3 This is wimlib version 1.7.1-BETA (June 2014). wimlib is a C library for
4 creating, modifying, extracting, and mounting files in the Windows Imaging
5 Format (WIM files). These files are normally created using the ImageX
6 (imagex.exe) or Dism (Dism.exe) utilities on Windows, but wimlib is distributed
7 with a free implementation of ImageX called "wimlib-imagex" for both UNIX-like
12 To install wimlib and wimlib-imagex on Windows you simply need to download and
13 extract the ZIP file containing the latest binaries from the SourceForge page
14 (http://sourceforge.net/projects/wimlib/), which you may have already done.
16 To install wimlib and wimlib-imagex on UNIX-like systems (with Linux being the
17 primary supported and tested platform), you must compile the source code, which
18 is also available at http://sourceforge.net/projects/wimlib/. Alternatively,
19 check if a package has been prepared for your Linux distribution. Example files
20 for Debian and RPM packaging are in the debian/ and rpm/ directories.
24 A Windows Imaging (WIM) file is an archive designed primarily for archiving
25 Windows filesystems. However, it can be used on other platforms as well, with
26 some limitations. Like some other archive formats such as ZIP, files in WIM
27 archives may be compressed. WIM files support multiple compression formats,
28 including LZX, XPRESS, and LZMS. All these formats are supported by wimlib.
30 A WIM file consists of one or more "images". Each image is an independent
31 top-level directory structure and is logically separate from all other images in
32 the WIM. Each image has a name as well as a 1-based index in the WIM file. To
33 save space, WIM archives automatically combine all duplicate files across all
36 A WIM file may be either stand-alone or split into multiple parts. Split WIMs
37 are read-only and cannot be modified.
39 Since version 1.6.0, wimlib also supports ESD (.esd) files, except when
40 encrypted. These are still WIM files but they use a newer version of the file
45 wimlib itself is a C library, and it provides a documented public API (See:
46 http://wimlib.sourceforge.net) for other programs to use. However, it is also
47 distributed with a command-line program called "wimlib-imagex" that uses this
48 library to implement an imaging tool similar to Microsoft's ImageX.
49 wimlib-imagex supports almost all the capabilities of Microsoft's ImageX as well
50 as additional capabilities. wimlib-imagex works on both UNIX-like systems and
51 Windows, although some features differ between the platforms.
53 Run `wimlib-imagex' with no arguments to see an overview of the available
54 commands and their syntax. For additional documentation:
56 * If you have installed wimlib-imagex on a UNIX-like system, you will find
57 further documentation in the man pages; run `man wimlib-imagex' to get
60 * If you have downloaded the Windows binary distribution, you will find the
61 documentation for wimlib-imagex in PDF format in the "doc" directory,
62 ready for viewing with any PDF viewer. Please note that although the PDF
63 files are converted from UNIX-style "man pages", they do document
64 Windows-specific behavior when appropriate.
68 wimlib (and wimlib-imagex) can create XPRESS, LZX, or LZMS compressed WIM files.
69 wimlib includes its own compression codecs and does not use the compression API
70 available on some versions of Windows.
72 I have gradually been improving the compression codecs in wimlib, and in some
73 cases they now outperform and outcompress the equivalent Microsoft
74 implementations. Although results will vary depending on the data being
75 compressed, in the table below I present the results for a common use case:
76 compressing an x86 Windows PE image. Each row displays the compression type,
77 the size of the resulting WIM file in bytes, and how many seconds it took to
78 create the file. When applicable, the results with the equivalent Microsoft
79 implementation in WIMGAPI is included.
81 =============================================================================
82 | Compression type || wimlib (v1.7.0) | WIMGAPI (Windows 8.1) |
83 =============================================================================
84 | None [1] || 361,182,560 in 7.6s | 361,183,674 in 11.9s |
85 | XPRESS [2] || 138,992,007 in 11.5s | 140,416,657 in 13.9s |
86 | LZX (quick) [3] || 131,428,113 in 15.2s | N/A |
87 | LZX (normal) [4] || 126,807,853 in 58.9s | 127,259,566 in 55.2s |
88 | LZX (slow) [5] || 126,085,971 in 125s | N/A |
89 | LZMS (non-solid) [6] || 122,082,982 in 59.8s | N/A |
90 | LZMS (solid) [7] || 92,171,126 in 260s | 88,742,356 in 556s |
91 =============================================================================
94 [1] '--compress=none' for wimlib-imagex;
95 '/compress:none' for DISM.
