1
## modifications dd-mm-yyyy
2
---------------------- PATCH FOR PARTED 1.6.5 ----------------------------
3
1 initial revision 07-04-2003
4
2 one pass resizing, removal of debug info 08-04-2003
5
3 safe abort if resize failed, code cleanups, timer, 10-04-2003
6
source file split, won't resize if not unmounted,
7
only relocate data if needed, minimize disk operations
8
4 memory leaks removal, code cleanups, resize hfs+ code, 17-04-2003
9
more checks, minor hfs resize bugfix, probe code
11
5 hfs+ resize bugfixes : 19-04-2003
12
* fragmented fs could be corrupted
13
* VH wasn't written on error during alloc map writing
14
* attributes file could be corrupted
15
6 Use PedSector to be able to use 2To+ HD 23-04-2003
16
Minor probe bugfix, Cleanups, HFS+ Timer tuning,
17
7 80 columns indentation 23-04-2003
18
8 Bugfix free blocks calculation in wrapper
19
(makes Mac OS boot !) 28-04-2003
20
---------------------- PATCH FOR PARTED 1.6.6 ----------------------------
21
9 Don't destroy the file being worked on in case of
22
interruption of Parted 28-10-2003
23
---------------------- PATCH FOR PARTED 1.6.10 ---------------------------
24
10 Regression tests, URL correction, In effect_move_extent :
25
corrected memory leak & corrected a bug in precondition checks
26
Added error messages, Check ped_alloc results
27
Use macro for test / set / clear in the allocation bitmap
28
Light probe correction, Check return value of get_empty_end
29
Moved dynamic memory allocation out of effect_move_extent
30
Check HFS+ version, Set implementation creator code
31
Check journal absence, Corrected a bug in HFS+ block number
32
calculation 24-04-2004
33
--------------------- PATCH FOR PARTED 1.6.11 ----------------------------
34
11-Some pointer dereference moved after non nul assertion
35
-Error messages for HFS(+) file IO
36
-Memory leak correction in hfs(plus)_read_bad_blocks
37
-Mark out of volume blocks as used
38
(improve compatibility with broken HFS+ Linux
40
WARNING : this fix is not 100% tn1150 compatible :
41
"The allocation file may be larger than the minimum
42
number of bits required for the given volume size.
43
Any unused bits in the bitmap must be set to _zero_."
44
Anyway neither is the Linux implementation, nor was my
45
previous implementations
46
Maybe I should ask Apple to change the specifications
47
-HISTORY, DOC and TODO files 29-04-2004
48
12 Corrected a bug in hfsplus_volume_resize : size of alloc
49
bitmap could be miscalculated 29-04-2004
50
--------------------- PATCH FOR PARTED 1.6.12 ----------------------------
51
13-Finally partial rewrite of *_search_move_*
52
Easier to maintain and prepare for extent search and
53
relocation algorithm changes for better ones.
54
-"An extent has not been relocated!" message now only when
56
-Slightly better and simpler relocation algorithm
57
-Update of Makefile.in and Makefile.am in fs_hfs
58
-Sign correction for some 8bits HFS integers
59
-Added the option --enable-hfs-extract-fs in 'configure'
60
-Added every ped_geometry_sync where needed
61
-Bugfix : "A root node does not need to exist
62
(if the tree is empty)."
63
- now handled correctly in btree_search
64
-Bugfix : failure wasn't detected in some cases
65
during 2 pass relocation (*_search_move_*)
66
-Bugfix : The extent key comparaison was done in a wrong order
67
and a pad field was used during the comparaison
68
-Bugfix : in hfs_file_find_sector and hfsplus_file_find_sector
69
the absolute position of a file sector could be
70
miscalculated in case of fragmentation, resulting
71
in potential data corruption, or various errors
72
-Bugfix : The end of the HFS bitmap compatibility block was
73
miscalculated ( (1<<16)/8 instead of (1<<16) )
76
--------------------- PATCH FOR PARTED 1.6.14 ----------------------------
77
14 Port of Patch 13 for Parted 1.6.14 (update timestamps)
79
--------------------- PATCH FOR PARTED 1.6.15 ----------------------------
80
15-hfsplus_open : added a warning message if the "attributes"
82
-hfsplus_open : added a test to check if the "allocation"
83
special file has been correctly opened
84
-optimisation of hfs+ block access : don't recalculate
85
the address of each sector, and avoid checking the cache if
87
( hfsplus_file_read && hfsplus_file_write
88
&& hfsplus_file_find_extent && hfs_file_find_sector)
89
-cut the "hfs.c" file in several parts
90
-Bugfix: in hfsplus_do_move_primary, hfs_effect_move_extent
91
was called instead of hfsplus_effect_move_extent !!!
92
This could not produce data corruption, because of a welcome
93
ASSERT in *_effect_move_extent that would detect the bug :)
94
-Bugfix: in hfs_effect_move_extent, do
95
PED_ASSERT(*ptr_to_fblock <= *ptr_fblock, return -1);
97
PED_ASSERT(*ptr_to_fblock < *ptr_fblock, return -1);
98
and added that assertion to hfsplus_effect_move_extent
99
-Bugfix: bugs introduced in rewrite of hfsplus_file_read
100
&& hfsplus_file_write : last sector was incorrectly detected
102
-Cache the extent references (speed improvement ?)
104
16-Bugfix: in hfsplus_do_move (reloc_plus.c), case CR_BTREE_EXT_ATTR
105
incorrectly updated the cached part of priv_data->catalog_file
106
instead of priv_data->attributes_file
107
-Bugfix: in hfs_read_bad_blocks && hfsplus_read_bad_blocks,
108
now generate an error if file_ID or type mismatch after the
110
Also check return value of ped_malloc
111
-Bugfix: in hfsplus_btree_search, check return value of ped_malloc
113
---------------- INTEGRATION IN PARTED 1.6.22 (cvs) ----------------------
114
Futur changes will be described in ../../ChangeLog