Ghost: option to disable unicode character conversion in post slugs.

[zumoshi]

hi,
after updating ghost i noticed my titles are converted in a weird way , some search gave me 0.4 change log which stated unicode characters in post slugs are converted to ascii as a bugfix. although this might be useful for some users to me it is more of a bug than a bugfix.

مشاهده محتویات فایل ها قبل از دانلود

is converted to

which even if i leave alone the fact that conversion its unwanted , its not even correct , that text written phonetically with latin letters would be

the converted version is not even readable (no vowels)

[javorszky]

Hi @zumoshi,

Sorry we haven't got back to you in 11 days. Just to make sure, are your titles (the title text when you view the post itself) converted to ascii characters too, or only the post slugs that appear in the address bar?

If it's the post slugs in the address bar, that is a necessary conversion as links with non-unicode characters are fairly hard to handle. Most modern browsers on modern computers are able to do that, however less capable devices aren't. The arabic - latin conversion is what I assume a "stupid straight" conversion. Computers possibly do not have the capability to interpret arabic text and produce a phonetically correct latin string out of it, although as I'm not familiar with the language, I could be wrong.

Hope this helps.

[zumoshi]

hi,
the text is Persian , and its only in the address , it is displayed correctly in the post , however about your argument i do not have the same opinion , 95% of iranian websites do use Persian text in address and there is no problem ,new browsers show persian text in address bar while older ones show the Escaped version , example :

`تاپیک-جامع-مدیالیچ-▓`

becomes :

%D8%AA%D8%A7%D9%BE%DB%8C%DA%A9-%D8%AC%D8%A7%D9%85%D8%B9-%D9%85%D8%AF%DB%8C%D8%A7%D9%84%DB%8C%DA%86-%E2%96%93

in node functions responsible are encodeURI and decodeURI. you can convert the example above back and forth with these two functions . it leaves acsii characters intact but convert characters that have a chance to fail in older browsers to %XX where xx represents character code.
i suggest converting titles to a safe alternative using this method rather just replacing letters with some random latin letters which may sound like title if readable .

also sorry about late reply , i just saw your reply :-s

edit: wordpress also encodes non-unicode URL's , example from official Persian Wordpress website. the links shows persian letters in address bar of chrome

[javorszky]

Hi,

Okay, this is interesting. @ErisDS ^ ?

[ErisDS]

This is another one to close against #3801. Ghost's handling of foreign languages is pretty much non-existent, and we need to tackle this as a coherent attempt to introduce support.

@zumoshi the conversion was considered a bug fix, because Ghost isn't intended to support non-latin characters at present and so them appearing in slugs was 1) unexpected and 2) causing other things to break. When support is added properly, we'll make sure it's done in such a way that it doesn't break other things.

[zumoshi]

ok , good to know you are working on adding foreign language support to ghost , i would be happy to help with Persian translations when you were ready to add other languages.

[jdssem]

any progress in this issue? I'm Korean user and I hope this feature will be included sooner or later

[karlvlam]

any progress of this issue? I'm a Chinese user and this feature is very useful (and make sense) to Asian language users.