Stan Hoeppner wrote:
I run into a similar problem frequently when saving downloads to my XFS samba
shares via a Windows client. I don't recall ever seeing XFS corruption,
Just a note:
Corrupt isn't exactly the right word -- as it doesn't spread to the rest
of the file system as a real corrupt likely would ... It seems to be limited
to these filenames -- where you can't remove them, or 'address' them, or rename
them. -- and their information in the owner and date fields shows up as
marks. It could be that once I retired any of these disks that contain these
files the problem will stay on them.
do have problems manipulating these file names through bash. Quite often I
end up having to rename the files through Windows Explorer to something bash
can handle. That usually fixes the problem--not always, but usually.
These files are not visible to explorer -- I tried/looked.
I'm running suse -- and they default to UTF-8, which is compatible with Win7
speaking UTF16. It's just my older WinXP clients that have problems emitting
characters that become incomprehensible.
As I've never seriously dealt with character encoding issues (i.e. changed
anything related in Debian), I don't even know where/how to find my servers
default character encoding. Google isn't being very friendly here. I'm using
whatever character encoding is the default for US English Debian Lenny.
I started to some number of years ago due to having music from foreign lands --
it's only gotten worse from there :-)...now half my music and video collection
foreign chars in it.
But even in english, you there there are unicode display values for colon,
backslash, so you can use those usually forbidden characters in filenames --
are different characters. They do look slightly different since it's not
it's better than putting in some substitute.
Lots of places a colon is called for in music and movie titles where you can
full width colon -- don't even need spaces around it：see? (this is where I use
alot!...when I want to insert my own...). You can find the reverse of
Firefox with the character identify extension. Something I didn't know -- if
the Japanese characters, you can use the extension -- and what they call 'Romaji' --
the anglicized version of Japanese -- is just the names of the characters -- which
you can read out in the character identifier. Interesting but slow translating. AT
least you can sound out the words that way...:-).
Read the book "Fonts and encodings", and you'll get hooked, I guarantee it!
(google it, it's the first result).