3 This is wimlib version 1.3.2 (March 2013). wimlib can be used to read,
4 write, and mount files in the Windows Imaging Format (WIM files). These files
5 are normally created by using the `imagex.exe' utility on Windows, but this
6 library provides a free implementation of imagex for UNIX-based systems.
8 wimlib 1.3.0 has added experimental support for Windows. See the file
9 "README.WINDOWS" for more details.
13 A Windows Imaging (WIM) file is an archive. Like some other archive formats
14 such as ZIP, files in WIM archives may be compressed. WIM archives support two
15 Microsoft-specific compression formats: LZX and XPRESS. Both are based on LZ77
16 and Huffman encoding, and both are supported by wimlib.
18 Unlike ZIP files, WIM files can contain multiple independent toplevel directory
19 trees known as images. While each image has its own metadata describing a
20 directory tree and file access modes, files are not duplicated for each image;
21 instead, each file is included only once in the entire WIM. Microsoft did this
22 so that in one WIM file, they could do things like have 5 different versions of
23 Windows that are almost exactly the same.
25 Microsoft provides documentation for the WIM file format, XPRESS compression
26 format, and LZX compression format. The XPRESS documentation is acceptable, but
27 the LZX documentation is not entirely correct, and the WIM documentation itself
30 A WIM file may be either stand-alone or split into multiple parts.
34 wimlib provides a public API for other programs to use, but also comes with two
35 programs: `imagex' and `mkwinpeimg'.
37 `imagex' is intended to be like the imagex.exe program from Windows. `imagex'
38 can be used to create, extract, and mount WIM files. Both read-only and
39 read-write mounts are supported. See the man page `doc/imagex.1' for more
42 `mkwinpeimg' is shell script that makes it easy to create a customized bootable
43 image of Windows PE that can be put on a CD or USB drive, or published on a
44 server for PXE booting. See the main page `doc/mkwinpeimg.1' for more details.
46 There is an additional program, `wimapply', that is not installed by default.
47 It can be used to build a small executable with the ability to apply a WIM image
48 from a standalone WIM, without having to build the whole shared library. This
49 could be useful on Linux boot clients that only need to be able to apply a WIM,
50 not capture/split/join/append/export/mount a WIM. See `programs/wimapply.c'.
54 wimlib can create XPRESS or LZX compressed WIM archives. Currently, the XPRESS
55 compression ratio is slightly better than that provided by Microsoft's software,
56 while the LZX compression ratio is approaching that of Microsoft's software but
57 is not quite there yet. Running time is as good as or better than Microsoft's
58 software, especially with multithreaded compression, available in wimlib v1.1.0
61 The following tables compare the compression ratio and performance for creating
62 a compressed Windows PE image (disk usage of about 524 MB, uncompressed WIM size
67 XPRESS Compression LZX Compression
68 wimlib imagex (v1.2.1): 138,971,353 bytes 131,379,943 bytes
69 Microsoft imagex.exe: 140,406,981 bytes 127,249,176 bytes
71 Table 2. Time to create WIM
73 XPRESS Compression LZX Compression
74 wimlib imagex (v1.2.1, 2 threads): 11 sec 17 sec
75 Microsoft imagex.exe: 25 sec 89 sec
79 As of version 1.0.0, wimlib supports capturing and applying images directly to
80 NTFS volumes. This was made possible with the help of libntfs-3g from the
81 NTFS-3g project. This feature supports capturing and restoring NTFS-specific
82 data such as security descriptors, alternate data streams, and reparse point
85 The code for NTFS image capture and image application is complete enough that it
86 is possible to apply an image from the "install.wim" contained in recent Windows
87 installation media (Vista, Windows 7, or Windows 8) directly to a NTFS volume,
88 and then boot Windows from it after preparing the Boot Configuration Data. In
89 addition, a Windows installation can be captured (or backed up) into a WIM file,
90 and then re-applied later.
94 A major use for this library is to create customized images of Windows PE, the
95 Windows Preinstallation Environment, without having to rely on Windows. Windows
96 PE is a lightweight version of Windows that can run entirely from memory and can
97 be used to install Windows from local media or a network drive or perform
98 maintenance. Windows PE is the operating system that runs when you boot from
99 the Windows installation media.
101 You can find Windows PE on the installation DVD for Windows Vista, Windows 7, or
102 Windows 8, in the file `sources/boot.wim'. Windows PE can also be found in the
103 Windows Automated Installation Kit (WAIK), which is free to download from
104 Microsoft, inside the `WinPE.cab' file, which you can extract if you install
105 either the `cabextract' or `p7zip' programs.
107 In addition, Windows installations and recovery partitions frequently contain a
108 WIM containing an image of the Windows Recovery Environment, which is similar to
114 This is a commonly used free library to read and write XML files. You
115 likely already have it installed as a dependency for some other program.
