WordPress “chokes” on Turkey emoji for Thanksgiving

Nope, WordPress does indeed have and support the turkey emoji. Tested on trunk, latest version.

If your browser doesn’t render the turkey, then the core code converts it to this image:
https://s.w.org/images/core/emoji/2.2.1/svg/1f983.svg

Or equivalent pngs and such.

Here’s what it looks like on a test site running 2016:

I can replicate this, on Godaddy Managed WordPress hosting – the emoji are stripped from the post title. Also see the same issue with the Instagram importing plugin.

Could this be related to database collation?
MySQL charset: UTF-8 Unicode (utf8)

Posts table collation is utf8_unicode_ci
Post Meta collation is utf8_unicode_ci
Term Meta collation is utf8mb4_unicode_ci

Other tables are utf8_general_ci or latin1_swedish_ci

Any thoughts @otto42 – what should be collation, could it be connected?

Thanks,

Ian.

If you’re using a recent installation of WordPress, and a recent version of MySQL, then the CHARSET should be “utf8mb4” and the COLLATE should be “utf8mb4_unicode_ci”.

Older installs may have different table layouts. WordPress contains an update routine that updates the relevant tables when you move from a version prior to 4.4 to one later than 4.4.

If your tables are not all properly converted to utf8mb4, then yeah, you probably cannot properly store emoji in them. It should work most of the time due to the compatibility layers, but perhaps not all of the time.

You can convert your tables to utfmb4 manually if you like.

Hi Otto,

Thanks, I’ve now updated the collation and charset.

Error remains.

Post title I’m testing on ‘Today be like… ????’

(looks like the emoji don’t show here, but it’s the caption from this Instagram Image
https://www.instagram.com/p/BO7J6spgmsl/ )

When editing directly via PhpMyAdmin MySQL throws the following error.

Affected rows: 1
Warning: #1300 Invalid utf8 character string: ‘F09F91’
Warning: #1366 Incorrect string value: ‘\xF0\x9F\x91\x99\xF0\x9F…’ for column ‘post_title’ at row 1

Any thoughts?

/

For others looking on the collation update:
I ran this query to update the database defaults:
ALTER DATABASE YourDatabaseName CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci

Then this query to generate the query to update the collation on each table, credit:
https://stackoverflow.com/questions/10859966/how-to-convert-all-tables-in-database-to-one-collation#10860122

SELECT CONCAT(“ALTER TABLE “, TABLE_NAME,” COLLATE utf8mb4_unicode_ci”) AS ExecuteTheString
FROM INFORMATION_SCHEMA.TABLES
WHERE TABLE_SCHEMA=”YourDatabaseName”
AND TABLE_TYPE=”BASE TABLE”

• This reply was modified 8 years, 2 months ago by ianatkins.

looks like the emoji don’t show here

Umm.. I’m seeing emoji in my browser here just fine.

Perhaps this is something in your client not displaying these?

This is bunch of turkey emoji’s. They do indeed show up on this forum:
??????????????????

See them?

In your rather direct method of conversion, you may have some mangled characters in your database, because conversion is more than just an “alter table”. You may need to correct those characters manually.

• This reply was modified 8 years, 2 months ago by Samuel Wood (Otto).