[4.3. Bug] Generate aliases for pages with russian names is broken

[Serhii-DV]

Hi.

It doesn't generate at all for russian words.

For example:

• create any page in Site structure with names like Первая страница, Вторая страница;
• generate aliases for them;
• the result will be an empty alias value, but excepted to see Pervaja-stranica (or something like this).

[aschempp]

Did that work in a previous version?

[fritzmg]

It should work in Contao 3, yes: https://github.com/contao/core/blob/master/system/helper/utf8_lookup.php#L189-L196

Probably duplicate of #560. The PR is still WIP: #569

[leofeyer]

It is true that transliteration is broken in Contao 4 (since switching to patchwork/utf8 to be exact). And it can only be solved by #569, which we forgot to merge. But there are two solutions IMO:

1. Do not transliterate at all.

We have already dropped transliteration for file and folder names in Contao 4.4.0-beta1, so why not just drop it for aliases, too? Any browser should be able to handle URLs like https://localhost/app_dev.php/ru/первая-страница.html.

2. Use the Slugify service to transliterate.

The service might not handle all languages, otherwise its fine.

@contao/developers /cc

[leofeyer]

I have removed the broken transliteration in 1dbbf9f. We might add Slugify in Contao 4.5, because it implies further changes (such as moving the forlderUrls setting from the localconfig.php file to the config.yml file) that I don't want to add this short before the final release.

[Toflar]

To be honest, I think it depends. It's something that should be a setting of a root page. If you think in German, somebody might want to have URI's like foobar.de/über-uns.html and an other one would like to have foobar.de/ueber-uns.html. It's a matter of taste. Also I think prepend_locale as well as folderUrls are both settings that should be a root page setting, not a system setting. There are other tickets where people would like to have the main language to be accessible via www.domain.de and all translations via www.domain.de/{locale}. If it was configurable in the root page settings, I guess this would be possible too then. So I think we should mark that as a feature for 4.5.

[leofeyer]

You should copy your comment to #569. This is just the bug report for the broken transliteration caused by Utf8::toAscii().

[tsarma]

Please make $arrSearch and $arrReplace customizable somewhere in global array. Striping away special characters in Tibetan language lose the meaning of the word.