97 [2] '--compress=fast' or '--compress=XPRESS' for wimlib-imagex;
98 '/compress:fast' for DISM.
99 Compression chunk size is 32768 bytes, the default for XPRESS.
101 [3] No compression option specified to wimlib-imagex;
102 no known equivalent for DISM.
103 Compression chunk size is 32768 bytes, the default for LZX.
105 [4] '--compress=maximum' or '--compress=LZX' for wimlib-imagex;
106 '/compress:maximum' for DISM.
107 Compression chunk size is 32768 bytes, the default for LZX.
109 [5] '--compress=maximum --compress-slow' for wimlib-imagex;
110 no known equivalent for DISM.
111 Compression chunk size is 32768 bytes, the default for LZX.
113 [6] '--compress=recovery' or '--compress=LZMS' for wimlib-imagex;
114 no known equivalent for DISM.
115 Compression chunk size is 131072 bytes, the default for LZMS.
117 [7] '--compress=recovery --solid' or '--compress=LZMS --solid' for
118 wimlib-imagex. Should be '/compress:recovery' for DISM, but it doesn't
119 work; I called WIMGAPI directly instead.
120 Compression chunk size in solid blocks is 67108864 bytes in both cases.
122 I did the timings on Windows 8.1 running in a virtual machine so that
123 side-by-side comparisons with the Microsoft implementation would be possible.
124 However, do note that wimlib may have even better performance on other operating
125 systems such as Linux. I used the 64-bit builds of both programs.
127 The compression ratio provided by wimlib is also competitive with commonly used
128 archive formats. Below are file sizes that result when the Canterbury corpus is
129 compressed with wimlib (v1.7.0), WIMGAPI (Windows 8), and some other
132 =================================================
133 | Format | Size (bytes) |
134 =================================================
136 | WIM (WIMGAPI, None) | 2,814,278 |
137 | WIM (wimlib, None) | 2,813,856 |
138 | WIM (WIMGAPI, XPRESS) | 825,410 |
139 | WIM (wimlib, XPRESS) | 792,024 |
140 | tar.gz (gzip, default) | 738,796 |
141 | ZIP (Info-ZIP, default) | 735,334 |
142 | tar.gz (gzip, -9) | 733,971 |
143 | ZIP (Info-ZIP, -9) | 732,297 |
144 | WIM (wimlib, LZX quick) | 722,196 |
145 | WIM (WIMGAPI, LZX) | 651,766 |
146 | WIM (wimlib, LZX normal) | 649,204 |
147 | WIM (wimlib, LZX slow) | 639,618 |
148 | WIM (wimlib, LZMS non-solid) | 592,136 |
149 | tar.bz2 (bzip, default) | 565,008 |
150 | tar.bz2 (bzip, -9) | 565,008 |
151 | WIM (wimlib, LZMS solid) | 525,270 |
152 | WIM (wimlib, LZMS solid, slow) | 521,700 |
153 | WIM (WIMGAPI, LZMS solid) | 521,232 |
154 | tar.xz (xz, default) | 486,916 |
155 | tar.xz (xz, -9) | 486,904 |
156 | 7z (7-zip, default) | 484,700 |
157 | 7z (7-zip, -9) | 483,239 |
158 =================================================
160 Note: WIM does even better on directory trees containing duplicate files, which
161 the Canterbury corpus doesn't have.
165 WIM images may contain data, such as alternate data streams and
166 compression/encryption flags, that are best represented on the NTFS filesystem
167 used on Windows. Also, WIM images may contain security descriptors which are
168 specific to Windows and cannot be represented on other operating systems.
