Research better solution for detecting languages with different word count

[Zodiac1978]

We are using a core string to detect languages with different word counts (like asian langauges) in:

antispam-bee/antispam_bee.php

Lines 1860 to 1873 in b21d41f

	/*
	* translators: If your word count is based on single characters (e.g. East Asian characters),
	* enter 'characters_excluding_spaces' or 'characters_including_spaces'. Otherwise, enter 'words'.
	* Do not translate into your own language.
	*/
	if ( strpos( _x( 'words', 'Word count type. Do not translate!' ), 'characters' ) === 0 && preg_match( '/^utf\-?8$/i', get_option( 'blog_charset' ) ) ) {
	preg_match_all( '/./u', $text, $words_array );
	if ( isset( $words_array[0] ) ) {
	$word_count = count( $words_array[0] );
	}
	} else {
	$words_array = preg_split( "/[\n\r\t ]+/", $text, -1, PREG_SPLIT_NO_EMPTY );
	$word_count = count( $words_array );
	}

There are at least two problems with this approach:

• If the core string changes we do need to change this test accordingly. If we do not have any tests for this to check it automatically, this will easily slip through.
• The string also will be parsed by the PO-Parser from WordPress.org and is therefore on translate.wordpress.org. I think this will be not used, but I do not found a way to disable parsing it.

Additionally there is problem with the translator comment which is not shown:
https://translate.wordpress.org/projects/wp-plugins/antispam-bee/stable/de/default/?filters%5Bstatus%5D=either&filters%5Boriginal_id%5D=7264244&filters%5Btranslation_id%5D=61317391

[Zodiac1978]

Maybe @swissspidy can help us here? Or @ocean90?

[swissspidy]

• If the core string changes we do need to change this test accordingly. If we do not have any tests for this to check it automatically, this will easily slip through.

I'd just add the antispam-bee text domain.

• The string also will be parsed by the PO-Parser from WordPress.org and is therefore on translate.wordpress.org. I think this will be not used, but I do not found a way to disable parsing it.

This can't be disabled.

Additionally there is problem with the translator comment which is not shown:
translate.wordpress.org/projects/wp-plugins/antispam-bee/stable/de/default/?filters%5Bstatus%5D=either&filters%5Boriginal_id%5D=7264244&filters%5Btranslation_id%5D=61317391

I see the comment:

[Zodiac1978]

I'd just add the antispam-bee text domain.

This is not handled by the locale options? It looks like a lot redundant work if this needs to be added to every plugin which needs it. The locale is saving the starting day of the week, why not something like this? Couldn't we add this to the locale?

This can't be disabled.

We could use the underlying internal translation function, but I think this is not recommended too.

I see the comment:

Thanks for checking. Maybe I just overseen it, because it was collapsed ...

[swissspidy]

The locale is saving the starting day of the week, why not something like this? Couldn't we add this to the locale?

Yeah why not.

Found https://core.trac.wordpress.org/ticket/39733 and https://core.trac.wordpress.org/ticket/47511 that look related.

[Zodiac1978]

I've added my comment to the trac ticket https://core.trac.wordpress.org/ticket/47511#comment:19

Summary: This solution is used on 65 plugins at the moment and it is used inconsistently. WooCommerce is using its own translation, but Jetpack is using the core translation. Although this string is added to GlotPress and remains unused. Solution could be to add this information to the WP_Locale object which could be easily used in this case.

[Zodiac1978]

There is no traction on the ticket and no action item for this issue, so I close this, as we cannot do anything here.

If someone has a strong opinion on adding the ASB text domain, please re-open this issue.

[pedro-mendonca]

@Zodiac1978 I hope this is a solution :)

https://core.trac.wordpress.org/ticket/56698

[Zodiac1978]

The ticket got accepted and comes with WP 6.2, so we could use the new way to get the info. Re-opening.