mirror of
https://review.haiku-os.org/buildtools
synced 2024-11-23 07:18:49 +01:00
88896a1144
Change-Id: I42dc2a783b3e0082a7583cc7ecaa6b043ba162f4 Reviewed-on: https://review.haiku-os.org/c/buildtools/+/3021 Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
720 lines
31 KiB
Plaintext
720 lines
31 KiB
Plaintext
Copyright 1999-2020 Free Software Foundation, Inc.
|
|
Contributed by the AriC and Caramba projects, INRIA.
|
|
|
|
This file is part of the GNU MPFR Library.
|
|
|
|
The GNU MPFR Library is free software; you can redistribute it and/or modify
|
|
it under the terms of the GNU Lesser General Public License as published by
|
|
the Free Software Foundation; either version 3 of the License, or (at your
|
|
option) any later version.
|
|
|
|
The GNU MPFR Library is distributed in the hope that it will be useful, but
|
|
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
|
|
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
|
|
License for more details.
|
|
|
|
You should have received a copy of the GNU Lesser General Public License
|
|
along with the GNU MPFR Library; see the file COPYING.LESSER. If not, see
|
|
https://www.gnu.org/licenses/ or write to the Free Software Foundation, Inc.,
|
|
51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA.
|
|
|
|
|
|
|
|
Installing GNU MPFR
|
|
===================
|
|
|
|
Note: In case of problem, please read this INSTALL file carefully before
|
|
reporting a bug, in particular Section "In case of problem" below. Some
|
|
problems are due to bad configuration on the user side (not specific to
|
|
MPFR).
|
|
|
|
0. You first need to install GMP. See <https://gmplib.org/>.
|
|
MPFR requires GMP version 5.0.0 or later.
|
|
|
|
1. Extract the files from the archive.
|
|
|
|
2. It is strongly advised to apply the latest patches if this has
|
|
not been done yet and if patches are available. You can check
|
|
on the release page:
|
|
|
|
https://www.mpfr.org/mpfr-4.1.0/
|
|
|
|
which may have additional information. The patches can be applied
|
|
with commands like:
|
|
|
|
wget https://www.mpfr.org/mpfr-4.1.0/allpatches
|
|
patch -N -Z -p1 < allpatches
|
|
|
|
or
|
|
|
|
curl https://www.mpfr.org/mpfr-4.1.0/allpatches | patch -N -Z -p1
|
|
|
|
(Those instructions are for the GNU patch command, for example
|
|
/usr/bin/gpatch on Solaris.)
|
|
|
|
3. In the MPFR directory, to detect your system, type:
|
|
./configure
|
|
possibly with options (see below, in particular if this step or
|
|
one of the following fails). You should also check whether WARNING
|
|
lines have been output (such a problem may cause a failure in one
|
|
of the following steps).
|
|
Note: paths provided in configure options must always be absolute
|
|
(relative paths are not supported).
|
|
|
|
4. To build the library, type:
|
|
make
|
|
|
|
[optional] if you want to tune MPFR for your specific architecture, see
|
|
the section "Tuning MPFR" below. Note that for most common architectures,
|
|
MPFR includes some default tuning parameters which should be near from
|
|
optimal.
|
|
|
|
5. To check the built library (runs the test files), type:
|
|
make check
|
|
|
|
Note: If any test fails, information about this failure can be found in
|
|
the tests/test-suite.log file; you should provide this file in your bug
|
|
reports (in addition to other useful information, as mentioned later).
|
|
If you want the contents of this file to be automatically output in case
|
|
of failure, you can set the VERBOSE environment variable to 1 before
|
|
running "make check", for instance by typing:
|
|
VERBOSE=1 make check
|
|
|
|
6. [Optional / experimental] Binary distributions may also want to run:
|
|
make check-gmp-symbols
|
|
|
|
This will check that MPFR does not use GMP internal symbols, which
|
|
could yield failures in case of GMP upgrade without a MPFR rebuild.
|
|
But note that this is a heuristic and might give false positives or
|
|
false negatives. Please report any problem to the MPFR developers.
|
|
End users may also be interested in this check, unless they have
|
|
allowed GMP internals with configure options (see below).
