There is a problem with the article alias (İ letters) Topic is solved

[ithaki]

Article Title;
İskenderun
Auto Alias
i-skenderun

Must be
iskenderun

[pe7er]

I cannot reproduce your issue. Do you have any 3rd party plugins active that might change it?

In Joomla 3.9.19 with en-GB English as default admin language,
I've created a new article "Is this really an issue?" and Joomla creates automatically the alias "is-this-really-an-issue".

In Joomla 3.9.19 with tr-TR (version 3.9.4.1) as default admin language,
I've tried "İskenderun" which got the alias "iskenderun".

Could you please try to reproduce this error with a new Joomla 3.9.19 installation without any 3rd party extensions (like plugins)?
And first with en_GB as default admin language, then with tr_TR as default admin language?

[ithaki]

ok. experimenting, I'll give you informatio. thx

[ithaki]

Hi pe7er

I installed a clean Joomla 3.9.19
The error continues

Admin Language En;
İkarus
i-karus

Admin Language Tr;
İletişim
i-letisim


This is a new problem. Did not exist before

[pe7er]

Please use Forum Post Assistant to give somewhat more info about your setup:
https://forumpostassistant.github.io/docs/

[Webdongle]

What is the first letter? It looks like an I (capital i) but with a dot over it like a lower case i

[pe7er]

What is the first letter? It looks like an I (capital i) but with a dot over it like a lower case i

Yes, "Dotted İi and dotless Iı are separate letters in Turkish and Azerbaijani"
https://en.wikipedia.org/wiki/Dotted_and_dotless_I

[pe7er]

I think that it's inevitable because if dotted İi and dotless Iı are both converted to lowercase "i",
then you will run into problems.

I suppose that
dotted İ is seen as a special symbol and converted to i- (i + dash)
and dotless I to i (i without a dash)

[Per Yngve Berg]

I can confirm the issue with the Turkish Capital i (İ).

Enabling Unicode aliases made the issue go away.

I could not find any issued with the non dotted i.

[Webdongle]

...

I suppose that
dotted İ is seen as a special symbol and converted to i- (i + dash)
and dotless I to i (i without a dash)

That was my thinking

[ithaki]

"Enabling Unicode aliases has resolved the issue."
But it will cause problems on multilingual sites

It was working fine before a few versions

[ithaki]

Interesting
The last character in the word does not cause problems

Testİ İleri denemİler Testİ
testi-i-leri-denemi-ler-testi

[sozzled]

"Enabling unicode aliases has resolved the issue" but it will cause problems on multilingual sites.

I don't know what problems this will cause.

It was working fine before a few versions

I tested, created an article with the title "İskenderun" with J! 3.9.19 and it created the article with the alias "i-skenderun". I tested the same thing on a test site running J! 3.8.13—that's 20 version ago—and it created the article with the alias "i-skenderun". So this is not a new thing.

Global Configuration » Site » SEO Settings » Unicode Aliases » Yes will "fix" this but, as I wrote earlier, I do not know what other "problems" this will cause.

Different things happen with accented characters:

Österreich ==> oesterreich
Éclairs ==> eclairs
Æða ==> aedha
Ångström ==> angstroem
Øresund bridge ==> oresund-bridge

From what I can see (and I could be wrong) the alias URL involves stringURLSafe that has been around since J! 1.7 (see https://api.joomla.org/cms-3/classes/Jo ... ilter.html). Also see discussion on GitHub about the OP's issue: https://github.com/joomla/joomla-cms/issues/27875; perhaps you used @infograf768's patch and did not re-apply it after you updated your J! software recently?

[ithaki]

Enabling "enabling Unicode aliase" is risky for Seo.

-Enabling unicode aliase
Title; Şiş Kebap
Auto alias; şiş-kebap <-- Bad url for seo. Non-english characters. (This topic is controversial. But we shouldn't risk)

-Disable unicode aliase;
Title; İce
Auto alias; i-ce <-- Bad url for seo. Meaningless words.

[sozzled]

Enabling "enabling Unicode aliase" is risky for Seo.

-Enabling unicode aliase
Title; Şiş Kebap
Auto alias; şiş-kebap <-- Bad url for seo. Non-english characters.

I tested.

Global Configuration » Site » SEO Settings » Unicode Aliases » No (default setting): Şiş Kebap ==> sis-kebap

Global Configuration » Site » SEO Settings » Unicode Aliases » Yes: Şiş Kebap ==> şiş-kebap.

I agree that the alias (generated when Unicode Aliases » Yes) in this case is incorrect because it contains unsafe ASCII characters (see https://www.urlencoder.io/learn/):

A URL is composed from a limited set of characters belonging to the US-ASCII character set. These characters include digits (0-9), letters(A-Z, a-z), and a few special characters ("-", ".", "_", "~").

ASCII control characters (e.g. backspace, vertical tab, horizontal tab, line feed etc), unsafe characters like space, \, <, >, {, } etc, and any character outside the ASCII charset is not allowed to be placed directly within URLs.

Thank you, @ithaki, for explaining your point. The question about whether "İ" should be replaced with "i" or "i-" is a good question but I don't know what will be done about it. The question about whether people should use

• Global Configuration » Site » SEO Settings » Unicode Aliases » Yes, or
• Global Configuration » Site » SEO Settings » Unicode Aliases » No

is also interesting if (in the first-mentioned cased) accented characters not replaced with "safe" US-ASCII equivalent characters.

[infograf768]

Folks, this is a well known issue.
Please read
https://github.com/joomla/joomla-cms/issues/27875
and specially https://github.com/joomla/joomla-cms/is ... -588080115
and
https://github.com/joomla/joomla-cms/pull/16880

I am explaining all about it there.