible as well, yes.
Though this is only true for the functions that actually manipulate the string data that contains UTF-8 characters, it's better to treat all strings as if they were UTF-8. That way you are ensured that you will not encounter any future bugs, related to differing charsets.

That said, I suspect there are one or two places where you've might have forgotten to define the proper charset:

• On the database connection, with *_set_charset ().
• In the HTTP headers themselves, with the proper header () call.

The latter one is necessary because the meta-HTML headers are ignored by the browser, if they conflict with a HTTP header. So if your web-server is set to serve HTML as ISO-8859-1 by default (which a lot of them are), then your "meta http-equiv" HTML header is quite ineffective.

I also recommend reading the article "The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)". As it says: It's the absolute minimum, and no excuses!

Thanks.  I tried this:

/* change character set to utf8 */
if (!mysqli_set_charset($con, "utf8")) {
    printf("Error loading character set utf8: %s\n", mysqli_error($con));
} else {
    printf("Current character set: %s\n", mysqli_character_set_name($con));
}

to test, and the output was "Error loading character set utf8:"  I don't understand.

[Christian F.]

Me using the * in the function name like that, and linking to the MySQLI version, was a subtle hint to tell you that:

• I don't know what library you're using to communicate with the database.
• And in any case, you really should be using MySQLI or PDO.

Since you got the error message, I'm assuming that you're still using the old and outdated mysql library. I recommend updating your scripts to use the current libraries, before the old one is completely removed from PHP. Alternatively, use the old mysql_* version of the function, in the meantime.

[Janj]

I found out that putting the HTML code for the character (such as  ) works without messing up.  Thanks for your help, I learned something!