1			ACKNOWLEDGEMENTS
2
3Thanks to everyone who have downloaded Squashfs.  I appreciate people
4using it, and any feedback you have.
5
6The following have provided useful feedback, which has guided
7some of the extra features in squashfs.  This is a randomly ordered
8(roughly in chronological order) list, which is updated when
9I remember...
10
11Acknowledgements for Squashfs 4.3
12---------------------------------
13
14Thanks to Bruno Wolff III and Andy Lutomirski for useful feedback
15during the long development process of Squashfs 4.3.
16
17Acknowledgements for Squashfs 4.2
18---------------------------------
19
20Thanks to Lasse Collin (http://tukaani.org/xz/) for mainlining XZ
21decompression support.
22
23Acknowledgements for Squashfs 4.1
24---------------------------------
25
26Thanks to Chan Jeong <chan.jeong@lge.com> and LG for the patches to support LZO
27compression.
28
29Acknowledgements for Squashfs 4.0
30---------------------------------
31
32Thanks to Tim Bird and CELF (Consumer Electronics Linux Forum) for helping
33fund mainstreaming of Squashfs into the 2.6.29 kernel and the
34changes to the Squashfs tools to support the new 4.0 file system layout.
35
36Acknowledgements for Squashfs-3.3
37------------------------------------
38
39Peter Korsgaard and others sent patches updating Squashfs to changes in the
40VFS interface for 2.6.22/2.6.23/2.6.24-rc1.  Peter also sent some small patches
41for the Squashfs kernel code.
42
43Vito Di Leo sent a patch extending Mksquashfs to support regex filters.
44While his patched worked, it unfortunately made it easy to make Mksquashfs
45perform unpredictably with poorly choosen regex expressions.  It, however,
46encouraged myself to add support for wildcard pattern matching and regex
47filters in a different way.
48
49Acknowledgements for Squashfs-3.2-r2
50------------------------------------
51
52Junjiro Okajima discovered a couple of SMP issues, thanks.
53
54Junjiro Okajima and Tomas Matejicek have produced some good LZMA patches
55for Squashfs.
56
57Acknowledgements for Squashfs-3.2
58---------------------------------
59
60Peter Korsgaard sent a patch updating Squashfs to changes in the VFS interface
61in Linux 2.6.20.
62
63Acknowledgements for Squashfs-3.1
64---------------------------------
65
66Kenneth Duda and Ed Swierk of Arastra Inc. identified numerous bugs with
67Squashfs, and provided patches which were the basis for some of the
68fixes.  In particular they identified the fragment rounding bug, the
69NFS bug, the initrd bug, and helped identify the 4K stack overflow bug.
70
71Scott James Remnant (Ubuntu) also identified the fragment rounding bug,
72and he also provided a patch.
73
74Ming Zhang identified the Lseek bug in Mksquashfs.  His tests on the
75performance of Mksquashfs on SMP systems encouraged the rewrite of
76Mksquashfs.
77
78Peter Korsgaard, Daniel Olivera and Zilvinas Valinskas noticed
79Squashfs 3.0 didn't compile on Linux-2.6.18-rc[1-4] due to changes
80in the Linux VFS interfaces, and provided patches.
81
82Tomas Matejicek (SLAX) suggested the -force option on Unsquashfs, and noticed
83Unsquashfs didn't return the correct exit status.
84
85Yann Le Doare reported a kernel oops and provided a Qemu image that led
86to the identification of the simultaneously accessing multiply mounted Squashfs
87filesystems bug.
88
89
90Older acknowledgements
91----------------------
92
93Mark Robson - pointed out early on that initrds didn't work
94
95Adam Warner - pointed out that greater than 2GB filesystems didn't work.
96
97John Sutton - raised the problem when archiving the entire filesystem
98(/) there was no way to prevent /proc being archived.  This prompted
99exclude files.
100
101Martin Mueller (LinuxTV) - noticed that the filesystem length in the
102superblock doesn't match the output filesystem length.  This is due to
103padding to a 4K boundary.  This prompted the addition of the -nopad option.
104He also reported a problem where 32K block filesystems hung when used as
105initrds.
106
107Arkadiusz Patyk (Polish Linux Distribution - PLD) reported a problem where 32K
108block filesystems hung when used as a root filesystem mounted as a loopback
109device.
110
111Joe Blow emailed me that I'd forgotten to put anything in the README about
112mounting the squashfs filesystem.
113
114David Fox (Lindows) noticed that the exit codes returned by Mksquashfs were
115wrong.  He also noticed that a lot of time was spent in the duplicate scan
116routine.
117
118Cameron Rich complained that Squashfs did not support FIFOs or sockets.
119
120Steve Chadsey and Thomas Weissmuller noticed that files larger than the
121available memory could not be compressed by Mksquashfs.
122
123"Ptwahyu" and "Hoan" (I have no full names and I don't like giving people's
124email addresses), noticed that Mksquashfs 1.3 SEGV'd occasionally.  Even though
125I had already noticed this bug, it is useful to be informed by other people.
126
127Don Elwell, Murray Jensen and Cameron Rich, have all sent in patches.  Thanks,
128I have not had time to do anything about them yet...
129
130Drew Scott Daniels has been a good advocate for Squashfs.
131
132Erik Andersen has made some nice suggestions, unfortunately, I have
133not had time to implement anything.
134
135Artemiy I. Pavlov has written a useful LDP mini-howto for Squashfs
136(http://linuxdoc.artemio.net/squashfs).
137
138Yves Combe reported the Apple G5 bug, when using Squashfs for
139his PPC Knoppix-mib livecd project.
140
141Jaco Greeff (mklivecd project, and maintainer of the Mandrake
142squashfs-tools package) suggested the new mksquashfs -ef option, and the
143standalone build for mksquashfs.
144
145Mike Schaudies made a donation.
146
147Arkadiusz Patyk from the Polish Linux Distribution reported that Squashfs
148didn't work on amd64 machines. He gave me an account on a PLD amd64 machine
149which allowed myself to track down these bugs.
150
151Miles Roper, Peter Kjellerstedt and Willy Tarreau reported that release 2.1 did
152not compile with gcc < 3.x.
153
154Marcel J.E. Mol reported lack of kernel memory issues when using Squashfs
155on small memory embedded systems.  This prompted the addition of the embedded
156system kernel configuration options.
157
158Era Scarecrow noticed that Mksquashfs had not been updated to reflect that
159smaller than 4K blocks are no longer supported.
160
161Kenichi Shima reported the Kconfig file had not been updated to 2.2.
162
163Aaron Ten Clay made a donation!
164
165Tomas Matejicek (SLAX) made a donation!
166