|
|
|
|
7. To install it (default "/usr/local" | see "--prefix" option), type:
|
|
make install
|
|
|
|
If you installed MPFR (header and library) in directories that are
|
|
not searched by default by the compiler and/or linking tools, then,
|
|
like with other libraries, you may need to set up some environment
|
|
variables such as C_INCLUDE_PATH (to find the header mpfr.h),
|
|
LIBRARY_PATH (to find the library), and if the shared library has
|
|
been installed, LD_LIBRARY_PATH (before execution) or LD_RUN_PATH
|
|
(before linking); this list is not exhaustive and some environment
|
|
variables may be specific to your system. "make install" gives some
|
|
instructions; please read them. You can also find more information
|
|
in the manuals of your compiler and linker. The MPFR FAQ may also
|
|
give some information.
|
|
|
|
Remember that if you have several MPFR (or GMP) versions installed
|
|
(e.g., one with the system, and one, newer, by you), you will not
|
|
necessarily get a compilation/linking error if a wrong library is
|
|
used (e.g., because LD_LIBRARY_PATH has not been set correctly).
|
|
But unexpected results may occur.
|
|
|
|
Under Mac OS X, if the shared library was not installed and you use
|
|
Apple's linker (this is the default), you will also need to provide
|
|
the -search_paths_first linker flag ("-Wl,-search_paths_first" when
|
|
you link via gcc) to make sure that the right library is selected,
|
|
as by default, Apple's linker selects a shared library preferably,
|
|
even when it is farther in the library paths. We recall that if a
|
|
wrong library is selected due to this behavior, unexpected results
|
|
may occur.
|
|
|
|
|
|
Building the documentation
|
|
==========================
|
|
|
|
To build the documentation in various formats, you may first need to
|
|
install recent versions of some utilities such as texinfo.
|
|
|
|
* Type "make info" to produce the documentation in the info format.
|
|
|
|
* Type "make pdf" to produce the documentation in the PDF format.
|
|
|
|
* Type "make dvi" to produce the documentation in the DVI format.
|
|
|
|
* Type "make ps" to produce the documentation in the Postscript format.
|
|
|
|
* Type "make html" to produce the documentation in the HTML format
|
|
(in several pages); if you want only one output HTML file, then
|
|
type "makeinfo --html --no-split mpfr.texi" from the doc directory
|
|
instead.
|
|
|
|
|
|
Building MPFR with internal GMP header files
|
|
============================================
|
|
|
|
MPFR built with internal GMP header files is a bit faster, so you may want
|
|
to build it with them. Just do this in step 1:
|
|
|
|
./configure --with-gmp-build=GMPBUILD
|
|
|
|
where GMPBUILD is the GMP build directory. The needed header files are:
|
|
gmp-impl.h, longlong.h and all the necessary headers to use them, which
|
|
may be located either in the GMP source directory or in the GMP build
|
|
directory, in case they are different (MPFR takes care of that, as of
|
|
MPFR 3.1.0).
|
|
|
|
Warning: the library obtained in this way may use some internal GMP
|
|
symbols, and thus dynamically linking your software with a different
|
|
version of GMP might fail, even though it is declared as compatible
|
|
by Libtool's versioning system.
|
|
|
|
|
|
Tuning MPFR
|
|
===========
|
|
|
|
[For the current GMP version (6.1.0), a Unix-like OS is required.]
|
|
|
|
For this, you need to build MPFR with a GMP build directory (see above).
|
|
In the GMP build directory, you also need to go into the "tune" subdirectory
|
|
and type "make speed". This will build the GMP speed library, which is used
|
|
by the MPFR tuning mechanism.
|
|
|
|
Then go back to the MPFR build directory, go into the "tune" subdirectory and
|
|
type "make tune". This will build an optimized file "mparam.h" for your
|
|
specific architecture.
|
|
|
|
|
|
./configure options
|
|
===================
|
|
|
|
--prefix=DIR installs MPFR headers and library in DIR/include and
|
|
DIR/lib respectively (the default is "/usr/local").
|
|
|
|
--with-gmp-include=DIR assumes that DIR contains gmp.h
|
|
--with-gmp-lib=DIR assumes that DIR contains the GMP library
|
|
--with-gmp=DIR assumes that DIR is where you have installed GMP.
