[openbeos] Re: A note to file system writers

On 2004-02-09 at 02:08:40 [+0100], François Revol wrote:
> I thought this bug was located in Tracker... wasn't there a
> similar illed behaviour before in R5 and it was fixed in OT ?
> or was it just a side effect that got fixed ? now I wonder why
> it doesn't trigger in R5.
> 
> I'll go update NFS since it has that problem in Zeta.
> Though I'm pretty sure there is ..
> do I have to fill in the correct block dev id and inodes for
> the parent or does the vfs correct it by itself ?

In my ReiserFS add-on I fill d_dev and d_ino with the values for the 
directory itself (as done in the dosfs sample code), but since the correct 
values appear in the dirent returned by the readdir() syscall they 
apparently get overridden with the correct values. :-)

CU, Ingo

Other related posts: