8 1.2. Platform-specific notes
15 1.3. Adding support for new platforms
17 3. xzgrep and other scripts
21 4.1. "No C99 compiler was found."
22 4.1. "No POSIX conforming shell (sh) was found."
23 4.2. configure works but build fails at crc32_x86.S
29 If you aren't familiar with building packages that use GNU Autotools,
30 see the file INSTALL.generic for generic instructions before reading
33 If you are going to build a package for distribution, see also the
34 file PACKAGERS. It contains information that should help making the
35 binary packages as good as possible, but the information isn't very
36 interesting to those making local builds for private use or for use
37 in special situations like embedded systems.
40 1. Supported platforms
41 ----------------------
43 XZ Utils are developed on GNU/Linux, but they should work on many
44 POSIX-like operating systems like *BSDs and Solaris, and even on
45 a few non-POSIX operating systems.
50 A C99 compiler is required to compile XZ Utils. If you use GCC, you
51 need at least version 3.x.x. GCC version 2.xx.x doesn't support some
52 C99 features used in XZ Utils source code, thus GCC 2 won't compile
55 XZ Utils takes advantage of some GNU C extensions when building
56 with GCC. Because these extensions are used only when building
57 with GCC, it should be possible to use any C99 compiler.
60 1.2. Platform-specific notes
64 MIPSpro 7.4.4m has been reported to produce broken code if using
65 the -O2 optimization flag ("make check" fails). Using -O1 should
71 If you try to use the native C compiler on Tru64 (passing CC=cc to
72 configure), it is possible that the configure script will complain
73 that no C99 compiler was found even when the native compiler supports
74 C99. You can safely override the test for C99 compiler by passing
75 ac_cv_prog_cc_c99= as the argument to the configure script.
80 Building XZ Utils on Windows is supported under MinGW + MSYS and
81 Cygwin. There is windows/build.bash to ease packaging XZ Utils with
82 MinGW + MSYS into a redistributable .zip or .7z file. See
83 windows/INSTALL-Windows.txt for more information.
85 It might be possible to build liblzma with a non-GNU toolchain too,
86 but that will probably require writing a separate makefile. Building
87 the command line tools with non-GNU toolchains will be harder than
88 building only liblzma.
90 Even if liblzma is built with MinGW, the resulting DLL or static
91 library can be used by other compilers and linkers, including MSVC.
92 Thus, it shouldn't be a problem to use MinGW to build liblzma even
93 if you cannot use MinGW to build the rest of your project. See
94 windows/README-Windows.txt for details.
99 There is an experimental Makefile in the "dos" directory to build
100 XZ Utils on DOS using DJGPP. Support for long file names (LFN) is
101 needed. See dos/README for more information.
103 GNU Autotools based build hasn't been tried on DOS. If you try, I
104 would like to hear if it worked.
109 To omit large number of harmless warnings about visibility support,
110 pass gl_cv_cc_visibility=no as an argument to the configure script.
111 This isn't mandatory since it should have no effect on the resulting
117 XZ Utils can be built for OpenVMS, but the build system files are
118 currently not included in the XZ Utils source package. The required
119 OpenVMS-specific files are maintained by Jouk Jansen and can be
122 http://nchrem.tnw.tudelft.nl/openvms/software2.html#xzutils
125 1.3. Adding support for new platforms
127 If you have written patches to make XZ Utils to work on previously
128 unsupported platform, please send the patches to me! I will consider
129 including them to the official version. It's nice to minimize the
130 need of third-party patching.
132 One exception: Don't request or send patches to change the whole
133 source package to C89. I find C99 substantially nicer to write and
134 maintain. However, the public library headers must be in C89 to
135 avoid frustrating those who maintain programs, which are strictly
142 In most cases, the defaults are what you want. Most of the options
143 below are useful only when building a size-optimized version of
144 liblzma or command line tools.
146 --enable-encoders=LIST
148 Specify a comma-separated LIST of filter encoders to
149 build. See "./configure --help" for exact list of
150 available filter encoders. The default is to build all
153 If LIST is empty or --disable-encoders is used, no filter
154 encoders will be built and also the code shared between
155 encoders will be omitted.
157 Disabling encoders will remove some symbols from the
158 liblzma ABI, so this option should be used only when it
159 is known to not cause problems.
161 --enable-decoders=LIST
163 This is like --enable-encoders but for decoders. The
164 default is to build all supported decoders.
166 --enable-match-finders=LIST
167 liblzma includes two categories of match finders:
168 hash chains and binary trees. Hash chains (hc3 and hc4)
169 are quite fast but they don't provide the best compression
170 ratio. Binary trees (bt2, bt3 and bt4) give excellent
171 compression ratio, but they are slower and need more
172 memory than hash chains.
174 You need to enable at least one match finder to build the
175 LZMA1 or LZMA2 filter encoders. Usually hash chains are
176 used only in the fast mode, while binary trees are used to
177 when the best compression ratio is wanted.
179 The default is to build all the match finders if LZMA1
180 or LZMA2 filter encoders are being built.
183 liblzma support multiple integrity checks. CRC32 is
184 mandatory, and cannot be omitted. See "./configure --help"
185 for exact list of available integrity check types.
187 liblzma and the command line tools can decompress files
188 which use unsupported integrity check type, but naturally
189 the file integrity cannot be verified in that case.
191 Disabling integrity checks may remove some symbols from
192 the liblzma ABI, so this option should be used only when
193 it is known to not cause problems.