|
|
same as --with-gmp-lib=DIR/lib
|
|
and --with-gmp-include=DIR/include
|
|
(use either --with-gmp alone or one or both of
|
|
--with-gmp-lib/--with-gmp-include)
|
|
Warning! Do not use these options if you have
|
|
CPPFLAGS and/or LDFLAGS containing a -I or -L
|
|
option with a directory that contains a GMP
|
|
header or library file, as these options just
|
|
add -I and -L options to CPPFLAGS and LDFLAGS
|
|
*after* the ones that are currently declared,
|
|
so that DIR will have a lower precedence. Also,
|
|
this may not work if DIR is a system directory
|
|
(typically /usr or /usr/local); see below.
|
|
|
|
--with-gmp-build=DIR assumes that DIR contains the GMP build directory,
|
|
and enables the use of GMP internals (see above).
|
|
Warning! This option and the group of options
|
|
--with-gmp are mutually exclusive.
|
|
|
|
--enable-assert build MPFR with assertions.
|
|
|
|
--enable-thread-safe build MPFR as thread safe, using compiler-level
|
|
Thread-Local Storage (TLS). Note: TLS support is
|
|
roughly tested by configure. If configure detects
|
|
that TLS does not work (because of the compiler,
|
|
linker or system libraries), it will output an
|
|
error message, telling you to build MPFR without
|
|
thread safe. For instance, though Mac OS X uses
|
|
GCC, it may not currently support GCC's __thread
|
|
storage class.
|
|
|
|
--disable-thread-safe build MPFR without TLS. By default, TLS support
|
|
is detected automatically, and MPFR is built as
|
|
thread safe if supported. However this detection
|
|
is only a heuristic: TLS can be detected as
|
|
supported while its support is incomplete or
|
|
buggy (MPFR tests may fail). In such a case,
|
|
this option is useful.
|
|
|
|
--enable-decimal-float build conversion functions from/to decimal floats.
|
|
Note that detection by the configure script is
|
|
limited in case of cross-compilation.
|
|
Accepted arguments:
|
|
yes Decimal support is requested and the configure
|
|
script fails if it detects that decimals do not
|
|
work.
|
|
The encoding (BID or DPD) will automatically be
|
|
detected at configure time or at compile time if
|
|
possible (if not, generic code will be used).
|
|
no Decimal support is explicitly disabled.
|
|
auto Decimal support is enabled if the configure script
|
|
detects that it works. This is the default when
|
|
--{enable,disable}-decimal-float is not given.
|
|
bid Decimal support is requested and the encoding is
|
|
assumed to be BID (some check may be done).
|
|
dpd Decimal support is requested and the encoding is
|
|
assumed to be DPD (some check may be done).
|
|
generic Decimal support is requested and the generic code
|
|
is used (mainly for developers).
|
|
|
|
--enable-gmp-internals allows the MPFR build to use GMP's undocumented
|
|
functions (not from the public API). Note that
|
|
library versioning is not guaranteed to work if
|
|
this option is used. Thus it must not be used in
|
|
binary distributions.
|
|
|
|
--with-sysroot=DIR Search for dependent libraries within DIR (which
|
|
may be useful in cross-compilation). If you use
|
|
this option, you need to have Libtool 2.4+ on
|
|
the target system. See Libtool 2.4+'s NEWS file.
|
|
|
|
Note: By default, the configure script tries to set CC / CFLAGS to GMP's
|
|
ones from gmp.h (__GMP_CC / __GMP_CFLAGS) in order to ensure that MPFR is
|
|
built with the same ABI as GMP. The reason is that when GMP is built, it
|
|
may set CC / CFLAGS to select an ABI that is not the default one in order
|
|
to have a better performance. The -pedantic option in GMP's CFLAGS, when
|
|
present (which is the case by default), is removed, because the MPFR
|
|
build system uses some C extensions (when this script detects that they
|
|
are supported) and -pedantic yields too many useless warnings. However,
|
|
this setting from GMP is not guaranteed to work as the configure script
|
|
does some compiler tests earlier, and a conflict may arise. Also, the
|
|
values obtained from GMP may be incorrect for the MPFR build if GMP has
|
|
been built on a different machine; in such a case, the user may need to
|
|
specify CC / CFLAGS, as explained below.