116 For more information see http://xmlsoft.org/.
118 * libfuse (optional but highly recommended)
119 Unless configured with --without-fuse, wimlib requires a non-ancient
120 version of libfuse to be installed. Most GNU/Linux distributions
121 already include this, but make sure you have the libfuse package
122 installed, and also libfuse-dev if your distribution distributes header
123 files separately. FUSE also requires a kernel module. If the kernel
124 module is available it will automatically be loaded if you try to mount
125 a WIM file. For more information see http://fuse.sourceforge.net/.
126 FUSE is also available for FreeBSD.
128 * libntfs-3g (optional but highly recommended)
129 Unless configured with --without-ntfs-3g, wimlib requires the library
130 and headers for libntfs-3g version 2011-4-12 or later to be installed.
131 Versions dated 2010-3-6 and earlier do not work because they are missing
132 the header xattrs.h (and the file xattrs.c, which contains functions we
133 need). libntfs-3g version 2013-1-13 is compatible only with wimlib
136 * OpenSSL / libcrypto (optional)
137 wimlib can use the SHA1 message digest code from OpenSSL instead of
138 compiling in yet another SHA1 implementation. (See LICENSE section.)
142 * syslinux (optional)
143 * cabextract (optional)
144 The `mkwinpeimg' shell script will look for several other programs
145 depending on what options are given to it. Depending on your GNU/Linux
146 distribution, you may already have these programs installed, or they may
147 be in the software repository. Making an ISO filesystem requires
148 `mkisofs' from `cdrkit' (http://www.cdrkit.org). Making a disk image
149 requires `mtools' (http://www.gnu.org/software/mtools) and `syslinux'
150 (http://www.syslinux.org). Retrieving files from the Windows Automated
151 Installation Kit requires `cabextract' (http://www.cabextract.org.uk).
155 Besides the various well-known options, the following options can be passed to
156 wimlib's `configure' script:
159 If libntfs-3g is not available or is not version 2011-4-12 or later,
160 wimlib can be built without it, in which case it will not be possible to
161 apply or capture images directly to/from NTFS volumes.
164 If libfuse or the FUSE kernel module is not available, wimlib can be
165 compiled with --without-fuse. This will remove the ability to mount and
169 Build in functions for SHA1 rather than using external SHA1 functions
170 from libcrypto (part of OpenSSL). The default is to use libcrypto if it
171 is found on the system.
173 --enable-xattr, --disable-xattr
174 Enable or disable support for the extended-attributes interface to NTFS
175 alternate data streams in mounted WIMs. To support these, wimlib
176 requires that the setxattr() function and the attr/xattr.h header are
177 available. The default is to autodetect whether support is possible.
179 --disable-multithreaded-compression
180 By default, data will be compressed using multiple threads when writing
181 a WIM, unless only 1 processor is detected. Specify this option to
182 disable support for this.
185 Use a very fast assembly language implementation of SHA1 from Intel.
186 Only use this if the build target supports the SSSE3 instructions.
188 --disable-custom-memory-allocator
189 If this option is given, a very small amount of space will be saved by
190 removing support for the wimlib_set_memory_allocator() function. imagex
193 --disable-verify-compression
194 Unless this option is given, every time wimlib compresses a data block,
195 it will decompress it into a temporary buffer and abort the program with
196 an error message if the decompressed data does not exactly match the
197 original data. This only makes compression about 10% slower. However,
198 this checking can probably be safely disabled because there are no known
199 bugs in the compression code, and the SHA1 message digest of every
200 extracted file is checked anyway.
202 --disable-error-messages
203 Save some space by removing all error messages from the library.
206 Remove all assertions, even the ones that are included by default.
208 --enable-more-assertions
209 Enable assertions that are not included by default.
212 Include debugging messages. Only use this option if you have found a
216 Include more debugging messages. Only use this option if you have found
217 a bug in the library.
221 wimlib has mostly been developed and tested on x86_64 (64-bit) GNU/Linux.
223 It has been tested on x86 (32-bit) GNU/Linux occasionally.
225 wimlib may work on FreeBSD and Mac OS X. However, this is not well tested. If
226 you do not have libntfs-3g 2011-4-12 or later available, you must configure
227 wimlib with --without-ntfs-3g. On FreeBSD, before mounting a WIM you need to
228 load the POSIX message queue module (run `kldload mqueuefs').
230 The code pays attention to endianness, so it should work on big-endian
231 architectures, but I've never tested this so do not expect it to work.
233 As of wimlib 1.3.0, there is experimental support for Microsoft Windows. I
234 expect it to work on Windows Vista and later, although I have only tested it on
235 Windows 7. See README.WINDOWS for more information.
239 The WIM file format is specified in a document that can be found in the
240 Microsoft Download Center. There is a similar document that specifies the LZX
241 compression format, and a document that specifies the XPRESS compression format.
