We have developed Delphi projects in Lithuanian (Baltic) local Windows machines and our Delphi *.pas files have been in "format ANSI" and they, of course, contained Lithuanian constants as well.

Now we are starting to use Cursor IDE (+ the best models, like Claude Opus 4.6 Max or any other top model) to make AI developement on the codebase, but Cursor IDE + models are totally incapable to handle ANSI Lithuanian files. Cursor + models get confused and overwrites and overwrites Lithuanian texts and complete mess is the sole result. The opening of *.pas files with relevant encoding in Cursor IDE is not helping. The prompts are not helping either - models become unmanageable with this respect.

So - I start the pratice to convert *.pas file from ANSI to UTF8 format - Delphi editor offers command "Right click - File Format..." and there is menu to change the format. This conversion worked as far as AI is concerted - the Cursor IDE + models are able to program such files perfectly.

But if the UTF8 format *.pas files contains string constants with Lithuanian values and those constants are used in the legacy ANSI format *.pas file code, then Strings in both files are incompatible. Even worse, IBX components, that wrote those Lithuanina values into UTF-8 Firebird database perfectly, now are writing complete mess.

So - there is no other option, than to convert and retest all the *.pas files into UTF8 format. But how to do that conversion? Only working way is to do that manually from the Delphi IDE. I have not find any other way. Delphi IDE definition of ANSI/UTF8 file formats are somehow strange, because - if Cursor IDE is asked to detected them, then it decides that there are no BOMs etc. and gets confused. So - how to handle this conversion?