|
|
|
|
Moreover, even without --with-gmp-build and --enable-gmp-internals,
|
|
MPFR might use some GMP internals by mistake. This would be a bug,
|
|
which should be reported to the MPFR developers.
|
|
|
|
Run "./configure --help" to see the other options (default options
|
|
from Autoconf and Automake).
|
|
|
|
|
|
If 'gmp.h' and 'libgmp' do not match
|
|
====================================
|
|
|
|
Under some circumstances, the configure script may output a message
|
|
saying:
|
|
|
|
'gmp.h' and 'libgmp' seem to have different versions or
|
|
we cannot run a program linked with GMP (if you cannot
|
|
see the version numbers above). [...]
|
|
|
|
Even though the configure script does not fail in such a case, this
|
|
message most often indicates a real error, which needs to be avoided.
|
|
Possible causes are:
|
|
|
|
* The wanted GMP library does not have the same ABI as the one chosen
|
|
to build MPFR. The consequences may be:
|
|
|
|
- A different GMP library (with the correct ABI) has been found,
|
|
but does not have the same version as 'gmp.h'.
|
|
|
|
- No other GMP libraries have been found (in this case, no version
|
|
numbers should be printed above the warning message).
|
|
|
|
This is incorrect and one of the following steps (make, make check)
|
|
will probably fail. GMP (actually gmp.h) now provides CC and CFLAGS
|
|
information to select the correct ABI, so that this problem should
|
|
no longer occur; but if GMP was provided by a binary package, such
|
|
information may not be valid. See the
|
|
|
|
checking for CC and CFLAGS in gmp.h...
|
|
|
|
line in the configure output (about the 11th line) and the following
|
|
few ones for more information. You may need to reinstall GMP or to
|
|
provide your own CC and/or CFLAGS. See also the remaining of this
|
|
INSTALL file.
|
|
|
|
* A configure option like --with-gmp or --with-gmp-include was used
|
|
with a system include directory, e.g. one of the following:
|
|
|
|
--with-gmp=/usr
|
|
--with-gmp=/usr/local
|
|
--with-gmp-include=/usr/include
|
|
--with-gmp-include=/usr/local/include
|
|
|
|
GCC (and possibly other compilers) will ignore such a directive for
|
|
include directories (but this rule is not applied for the library
|
|
itself!). This means that the library search paths will be reordered
|
|
as declared, but the specified include directory will still be near
|
|
the end of the include search paths (thus with a low precedence).
|
|
This is not a problem if only one GMP version is installed, but
|
|
otherwise, a wrong gmp.h may be chosen, so that the versions of
|
|
gmp.h and libgmp may not match. The suggestions are the following:
|
|
|
|
- If you want to use the GMP version under /usr, then you should
|
|
uninstall all the other GMP versions (header and library files)
|
|
that may be seen in the search paths, in particular those under
|
|
/usr/local.
|
|
|
|
- If you want to use the GMP version under /usr/local, then you
|
|
should uninstall all the other GMP versions (header and library
|
|
files) that may be seen in the search paths, but *NOT* the one
|
|
under /usr (the version under /usr is provided by the OS vendor,
|
|
and changing/removing anything related to it may break your
|
|
system, and /usr should have a lower precedence than /usr/local
|
|
anyway).
|
|
|
|
To find where GMP versions have been installed:
|
|
|
|
$ locate libgmp (if you have a locate database)
|
|
|
|
and if the compiler is GCC:
|
|
|
|
$ gcc -print-file-name=libgmp.so (under most systems)
|
|
$ gcc -print-file-name=libgmp.dylib (under Mac OS X)
|
|
|
|
and if this does not work, you may try:
|
|
|
|
$ gcc -print-search-dirs
|
|
|
|
* An official GCC version was used under Debian GNU/Linux. Problems
|
|
may come from the fact that Debian chose a different convention
|
|
for library search paths concerning 32-bit and 64-bit libraries.
