Unexpected invisible Em Space \u2003 character dropped in the group phase

[lucbouge]

Dear Piotr,

I am intensively using your excellent software. Thanks for it!

As I am just retired, I took time to run an extensive dedupe process on my archives. I encountered a surprising behavior.

For some reason, one of my files was named something like "filename\u2003". Everything went well at the group phase.

ad5a0980fff1c9e57f3209102bae9c36, 41700 B (41.7 KB) * 2:
    filename/Francisco_Rivera 
    filename/Francisco_Rivera 

Note the invisible character just after Rivera.

However, I got a double error message at the dedupe phase. Somehow, the fclones.group files had dropped this invisible character.

[2024-09-21 05:54:13.255] fclones: warn: Failed to read metadata of 'filename/Francisco_Rivera': Failed to read metadata of 'filename/Francisco_Rivera': No such file or directory (os error 2)

[2024-09-21 05:54:13.255] fclones: warn: Failed to read metadata of 'filename/Francisco_Rivera': Failed to read metadata of 'filename/Francisco_Rivera': No such file or directory (os error 2)

I used a small Python program to look at the actual name of the file. Strangely enough, it is actually

PosixPath('filename/Francisco_Rivera\u2003')

Character '\u2003' is Em Space, pretty unexpected, indeed. However, it should have worked, shouldn't it?

Best regards, Luc.