Default character set and collation

Dear all,
I have some trables with installation related to char set.
So I would like to ensure:
Utf8 is the character set. Can we use utf32?
collation is utf8_general_ci. Am i correct can we use _universal_ci on whatever charset we choose?
Bets regards

why utf32? utf8 is a standard

I am just asking. I have some problems with Greek charcters on long text fields. But this is a minor problem.
I have already chosen utf8.
But there is a problem with collation. Mariadb10 starts with unicode_ci while v13 starts installation with general_ci. I search inside the v13 installation docs and everything is set up on general_ci.Very often I have problem with database migration due to different collation.

1 Like

ok. I suppose if your database supports it, Dolibarr will. Nevertheless most tools communicate for export and import in utf8.
Maybe test it and if it works for you, you’ll be fine.

Kind regards

PS: https://stackoverflow.com/questions/496321/utf-8-utf-16-and-utf-32

Thank you very mutch.
I know the story since we programmed EPROMs and EEPROMs for DOS 3.33 in order to have (ASCII) Greek characters on 8088 CPU baced PCs and dot matrix printers.
Utf8(that I already have choosen) supposed to work perfect, but they dont.
If I input a Greek text in a text filed both In the PHPmyadmin and in the related text filed the text is the same for example:
INPUT IN PRODUCT->LABEL : “ΕΙΚΟΝΙΚΟ ΠΡΟΙΟΝ” (means VIRTUAL PRODUCT)
(it is the same in both displaied field and PHPmyadmin llx_product->label field)
INPUT IN PRODUCT->DESCRIPTION: “ΕΙΚΟΝΙΚΟ ΠΡΟΙΟΝ” (label and description are the same)
but inside the PHPMYADMIN llx_product-> description is: “ΕΙΚΟΝΙΚΟ ΠΡΟΙΟΝ” (which unfortunately in the forum web page dispaied correctly thus I submit an image instance)
In most cases it displaied correctly but some times not. Especialy in the case that there is a language translate in the middle. That is my problem.

Hello there,

We are also using the Greek language in dolibarr (v14) and can confirm that in various instances with greek letters we get html encoded characters like the ones in your screenshot. These characters are also stored in the database like so. Our belief is that at some point, either the browser while posting, or the translation module did not convert the characters successfully to UTF8 and stored the HTML greek special characters instead.
You can reproduce the problem in the ticketing module when you try to mark a ticket are read.

Can you write an issue on Issues · Dolibarr/dolibarr · GitHub with the description of the process to do to reproduce the bug ?

Dear Eldy,

My apologies for the late response. I have just created an issue in the following url
https://github.com/Dolibarr/dolibarr/issues/20068#issue-1138313449
I will be following both threads if you need me to perform any other tests.

Petros

Dear Pilaftsis,

Did you solved the problem with the Greek characters ?