|
|
A possible problem can be, for instance:
|
|
|
|
[Debian's GCC, correct library path]
|
|
$ gcc -print-file-name=libgmp.so
|
|
/home/vlefevre/gmp/athlon64/lib/../lib/libgmp.so
|
|
|
|
[Official GCC, incorrect library path]
|
|
$ gcc-4.3.1 -print-file-name=libgmp.so
|
|
/usr/lib/../lib64/libgmp.so
|
|
|
|
The solution: use a GCC provided by Debian or add symbolic links
|
|
such as lib64 -> lib (on 64-bit machines) for your library paths.
|
|
|
|
* The problem may also be temporary and only due to the fact that
|
|
libtool could not be used at this time. This is unlikely, though.
|
|
|
|
|
|
In case of problem
|
|
==================
|
|
|
|
First, look for any warning message in the configure output.
|
|
|
|
Several documents may help you to solve the problem:
|
|
* this INSTALL file, in particular information given below;
|
|
* the FAQ (either the FAQ.html file distributed with MPFR, or the
|
|
on-line version <https://www.mpfr.org/faq.html>, which may be more
|
|
up-to-date);
|
|
* the MPFR web page for this version <https://www.mpfr.org/mpfr-4.1.0/>,
|
|
which lists bugs found in this version and provides some patches.
|
|
|
|
If the "configure" fails, please check that the C compiler and its
|
|
options are the same as those for the GMP build (specially the ABI).
|
|
You can see the latter with the following command:
|
|
|
|
grep "^CC\|^CFLAGS" GMPBUILD/Makefile
|
|
|
|
if the GMP build directory is available. Then type:
|
|
|
|
./configure <configure options> CC=<C compiler> CFLAGS="<compiler options>"
|
|
|
|
(quotes are needed when there are spaces or other special characters
|
|
in the CC/CFLAGS value) and continue the install. On some platforms,
|
|
you should provide further options to match those used by GMP, or set
|
|
some environment variables. For instance, see the "Notes on AIX/PowerPC"
|
|
section below.
|
|
|
|
Warning! Do NOT use optimization options that can change the semantics
|
|
of math operations, such as GCC's -ffast-math or Sun CC's -fast.
|
|
Otherwise conversions from/to double's may be incorrect on infinities,
|
|
NaN's and signed zeros. Since native FP arithmetic is used in a few
|
|
places only, such options would not make MPFR faster anyway.
|
|
|
|
If some "make check" tests fail, you can try the --disable-thread-safe
|
|
configure option (see the configure options above): it has been reported
|
|
that some platforms have buggy TLS support. Before trying this option,
|
|
you may want to check in the configure output whether MPFR was built
|
|
with TLS support; if yes, you will have a line:
|
|
checking for TLS support... yes
|
|
Alternatively "grep MPFR_USE_THREAD_SAFE config.log" will show that
|
|
MPFR_USE_THREAD_SAFE is defined to 1. If it is "no" (or the variable
|
|
is not defined), the --disable-thread-safe option would be useless.
|
|
|
|
Some tests failure may be due to other compiler bugs, in particular
|
|
in optimization code. You can try to build MPFR without compiler
|
|
optimizations by giving -O0 (letter O, digit 0) in CFLAGS. If the
|
|
MPFR tests no longer fail, this was probably due to a compiler bug,
|
|
though we cannot exclude a bug in MPFR. You may want to contact us
|
|
(see below), possibly after looking at:
|
|
https://members.loria.fr/PZimmermann/software/compilerbugs.html
|
|
|
|
On some platforms, try with "gmake" (GNU make) instead of "make".
|
|
Problems have been reported with the Tru64 make.
|
|
|
|
If the configure script reports that gmp.h version and libgmp version
|
|
are different, or if the build was OK, but the tests failed to link
|
|
with GMP or gave an error like
|
|
|
|
undefined reference to `__gmp_get_memory_functions'
|
|
|
|
meaning that the GMP library was not found or a wrong GMP library was
|
|
selected by the linker, then your library search paths are probably
|
|
not correctly set (some paths are missing or they are specified in an
|
|
incorrect order).
