File: NEWS

package info (click to toggle)
lphdisk 0.9.1-3
  • links: PTS
  • area: main
  • in suites: etch, etch-m68k, lenny
  • size: 192 kB
  • ctags: 237
  • sloc: ansic: 1,789; makefile: 54
file content (139 lines) | stat: -rw-r--r-- 6,764 bytes parent folder | download | duplicates (4)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
NEWS
----

12 February 2002 - lphdisk 0.9.1 - Alex Stewart <alex@foogod.com>

  (Fixed a couple of minor bugs submitted by users)

23 August 2001 - Second release, lphdisk 0.9. "Oh yeah..."

16 July 2001 - lphdisk 0.4as1 - Alex Stewart <alex@foogod.com>

There are quite a few changes in this release.  First of all, it's been largely
rewritten (I didn't set out to do it, but it ended up that way, for a couple of
reasons).  In the process, much of the code was made to be more modular, with
well-defined APIs and removal of most of the global variables.  A good chunk of
the rewriting was needed to support some of the new features.  A side effect is
that it will now also be cleaner to add partition table modification should
anyone ever get around to that.

OK, some of the changes and reasons:

  = lphdisk can now handle hibernate partitions on any primary partition, not
    just partition 4.

    As far as I know, all versions of Phoenix NoteBIOS will handle a hibernate 
    partition on any of the primary partitions, and some new notebooks are even 
    coming with it on partition 1 now.  In some cases it's also more 
    convenient to have it on another partition, so it seemed reasonable to 
    remove this limitation.

  = It can now be used on drives besides /dev/hda and /dev/sda.

    There are several reasons for this, one of the biggies being the unix 
    philosophy that, yes, the user really does know best.  There is also the 
    point that on some strange systems, the /dev entries may be named 
    differently or even point to different disks (and on devfs systems they 
    will follow a completely different naming scheme, for example), and in 
    some configurations (prepping a disk for moving into another machine) it 
    may also be desirable to perform this on something that isn't the first 
    drive in the system.

    That's not to say that lphdisk will just use any destination blindly.  A 
    lot of additional checks have been built in (see below).  If something 
    other than the first IDE disk in the system is used, a warning will still 
    be given to the user that the BIOS may not be able to use it there, but it 
    won't outright prohibit the use of it.

    Oh, and you can also now use a normal file as an argument to lphdisk.
    Pointing it at anything other than a block device does require the use of 
    the '-f' (force) switch, and it does still require that the first 512 bytes 
    contain a valid partition table.  This, however, makes it possible to use 
    lphdisk on disk image files which can later be transferred to an actual 
    physical drive, etc.

  = lphdisk will now automagically calculate and print the required partition
    size.

    Detection of physical memory size (through /proc/mtrr and /proc/meminfo) 
    and video memory (via VESA BIOS Extensions) is now done to figure out the 
    right size for a hibernate partition.  This information is displayed to 
    the user and also checked against the actual partition used to make sure 
    it'll be big enough (a warning is given if it doesn't appear to be).

    A '-p' (--probeonly) option has also been added to tell lphdisk to just 
    calculate and print the reccomended size without doing any formatting.

  = lphdisk now performs all of its updates using only the (raw) disk 
    device/file specified (i.e. just /dev/hda instead of using both 
    /dev/hda and /dev/hda4)

    Making assumptions about relationships between different device filenames
    is potentially very bad, and this prevents somebody who's done something
    silly like 'mknod /dev/hda4 b 3 1' from accidentally getting something
    important clobbered.

    This does require using raw unix fd IO instead of stdio streams, which for 
    some reason I don't quite fathom turns out to be somewhere around 10 times 
    slower.  I'm still working on figuring out if there's a way to speed this 
    up, but on decent sized partitions (260MB) it still only takes a minute 
    and a half or so, which considering this is usually a one-time operation 
    isn't that bad, and the safety added by doing things this way outweighs 
    an extra minute of runtime, IMHO.

    This also makes it possible to use regular files the same as disks for this 
    purpose, as mentioned above.

  = A lot of error-checking has been added, as well as sanity-checking of some
    of the data it's getting from the partition table.

    lphdisk will now check some of the obvious things like making sure 
    partitions don't overlap, and checking of return values has been added to 
    the couple of system calls which previously didn't have it.

  = Sanity-checking against /proc/partitions

    For block devices, lphdisk will attempt to correlate what it's seeing in 
    the partition table against what the running kernel reports via 
    /proc/partitions.  This both double-checks that it's looking at something 
    that should be a partitionable device, and ensures that the running system
    is in sync with what's actually on the disk (i.e. the user has rebooted
    since changing things with fdisk).

    One exception is made if it appears that the user may have just run fdisk 
    to create the "A0" partition out of free space and hasn't yet rebooted, but
    all other partitions check out.  This avoids the need for one more reboot
    in this reasonably common scenario, and should be fairly safe as long as 
    everything else matches up.

  = Long option support added

    Added support for GNU-style long options, just for completeness/linuxness.

  = Added '-d' (--debug) flag

    All of the debug messages previously enabled by compiling with the DEBUG 
    define set have been changed to be runtime-selectable using the '-d' 
    command line option (and a lot of additional debug messages have been 
    added).  This is particularly useful in troubleshooting situations where 
    users may only have ready access to a precompiled binary (and in a 
    program like this there isn't a great deal of speed or size advantage to 
    conditional compilation).

  = Added '-n' (--nowrite) flag

    The '-n' flag will cause lphdisk to do everything it would normally do, 
    except actually writing sectors to the disk.  This can be used for 
    testing, or in combination with '-d' to gather information about a 
    potentially dangerous issue.

  = '-f' (--force) flag is now useful

    '-f' will now cause most sanity-checking errors to be demoted to warnings 
    an allow execution to proceed anyway.  This can be used to:
      a) Use a regular file instead of a block device.
      b) Ignore mismatches between the partition table and /proc/partitions


27 July 2000 - First release, lphdisk 0.4. "Spooooon!"