When using fruit:encoding=native, an additional catia mapping should be added: 0x2e:0xf029 This character is used then "mangle names = no" is used and MacOS creates a file with a terminating . (eg: test.). Adding this mapping permits the use of such names since a final . is allowed in MacOS. Otherwise, a file named "test(U+F029)" is created on the file system and it's not readable/writable by Mac clients. Even with this fix, Fruit still has trouble accessing resource forks for files ending with a period (like fonts), but this is caused by something else.
Unfortunately, it seems that in some circumstances, it causes bad side effects. Sometimes, the . and .. folders are listed as U+F029 and U+F029,U+F029. This can have tragic consequences for folder deletions (it becomes recursive up to the root). I'm not to sure how to fix file names ending with a . on MacOS then... I think a naive character swap like catia does is not enough.
A log level 10 of an incidence where . or .. are mapped would be really helpful (alongside a network trace an overview of whats going on). Instructions how to reproduce it as well.
Actually, I'm having trouble replicating it myself, but I saw it. I think it happens when stacking more vfs modules (shadow_copy2 maybe). I'm going to investigate more.
for native encoding samba would have to check for trailing period and map, an unconditional mapping of 0x2e:0xf029 is not correct. *** This bug has been marked as a duplicate of bug 11255 ***