|
|
|
|
Such problems commonly occur under some GNU/Linux machines, where the
|
|
default header and library search paths may be inconsistent: GCC is
|
|
configured to search /usr/local/include and /usr/local/lib by default,
|
|
while the dynamic linker ignores /usr/local/lib. If you have a GMP
|
|
version installed in /usr (provided by the OS vendor) and a new one
|
|
installed in /usr/local, then the header of the new GMP version and
|
|
the library of the old GMP version will be used! The best solution
|
|
is to make sure that the dynamic linker configuration is consistent
|
|
with GCC's behavior, for instance by having /usr/local/lib in
|
|
/etc/ld.so.conf or in some file from /etc/ld.so.conf.d (as Debian
|
|
did: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=395177). See
|
|
also https://gcc.gnu.org/legacy-ml/gcc-help/2010-01/msg00171.html for
|
|
more information. Alternatively you can use:
|
|
* environment variables. This may sometimes be necessary. If DIR
|
|
is the installation directory of GMP, add DIR/include to your
|
|
CPATH or C_INCLUDE_PATH (for compilers other than GCC, please
|
|
check the manual of your compiler), and add DIR/lib to your
|
|
LIBRARY_PATH and LD_LIBRARY_PATH (and/or LD_RUN_PATH);
|
|
* --with-gmp* configure options (described above), e.g.
|
|
--with-gmp=/opt/local (to use /opt/local/include for headers and
|
|
/opt/local/lib for libraries), but other software that uses GMP
|
|
and/or MPFR will need correct paths too, and environment variables
|
|
allow one to set such search paths in a global way.
|
|
Note about "--with-gmp=/usr/local". This option may appear to
|
|
solve the above inconsistency problem, but does not work as you
|
|
expect. Indeed it affects the library search path, in particular,
|
|
the one used by the dynamic linker (thus adding the missing
|
|
/usr/local/lib directory as wanted), but since /usr/local/include
|
|
is a "standard system include directory" for GCC, the include
|
|
search patch is not changed; this is often not a problem in this
|
|
particular case because usually, /usr/local/include is already
|
|
last in the include search patch, but this may fail under some
|
|
occasions and may trigger obscure errors.
|
|
|
|
For instance, under Unix, where paths are separated by a colon:
|
|
|
|
* With POSIX sh-compatible shells (e.g. sh, ksh, bash, zsh):
|
|
export C_INCLUDE_PATH="/usr/local/include:/other/path/include"
|
|
export LIBRARY_PATH="/usr/local/lib:/other/path/lib"
|
|
export LD_LIBRARY_PATH="$LIBRARY_PATH"
|
|
|
|
* With csh or tcsh:
|
|
setenv C_INCLUDE_PATH "/usr/local/include:/other/path/include"
|
|
setenv LIBRARY_PATH "/usr/local/lib:/other/path/lib"
|
|
setenv LD_LIBRARY_PATH "$LIBRARY_PATH"
|
|
|
|
If almost all the tests fail and the messages in the test-suite.log file
|
|
(or in the output, when running individual tests from the command line)
|
|
start with a line of the form:
|
|
|
|
Incorrect MPFR version! (xxx header vs yyy library)
|
|
|
|
then this means that an installed MPFR version is tested instead of the
|
|
one that has just been built. This is probably not a bug in MPFR, but a
|
|
problem caused by the user or system configuration (particular options,
|
|
environment variables, etc.) or a bug in the toolchain. In particular,
|
|
if LD_LIBRARY_PATH overrides the run path (set up by libtool) and an
|
|
installed ABI-compatible version of MPFR is in a directory listed in
|
|
the LD_LIBRARY_PATH search path, then this will break. An example with
|
|
GNU ld:
|
|
|
|
https://sourceware.org/bugzilla/show_bug.cgi?id=21476
|
|
|
|
If you can't solve your problem, you should contact us via the MPFR
|
|
mailing-list <mpfr@inria.fr>, indicating the machine and operating system
|
|
used (uname -a), the compiler and version used (gcc -v if you use gcc),
|
|
the configure options used if any (including variables such as CC and
|
|
CFLAGS), the version of GMP and MPFR used, and a description of the
|
|
problem encountered, in particular the tests/test-suite.log file if
|
|
"make check" failed. Please send us also the log of the "configure"
|
|
(config.log).
