home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Amiga GigaPD 3
/
Amiga_GigaPD_v3_3of3.iso
/
netbsd
/
contrib
/
ados
/
bffs1.25.lzh
/
bffs1.25
/
relnotes
< prev
Wrap
Text File
|
1993-04-10
|
2KB
|
47 lines
1.25 Release fixes over 1.2:
This is a quick patch to the 1.2 distribution which fixes a nasty
large block read bug. You will probably notice this bug if you
copy a large file off the disk, then compare it to the original.
This bug was brought to you complements of the gnarly death block
batch routines. Thanks to sjk@net.bio.net for pointing this out.
Next release? How about a device interface similar to UNIX? ie: /dev
1.2 Release fixes over 1.1:
Fixed preposterous (bad value) bootblock crashing filesystem
Seek of superblock is now much smarter - will look for alternate starts
Low memory allocations now much more robust
Assigned paths onto the filesytem now work correctly
previously, assigned paths were parsed as the root directory
Indirect blocks of (very) large files were incorrectly being calculated
Fixed cache low memory problem - fs would exit on low memory when
attempting to allocate cache space.
Filesystem now reports correct number of blocks consumed by files
File dates were five hours fast; they are now probably 5 slow
Filesystem now starts immediately if Mount=1 in mountlist
Setting Reserved=-1 will make BFFS not rely on information in the
boot block of the disk. Some cases where BFFS crashed
previously can be eliminated with this feature.
1.1 Release fixes over 1.0:
fsck now requires 50k of stack, instead of just crashing
sym links now show what they point to in the comment field
block and char special files now show type and device in comment field
fixed bug in filename path lookups which would cause a guru when
referencing objects that are not directories as if they were.
added automatic path lookups to follow symlinks
diskname will now default to BFFS{partition} and can be changed with
the relabel command
file permissions and file type are now considered when granting file
read requests
filename searches are now case-independent if a case-dependent match
is not found
fixed problem with return filename on root dir, should be compatible
with workbench now
Added code to handle SAME_LOCK packet. Unfortunately, I suspect a bug
in the way WorkBench handles that info so this routine has been
hacked to always return LOCK_SAME.
Fixed motor off problem on floppy drives
New program - dumpfs to show filesystem information
1.0 Release : beta only