Duplicate post_class() for different categories

[SergeyBiryukov]

1. Create two categories, "Первая рубрика" and "Вторая рубрика". They will have the following slugs:

   %d0%b2%d1%82%d0%be%d1%80%d0%b0%d1%8f-%d1%80%d1%83%d0%b1%d1%80%d0%b8%d0%ba%d0%b0
   %d0%bf%d0%b5%d1%80%d0%b2%d0%b0%d1%8f-%d1%80%d1%83%d0%b1%d1%80%d0%b8%d0%ba%d0%b0
   

2. Create a post in each category.
3. Both posts will have a category-- class.

sanitize_html_class() uses a fallback value (term ID) if the resulting string is empty. It should do the same if the resulting string only contains hyphens.

[sgrant]

Hiya! I took a crack at this. Added a new conditional in sanitize_html_class, and added three unit tests relating to the fallback in a new file (seems to fit the formatting unit test structure). This should swap any remaining sanitized string to the fallback if it only consists of hyphens. Patch attached. Thanks!

[TempAcc]

Also big problem is coming when you try create category with numbers-chars slug. For example "25кадр". In output we get category-25, but if category with ID=25 exists and has own css we have a new problem.

Maybe transliteration will help?

[davideugenepratt]

I installed sgrant's patch and it works for me. The unit test ran without any failures.

[TempAcc]

Replying to davideugenepratt:

I installed sgrant's patch and it works for me. The unit test ran without any failures.

The patch not covering all possible situations. One of these described above by me. Also, as minimum, patch not fixes underscore symbol _ (try instead of hyphen)

I think the best way use $fallback variable when $sanitized != $class or use term ID always.

[davideugenepratt]

Or maybe one last conditional that checks for a numerical value and if it's numerical will use the ID.

[sgrant]

If the issue is the non-letter characters, should the single additional check be whether or not the remaining characters after replacement are exclusively in [0-9_-] ? It seems redundant to have multiple checks; rather, maybe it's best to find the line in the sand and do the correct test.

[sgrant]

If only non-letter characters remain, use fallback. New tests.

[davideugenepratt]

I applied patch.2.diff and it works for me for categories "Первая рубрика", "Вторая рубрика", and "25кадр".

[DrewAPicture]

We're going to need a new patch here. patch.2.diff​ doesn't account for the "category" part of the class name (therefore removing the hyphens via the preg_replace() returns "category" not empty), and for that matter, we should probably handle it for any prospective taxonomy slug.

[A5hleyRich]

Refresh of patch 2 to work with taxonomy slug

[A5hleyRich]

Refreshed to work with any taxonomy. The taxonomy name should only contain lowercase letters and underscores, as per the codex (​http://codex.wordpress.org/Function_Reference/register_taxonomy). Therefore, do we need to sanitize the taxonomy name?

[DrewAPicture]

@SergeyBiryukov: patch.3.diff​ fixes the problem for me. I wonder what you think about this bit, however: $taxonomy . '-' . sanitize_html_class( $term->slug, $term->term_id ). What are your thoughts on essentially keeping sanitize_html_class() separate from the term slug logic as .3 does?

[SergeyBiryukov]

Replying to DrewAPicture:

What are your thoughts on essentially keeping sanitize_html_class() separate from the term slug logic as .3 does?

Actually, I don't think we should touch sanitize_html_class() at all, it works as expected.

This logic should be in post_class() and body_class().

30883.2.diff​ is a new patch with new tests.

[SergeyBiryukov]

In 31979:

Avoid duplicate classes for different terms with UTF-8 slugs in post_class() and body_class().

Fall back to term ID if the sanitized slug is numeric or only contains hyphens.

props SergeyBiryukov, A5hleyRich, sgrant, davideugenepratt.
fixes #30883.