Wrong non-latin slugs (in activity stream, group names etc.)

[motomac]

In activity stream link "View" doesn't work, if the name of blog post or name of forum topic contains cyrillic and is long.

See here: ​http://testbp.org/members/motomac/

[DJPaul]

How do we fix it? What should the link be?

[motomac]

For example, cyrillic string:

Проверка кириллицы. Проверка кириллицы. Проверка кириллицы. Проверка кириллицы

[motomac]

Simple solution is to increase limit of symbols for slugs tables in the database:
primary_link in wp_bp_activity make varchar(300), for example.

And make the same with other slug tables (group names, for example): slug in wp_bp_groups (varchar(900)).

It happens because non-latin symbols take more space then latinic (*6 times longer).

[boonebgorges]

Related: #880

[lenasterg]

Just to config this bug for 1.2.7.
The same behavior with long Greek titles of blog posts.

[DJPaul]

Obviously the blog post slugs aren't anything to do with BuddyPress. In 1.3, activity stream character limit has gone up to 200. See the notes on #880. Activity stream view link with the example above works okay for me. Could someone give trunk the once-over and find out what other slugs still break?

[SergeyBiryukov]

200 character limit allows up to 33 non-Latin characters. It's obviously better than the previous 100 character limit (which only allowed up to 16 non-Latin characters) and probably should be enough for group slugs, however forum topic titles can be longer, I guess.

motomac's example slug is being cut, but actually works on current trunk. We just need to make sure that cutting a slug doesn't break its last symbol, like it happened in #880. From what I've tested so far, that doesn't seem to happen now.

[DJPaul]

Oh, I see. Thank you for the explanation. FYI, 200 is the magic number because that's the size of the wp_posts table 'post_name' column which is relevant because BuddyPress will be moving to custom post types in the future.

[boonebgorges]

The important thing about the 200 character limit is that it is the size used in sanitize_title_with_dashes(). As long as we allow for at least that much space, WP will make sure that the uri-encoded non-latin slug is never cut off, and is always unique. It might not make for pretty URLs all the time, but it will at least ensure that links won't break.

[SergeyBiryukov]

So if that limit will ever be increased, it should happen in WordPress, right?

[boonebgorges]

Yes, that's my understanding, though it's something that you might want to confer with the WP l18n team about.

[DJPaul]

Closing this as fixed per the changes made back in ticket #880