169 wimlib handles this NTFS-specific or Windows-specific data in a
170 platform-dependent way:
172 * In the Windows version of wimlib and wimlib-imagex, NTFS-specific and
173 Windows-specific data are supported natively.
175 * In the UNIX version of wimlib and wimlib-imagex, NTFS-specific and
176 Windows-specific data are ordinarily ignored; however, there is also special
177 support for capturing and extracting images directly to/from unmounted NTFS
178 volumes. This was made possible with the help of libntfs-3g from the
181 For both platforms the code for NTFS capture and extraction is complete enough
182 that it is possible to apply an image from the "install.wim" contained in recent
183 Windows installation media (Vista, Windows 7, or Windows 8) directly to an NTFS
184 filesystem, and then boot Windows from it after preparing the Boot Configuration
185 Data. In addition, a Windows installation can be captured (or backed up) into a
186 WIM file, and then re-applied later.
190 A major use for wimlib and wimlib-imagex is to create customized images of
191 Windows PE, the Windows Preinstallation Environment, on either UNIX-like systems
192 or Windows without having to rely on Microsoft's software and its restrictions
195 Windows PE is a lightweight version of Windows that can run entirely from memory
196 and can be used to install Windows from local media or a network drive or
197 perform maintenance. It is the operating system that runs when you boot from
198 the Windows installation media.
200 You can find Windows PE on the installation DVD for Windows Vista, Windows 7, or
201 Windows 8, in the file `sources/boot.wim'. Windows PE can also be found in the
202 Windows Automated Installation Kit (WAIK), which is free to download from
203 Microsoft, inside the `WinPE.cab' file, which you can extract natively on
204 Windows, or on UNIX-like systems if you install either the `cabextract' or
207 In addition, Windows installations and recovery partitions frequently contain a
208 WIM containing an image of the Windows Recovery Environment, which is similar to
211 A shell script `mkwinpeimg' is distributed with wimlib on UNIX-like systems to
212 ease the process of creating and customizing a bootable Windows PE image.
216 This section documents the dependencies of wimlib and the programs distributed
217 with it, when building for a UNIX-like system from source. If you have
218 downloaded the Windows binary distribution of wimlib and wimlib-imagex then all
219 dependencies were already included and this section is irrelevant.
222 This is a commonly used free library to read and write XML files. You
223 likely already have it installed as a dependency for some other program.
224 For more information see http://xmlsoft.org/.
226 * libfuse (optional but highly recommended)
227 Unless configured with --without-fuse, wimlib requires a non-ancient
228 version of libfuse to be installed. Most Linux distributions already
229 include this, but make sure you have the libfuse package installed, and
230 also libfuse-dev if your distribution distributes header files
231 separately. FUSE also requires a kernel module. If the kernel module
232 is available it will automatically be loaded if you try to mount a WIM
233 file. For more information see http://fuse.sourceforge.net/. FUSE is
234 also available for FreeBSD.
236 * libntfs-3g (optional but highly recommended)
237 Unless configured with --without-ntfs-3g, wimlib requires the library
238 and headers for libntfs-3g version 2011-4-12 or later to be installed.
239 Versions dated 2010-3-6 and earlier do not work because they are missing
240 the header xattrs.h (and the file xattrs.c, which contains functions we
241 need). libntfs-3g version 2013-1-13 is compatible only with wimlib
244 * OpenSSL / libcrypto (optional)
245 wimlib can use the SHA1 message digest code from OpenSSL instead of
246 compiling in yet another SHA1 implementation. (See LICENSE section.)
250 * syslinux (optional)
251 * cabextract (optional)
252 The `mkwinpeimg' shell script will look for several other programs
253 depending on what options are given to it. Depending on your Linux
254 distribution, you may already have these programs installed, or they may
255 be in the software repository. Making an ISO filesystem requires
256 `mkisofs' from `cdrkit' (http://www.cdrkit.org). Making a disk image
257 requires `mtools' (http://www.gnu.org/software/mtools) and `syslinux'
258 (http://www.syslinux.org). Retrieving files from the Windows Automated
259 Installation Kit requires `cabextract' (http://www.cabextract.org.uk).
263 This section documents the most important options that may be passed to the
264 "configure" script when building from source:
267 If libntfs-3g is not available or is not version 2011-4-12 or later,
268 wimlib can be built without it, in which case it will not be possible to
269 apply or capture images directly to/from NTFS volumes.
272 If libfuse or the FUSE kernel module is not available, wimlib can be
273 compiled with --without-fuse. This will remove the ability to mount and
277 Build in functions for SHA1 rather than using external SHA1 functions
278 from libcrypto (part of OpenSSL). The default is to use libcrypto if it
279 is found on the system.
