Japanese PHP forms data entry

andrew972

Hi, we have a web site in English/Japanese and the Japanese portion is not working quite right. I've read several threads here and tried many things, and hoping someone can tell me what we're doing wrong. RedHat 7.3, PHP 4.3.3, MySQL 3.23.55.

Apache/PHP settings as follows:

    php_admin_flag output_buffering on
    php_admin_flag zlib.output_compression off
    php_admin_value output_handler  mb_output_handler
    php_admin_value mbstring.detect_order auto
    php_admin_value mbstring.func_overload  0
    php_admin_value mbstring.http_input   auto
    php_admin_value mbstring.http_output   SJIS
    php_admin_value mbstring.internal_encoding   EUC-JP
    php_admin_value mbstring.substitute_character  none

Problem: FIRST japanese data entry works, but EDITING is turning Japanese into garbage ("mojibake" .. see, I have been reading the threads.. :-).

First time: we pull up our content management form, paste Japanese text into it. Hit "submit" and then go to the web page: the Japanese is correctly displayed. So, the form was correctly putting the Japanese into the MySQL database.

Then, to re-edit the page, we call up the form again. This time, some PHP fetches the Japanese from the database and shows it for editing, i.e. it sets the form's value to the text from the database (same text which is displaying correctly on the web site).

However, the Japanese displayed in the form (just a simple form input type=TEXT), is now garbage. If we hit "submit" to update it again, it also looks like garbage when we refresh the web page.

Somehow, when our form is fetching the Japanese for purposes of Re-Editing, it is becoming garbage. The initial POST seems to be correct.

I would really appreciate anyone's help, and can provide you with links to the site so you can see what is happening.

Thanks,

Andrew

groovything

I dont know if this will work without being able to test it, but i know I have used <![CDATA[$string]]> to resolve similar issues.

It's kind of a shot in the dark, but maybe it will work for you...?

echo "<textarea name='test' cols='30' rows='6'><[CDATA[".$japanese_data."]]></textarea>";

andrew972

Hey, thanks for thee idea.

That didn't work, but it gave me an idea. The input form looked something like this (simplified):

print "<input type='text' name='foo' value=\"".htmlentities($inpVal)."\">\n";

I removed the htmlentities() wrapper and it prevented "garbage" from appearing for one round. If I saved it again, then it again turned into garbage.

Does htmlentities() mess things up? addslashes()?

.... anyone?.. Beuhler.. anyone..?

Thanks,
Andrew

groovything

If you figure it out, please post the solution. I've had similar issues with special characters and never really found a solution that solves them all. I'd love to stop replacing chars. I think if you figure this out it will be a solution that covers a lot of issues.

andrew972

Well, I will definitely post whatever I find out.

But no luck so far. At this point I'd be open to offering up a reward of some sort, maybe a $25 gift certificate to Amazon.com?

(I hope this does not count as a job posting!)

But really, any help on this tough issue would be appreciated.

Thanks,

Andrew

groovything

heres a long shot, but My programmer buddy here suggested it:

Put this in the header of the page:

you might need to put the htmlentities() back in the code.

andrew972

Hey Groove,

I already had this:

I suppose it could be Shift_JIS .. case sensitive..? Worth a tweak.. if it works you'll be the first to know.

Thanks!

Andrew

mooseday

Not much help but I do this using Postgres and have no problem at all.

The header of the input page is set to:

The only conversion I do before UPDATING a record it to AddSlashes() the field. On the display I do nothing

I know this doesn't really help ...

🙁