196 liblzma includes some assembler optimizations. Currently
197 there is only assembler code for CRC32 and CRC64 for
200 All the assembler code in liblzma is position-independent
201 code, which is suitable for use in shared libraries and
202 position-independent executables. So far only i386
203 instructions are used, but the code is optimized for i686
204 class CPUs. If you are compiling liblzma exclusively for
205 pre-i686 systems, you may want to disable the assembler
208 --enable-unaligned-access
209 Allow liblzma to use unaligned memory access for 16-bit
210 and 32-bit loads and stores. This should be enabled only
211 when the hardware supports this, i.e. when unaligned
212 access is fast. Some operating system kernels emulate
213 unaligned access, which is extremely slow. This option
214 shouldn't be used on systems that rely on such emulation.
216 Unaligned access is enabled by default on x86, x86-64,
217 and big endian PowerPC.
220 Reduce the size of liblzma by selecting smaller but
221 semantically equivalent version of some functions, and
222 omit precomputed lookup tables. This option tends to
223 make liblzma slightly slower.
225 Note that while omitting the precomputed tables makes
226 liblzma smaller on disk, the tables are still needed at
227 run time, and need to be computed at startup. This also
228 means that the RAM holding the tables won't be shared
229 between applications linked against shared liblzma.
231 This option doesn't modify CFLAGS to tell the compiler
232 to optimize for size. You need to add -Os or equivalent
233 flag(s) to CFLAGS manually.
235 --enable-assume-ram=SIZE
236 On the most common operating systems, XZ Utils is able to
237 detect the amount of physical memory on the system. This
238 information is used by the options --memlimit-compress,
239 --memlimit-decompress, and --memlimit when setting the
240 limit to a percentage of total RAM.
242 On some systems, there is no code to detect the amount of
243 RAM though. Using --enable-assume-ram one can set how much
244 memory to assume on these systems. SIZE is given as MiB.
245 The default is 128 MiB.
247 Feel free to send patches to add support for detecting
248 the amount of RAM on the operating system you use. See
249 src/common/tuklib_physmem.c for details.
252 Disable threading support. This makes some things
253 thread-unsafe, meaning that if multithreaded application
254 calls liblzma functions from more than one thread,
255 something bad may happen.
257 Use this option if threading support causes you trouble,
258 or if you know that you will use liblzma only from
259 single-threaded applications and want to avoid dependency
262 --enable-dynamic=TYPE
263 Specify how command line tools should be linked against
264 liblzma. Possible TYPES:
266 yes All command line tools are linked against
267 shared liblzma (if shared liblzma was built).
268 This is equivalent to --enable-dynamic (i.e.
271 mixed Some tools are linked against static liblzma
272 and some against shared liblzma. This is the
273 default and recommended way.
275 no All command line tools are linked against
276 static liblzma (if static liblzma was built).
277 This is equivalent to --disable-dynamic.
279 This option is mostly useful for packagers, if distro
280 policy requires linking against shared libaries. See the
281 file PACKAGERS for more information about pros and cons
285 This enables the assert() macro and possibly some other
286 run-time consistency checks. It makes the code slower, so
287 you normally don't want to have this enabled.
290 If building with GCC, make all compiler warnings an error,
291 that abort the compilation. This may help catching bugs,
292 and should work on most systems. This has no effect on the
296 3. xzgrep and other scripts
297 ---------------------------
301 POSIX shell (sh) and bunch of other standard POSIX tools are required
302 to run the scripts. The configure script tries to find a POSIX
303 compliant sh, but if it fails, you can force the shell by passing
304 gl_cv_posix_shell=/path/to/posix-sh as an argument to the configure
307 Some of the scripts require also mktemp. The original mktemp can be
308 found from <http://www.mktemp.org/>. On GNU, most will use the mktemp
309 program from GNU coreutils instead of the original implementation.
310 Both mktemp versions are fine for XZ Utils (and practically for
311 everything else too).
316 The scripts assume that the required tools (standard POSIX utilities,
317 mktemp, and xz) are in PATH; the scripts don't set the PATH themselves.
318 Some people like this while some think this is a bug. Those in the
319 latter group can easily patch the scripts before running the configure
320 script by taking advantage of a placeholder line in the scripts.
322 For example, to make the scripts prefix /usr/bin:/bin to PATH:
324 perl -pi -e 's|^#SET_PATH.*$|PATH=/usr/bin:/bin:\$PATH|' \
331 4.1. "No C99 compiler was found."
333 You need a C99 compiler to build XZ Utils. If the configure script
334 cannot find a C99 compiler and you think you have such a compiler
335 installed, set the compiler command by passing CC=/path/to/c99 as
336 an argument to the configure script.
338 If you get this error even when you think your compiler supports C99,
339 you can override the test by passing ac_cv_prog_cc_c99= as an argument
340 to the configure script. The test for C99 compiler is not perfect (and
341 it is not as easy to make it perfect as it sounds), so sometimes this
342 may be needed. You will get a compile error if your compiler doesn't
346 4.1. "No POSIX conforming shell (sh) was found."
348 xzgrep and other scripts need a shell that (roughly) conforms
349 to POSIX. The configure script tries to find such a shell. If
350 it fails, you can force the shell to be used by passing
351 gl_cv_posix_shell=/path/to/posix-sh as an argument to the configure
355 4.2. configure works but build fails at crc32_x86.S
357 The easy fix is to pass --disable-assembler to the configure script.
359 The configure script determines if assembler code can be used by
360 looking at the configure triplet; there is currently no check if
361 the assembler code can actually actually be built. The x86 assembler
362 code should work on x86 GNU/Linux, *BSDs, Solaris, Darwin, MinGW,
363 Cygwin, and DJGPP. On other x86 systems, there may be problems and
364 the assembler code may need to be disabled with the configure option.
366 If you get this error when building for x86-64, you have specified or
367 the configure script has misguessed your architecture. Pass the
368 correct configure triplet using the --build=CPU-COMPANY-SYSTEM option
369 (see INSTALL.generic).