281 --disable-multithreaded-compression
282 By default, data will be compressed using multiple threads when writing
283 a WIM, unless only 1 processor is detected. Specify this option to
284 disable support for this.
287 Use a very fast assembly language implementation of SHA1 from Intel.
288 Only use this if the build target supports the SSSE3 instructions.
290 --disable-error-messages
291 Save some space by removing all error messages from the library.
294 Remove assertions included by default.
298 wimlib has primarily been tested on Linux and Windows (primarily Windows 7, but
299 also Windows XP and Windows 8).
301 wimlib may work on FreeBSD and Mac OS X. However, this is not well tested. If
302 you do not have libntfs-3g 2011-4-12 or later available, you must configure
303 wimlib with --without-ntfs-3g. On FreeBSD, before mounting a WIM you need to
304 load the POSIX message queue module (run `kldload mqueuefs').
306 The code has primarily been tested on x86 and x86_64 CPUs, but it's written to
307 be portable to other architectures and I've also tested it on ARM. However,
308 although the code is written to correctly deal with endianness, it has not yet
309 actually been tested on a big-endian architecture.
313 The WIM file format is partially specified in a document that can be found in
314 the Microsoft Download Center. However, this document really only provides an
315 overview of the format and is not a formal specification.
317 With regards to the supported compression formats:
319 - Microsoft has official documentation for XPRESS that is of reasonable quality.
320 - Microsoft has official documentation for LZX but it contains errors.
321 - There does not seem to be any official documentation for LZMS, so my comments
322 and code in src/lzms-decompress.c may in fact be the best documentation
323 available for this particular compression format.
325 The code in ntfs-3g_apply.c and ntfs-3g_capture.c uses the NTFS-3g library,
326 which is a library for reading and writing to NTFS filesystems (the filesystem
327 used by recent versions of Windows). See
328 http://www.tuxera.com/community/ntfs-3g-download/ for more information.
330 The LZX decompressor (lzx-decompress.c) was originally based on code from the
331 cabextract project (http://www.cabextract.org.uk). The LZX compressor
332 (lzx-compress.c) was originally based on code written by Matthew Russotto
333 (www.russotto.net/chm/). However I have since rewritten and made many
334 improvements to both the decompressor and compressor.
336 lz_binary_trees.c contains LZ77 match-finding code that uses binary trees. It
337 is based on code from liblzma but I have since rewritten it.
339 A limited number of other free programs can handle some parts of the WIM
342 * 7-zip is able to extract and create WIMs (as well as files in many
343 other archive formats). However, wimlib is designed specifically to handle
344 WIM files and provides features previously only available in Microsoft's
345 implementation, such as the ability to mount WIMs read-write as well as
346 read-only, the ability to create compressed WIMs, and the correct handling
347 of security descriptors and hard links.
348 * ImagePyX (https://github.com/maxpat78/ImagePyX) is a Python program that
349 provides similar capabilities to wimlib-imagex. One thing to note, though,
350 is that it does not support compression and decompression by itself, but
351 instead relies on external native code, such as the codecs from wimlib.
353 If you are looking for an archive format that provides features similar to WIM
354 but was designed primarily for UNIX, you may want to consider SquashFS
355 (http://squashfs.sourceforge.net/). However, you may find that wimlib works
356 surprisingly well on UNIX. It will store hard links and symbolic links, and it
357 has optional support for storing UNIX owners, groups, modes, and special files
358 such as device nodes and FIFOs. Actually, I use it to back up my own files on
361 LICENSE AND DISCLAIMER
363 See COPYING for information about the license.
365 wimlib is independently developed and does not contain any code, data, or files
366 copyrighted by Microsoft. It is not known to be affected by any patents.
368 On UNIX-like systems, if you do not want wimlib to be dynamically linked with
369 libcrypto (OpenSSL), configure with --without-libcrypto. This replaces the SHA1
370 implementation with built-in code and there will be no difference in
373 wimlib comes with no warranty whatsoever. Please submit a bug report (to
374 ebiggers3@gmail.com) if you find a bug in wimlib and/or wimlib-imagex.
376 Be aware that some parts of the WIM file format are poorly documented or even
377 completely undocumented, so I've just had to do the best I can to read and write
378 WIMs that appear to be compatible with Microsoft's software.