242 However, many parts of these formats are poorly documented, and some parts have
243 no documentation whatsoever. Some particularly poorly documented parts of the
244 formats have had comments added in various places in the library. Please see
245 the code and/or ask me if you have any questions about the WIM file format as it
246 exists in reality and not as it exists in Microsoft's poorly written
249 The code in ntfs-apply.c and ntfs-capture.c uses the NTFS-3g library, which is a
250 library for reading and writing to NTFS filesystems (the filesystem used by
251 recent versions of Windows). See
252 http://www.tuxera.com/community/ntfs-3g-download/ for more information.
254 lzx-decompress.c, the code to decompress WIM file resources that are compressed
255 using LZX compression, is originally based on code from the cabextract project
256 (http://www.cabextract.org.uk).
258 lzx-compress.c, the code to compress WIM file resources using LZX compression,
259 is originally based on code written by Matthew Russotto (www.russotto.net/chm/).
261 lz77.c, the code to find LZ77 matches (used for both XPRESS and LZX compression),
262 is based on code from zlib.
264 A very limited number of other free programs can handle some parts of the WIM
265 file format. 7-zip is able to extract and create WIMs (as well as files in many
266 other archive formats). However, wimlib is designed specifically to handle WIM
267 files and provides features previously only available in Microsoft's imagex.exe,
268 such as the ability to mount WIMs read-write as well as read-only, the ability
269 to create LZX or XPRESS compressed WIMs, and the correct handling of security
270 descriptors and hard links.
272 An earlier version of wimlib is being used to deploy Windows 7 from the Ultimate
273 Deployment Appliance. For more information see
274 http://www.ultimatedeployment.org/.
276 You can see the documentation about Microsoft's version of the imagex program at
277 http://technet.microsoft.com/en-us/library/cc749447(v=ws.10).aspx, so you can
278 see how it compares to the version provided by this library.
280 GNU/Linux equivalents of WIM format
282 What's the equivalent way to capture the filesystem of a GNU/Linux operating
283 system into an archive file? You have a few options:
286 SquashFS (http://squashfs.sourceforge.net/) provides a compressed,
287 read-only filesystem for Linux, and it's probably the closest equivalent
288 of the WIM format and better designed. Although you can't mount
289 SquashFS read-write, when wimlib does this for WIM files it's really an
290 illusion since the WIM isn't actually modified until the image is
291 unmounted. Multiple top-level images in SquashFS files are not
292 supported, although nothing stops you from just putting each image in a
296 FSArchiver (http://www.fsarchiver.org/Main_Page) is not widely used, but
297 it appears to have some features quite similar to the WIM format.
300 The well-known tar format can usually capture a UNIX filesystem just
301 fine, and compressing the tar file produces a good compression ratio
302 (better than WIM, especially if using XZ compression), but there is no
303 support for random access, file deduplication, multiple images per
304 archive, or extended attributes.
307 Zip shares some features with WIM but is not designed to store entire
311 The 7z format has some nice features but is unfortunately not designed
316 See the manual pages for `imagex', the manual pages for the subcommands of
317 `imagex', and the manual page for `mkwinpeimg'.
319 As of version 0.5.0, wimlib's public API is documented. Doxygen is required to
320 build the documentation. To build the documentation, run `configure', then
321 enter the directory `doc' and run `doxygen'. The HTML documentation will be
322 created in a directory named `html'.
326 As of version 1.0.0, wimlib is released under the GNU GPL version 3.0 or later.
327 This includes the files in the `programs' directory as well as the files in the
330 wimlib is independently developed and does not contain any code, data, or files
331 copyrighted by Microsoft. It is not known to be affected by any patents.
333 By default, wimlib will be linked to the system library "libcrypto", which
334 probably will be OpenSSL. Some people believe that GPL code cannot be linked to
335 OpenSSL without a linking exception. As far as I know, I cannot officially
336 include a linking exception with the license of this library because several
337 files could be considered derived works of code copyrighted by others. If you
338 believe this to be a problem, configure with --without-libcrypto to avoid
339 linking with OpenSSL. There is no difference in functionality--- there will
340 just be stand-alone SHA1 message digest code built into the library.
344 wimlib is experimental. Use Microsoft's `imagex.exe' if you want to make sure
345 your WIM files are made correctly (but beware: Microsoft's version contains some
348 Please submit a bug report (to ebiggers3@gmail.com) if you find a bug in wimlib.
350 Some parts of the WIM file format are poorly documented or even completely
351 undocumented, so I've just had to do the best I can to read and write WIMs that
352 appear to be compatible with Microsoft's software.