Broken Encoding after File Translation Upload

[SimbaStitch]

Describe the issue

Hello :)
Weblate version 5.10.4
Sometimes, some uploaded files with translations were read incorrectly.
In the majority of cases, we have correct translations uploaded.
Valid UTF-8 symbols were substituted for invalid ones. We faced this problem in different languages (like Czech, Slovak, French, Arabic (Egypt) and Russian (in attachment))
Looks like Weblate reads 4 bytes utf-8 symbols as 2-byte ones, at least in the hex editor (wich doesn't support 4 byte utf symbols) they look the same as in Weblate.
Example:

ui-suite-questionnaires-ru_RU.properties.txt

Could you please help with that? <3

I already tried

• I've read and searched the documentation.
• I've searched for similar filed issues in this repository.

Steps to reproduce the behavior

Upload the translation file in *.properties format in UTF-8 encoding (non-English)

Expected behavior

we have correct translations uploaded

Screenshots

Exception traceback

How do you run Weblate?

No response

Weblate versions

5.10.4

Weblate deploy checks

Additional context

No response

[nijel]

I guess you are using CSV format. Choose a variant with UTF-8 instead of the one that does encoding autodetection. The autodetection is never fully accurate.

[SimbaStitch]

Hello
No, it was UTF-8 as it said in the description
Also those symbols are stuck in Weblate memory

So we can’t use weblate memory because of this

Is there some way how we can rid off those consequences?

[nijel]

You can rebuild the translation memory once the issue is fixed.

Properties files are ISO-8859-1 with Unicode literals encoded, if you upload UTF-8 instead, this will probably happen. Configure the translation to use UTF-8 properties instead in the component configuration.

[SimbaStitch]

Now it is better for understanding 😃
Thank you!

[SimbaStitch]

Please reopen this issue