Allow emoji in contribution names

[ECYaz]

Picks up #390, which has been sitting as a draft since 2022. The diagnosis there is @battye's, and the approach here is @iMattPro's suggestion on that thread.

A four byte character in a contribution name reaches the utf8 column as-is and MySQL rejects the row, on create and on edit both:

Incorrect string value: '\xF0\x9F\x86\x95 E...' for column 'contrib_name' at row 1 [1366]

utf8_encode_ucr() replaces those characters with their numeric character reference, the same thing core does to report text in phpbb\report\controller\report. Encoding in titania_contribution::submit() rather than in the controllers covers create and edit together, and keeps the stored form out of everything that reads the name, so nothing has to decode it.

That last part is why this does not take the approach in #390. Parsing the name leaves s9e markup in the field, which then reaches the queue topic subject, and @battye ran into exactly that in June 2024:

Incorrect string value ... for column 'topic_subject' at row 1 [1366]
'Validation - <r>Test edit: <EMOJI seq="1f195"...'

A character reference has neither problem. Dropping them when generating a permalink keeps the digits out of the URL, otherwise the slug comes out as cool_127381;_extension.

Checked on a local board:

• creating and editing with an emoji in the name both succeed
• the release topic subject stores fine and stays within the column
• the permalink comes out as cool_extension
• four byte characters outside the emoji range are handled too, a rare CJK ideograph encodes the same way
• accented characters are untouched, only four byte sequences are touched
• re-saving does not encode twice, and names stored before this change are unaffected

#439 fixes the attention fields, which fail the same way, using the same function.

If @battye would rather carry this on #390 then I am happy for this to be closed in favour of it.

[iMattPro]

This still does not work because the un-encoded emoji is still trying to be stored to the contrib_name_clean field

SQL ERROR [ mysqli ]

Incorrect string value: '\xF0\x9F\x86\x95' for column 'contrib_name_clean' at row 1 [1366]

SQL

INSERT INTO cdb_contribs (contrib_id, contrib_type, contrib_name, contrib_name_clean, contrib_desc, contrib_desc_bitfield, contrib_desc_uid, contrib_desc_options, contrib_categories, contrib_demo, contrib_status, contrib_user_id, contrib_downloads, contrib_views, contrib_visible, contrib_rating, contrib_rating_count, contrib_creation_time, contrib_last_update, contrib_release_topic_id, contrib_faq_count, contrib_iso_code, contrib_local_name, contrib_clr_colors, contrib_limited_support, contrib_package_name) VALUES (0, 8, 'Emoji 🆕', 'emoji_🆕', 'Emoji <EMOJI seq="1f195" tseq="1f195">🆕', '', '', 7, '18', '', 1, 2, 0, 0, 1, 0, 0, 1785159808, 1785159808, '', '', '', '', '', 0, '')

However, I'm starting to think we should not allow emoji in contrib names. Emoji should only be allowed in descriptions. Contrib names should remain alpha-numeric IMO.

[ECYaz]

Good catch. The permalink is generated in validate(), which runs before submit() encodes, so the slug was still built from the raw name.

Now dropping four byte characters and character references in generate_slug() instead, which also covers the topic subject slug.

[iMattPro]

Check out #441 as a more complete solution, possibly