Hello, I see no reaction to my previous message. Is there anyone on this project ?
About this bug, I just found out that the problem is about collating sequences in some locales. It might not be a solution for text files if it could impact the way diff processes unicode files content, but I worked it around by adding LC_COLLATE=C before the diff command : Best regards, Philippe Errembault ----- Mail transféré ----- De : Errembault Philippe <[email protected]> À : "[email protected]" <[email protected]> Envoyé le : Dimanche 27 janvier 2013 6h18 Objet : Confusion between unicode names in diff 2.9.19-4065 Hello, I have been building a raid6 server with the last Debian version and ext3, I have copied some NTFS Hard Disks on it, and double checked the copy with diff. On one copy, diff -r has been finding differences on four files in the same folder, and it appears, that it has been confusing files whose filename differed on unicode characters only, probably partly because, they were not in the same order between the NTFS and ext3 directories. > diff -r '/raid/...' '/mnt/...' Les fichiers binaires /raid/.../エンドカード1.jpg et /mnt/.../ブックレット1.jpg sont différents. Les fichiers binaires /raid/.../ブックレット1.jpg et /mnt/.../エンドカード1.jpg sont différents. Les fichiers binaires /raid/.../ブックレット2.jpg et /mnt/.../エンドカード2.jpg sont différents. Les fichiers binaires /raid/.../エンドカード2.jpg et /mnt/.../ブックレット2.jpg sont différents. Best regards, Philippe Errembault