|
|
|
|
|
|
Notes about ABI
|
|
===============
|
|
|
|
On 64-bit computers, it may happen that the default ABI (Application Binary
|
|
Interface) chosen by MPFR does not correspond to the default one chosen by
|
|
the compiler.
|
|
|
|
In particular, this kind of message may indicate the problem:
|
|
/usr/bin/ld: skipping incompatible mpfr/src/.libs/libmpfr.a when searching for -lmpfr
|
|
|
|
In fact, since MPFR relies on GMP, it uses the same ABI as GMP.
|
|
To be sure that your program compiles and links correctly, use the same
|
|
compiler flags as MPFR does (look for CFLAGS in config.log).
|
|
|
|
You might also recompile GMP with a different ABI, with for example
|
|
./configure ABI=32.
|
|
|
|
|
|
Notes on Mac OS X
|
|
=================
|
|
|
|
If you get an error of the form
|
|
|
|
ld: pointer in read-only segment not allowed in slidable image...
|
|
|
|
this can mean that the link is done against a static (GMP) library.
|
|
In such a case, you should configure MPFR with --disable-shared to
|
|
disable the build of the shared library.
|
|
|
|
|
|
Notes on FreeBSD 4.3
|
|
====================
|
|
|
|
FreeBSD 4.3 is provided with an incorrect <float.h> header file, and
|
|
MPFR tests related to long double's may fail. If you cannot upgrade
|
|
the system, you can still use MPFR with FreeBSD 4.3, but you should
|
|
not use conversions with the long double type.
|
|
|
|
|
|
Notes on AIX/PowerPC
|
|
====================
|
|
|
|
The following has been tested on AIX 7.1.3 (gcc111.fsffrance.org)
|
|
with gcc 4.8.1 and GMP 6.1.0.
|
|
|
|
If GMP was built with the 64-bit ABI, before building and testing MPFR,
|
|
you should set the OBJECT_MODE environment variable to 64, e.g., with:
|
|
|
|
export OBJECT_MODE=64
|
|
|
|
(in a sh-compatible shell). Alternatively add the following to the configure
|
|
line: AR="ar -X64" NM="nm -B -X64".
|
|
|
|
|
|
Notes on Solaris
|
|
================
|
|
|
|
Do not put a -R<dir> option in the LD_OPTIONS environment variable, at least
|
|
if the directory <dir> can contain an MPFR library. Otherwise this MPFR
|
|
library may be chosen for the tests (make check) instead of the one that has
|
|
just been built, in which case, either you will get errors due to unmatched
|
|
versions or this problem may remain undetected. The reason is that this
|
|
option will appear before the -R options added by libtool, such as the one
|
|
to the src/.libs directory containing the MPFR library that has just been
|
|
built, and will have the precedence.
|
|
|
|
|
|
MPFR for use with Windows Applications
|
|
======================================
|
|
|
|
There are several ways of building MPFR on Windows, the most appropriate
|
|
approach depending on how you intend to use the resulting libraries.
|
|
|
|
a. Using MinGW
|
|
==============
|
|
|
|
1 - We advise to use MinGW (http://www.mingw.org/ for 32-bit, and
|
|
https://sourceforge.net/projects/mingw-w64/ for 32- and 64-bit),
|
|
which is simpler and less demanding than Cygwin. Contrary to Cygwin,
|
|
it also provides native Windows code.
|
|
|
|
If you also use MSYS, you should use "make" for MSYS instead of
|
|
the "make" utility from MinGW-W64 or from GCC, which causes the
|
|
following error:
|
|
|
|
libtool: warning: libobj name 'extract.Tpo -c -o extract.lo extract.lo'
|
|
may not contain shell special characters.
|
|
rm: unknown option -- c
|
|
|
|
References about this issue and solution:
|
|
https://sourceforge.net/p/msys2/tickets/223/
|
|
https://sympa.inria.fr/sympa/arc/mpfr/2016-07/msg00009.html
|
|
|
|
2 - If you just want to make a binary with gcc, there is nothing to do:
|
|
GMP, MPFR and the program compile exactly as under Linux. (It is
|
|
recommended to pass --build=xxx-yyy-mingw64 to the GMP configure command,
|
|
or --build=xxx with xxx containing mingw.)
|
|
|
|
3 - To avoid using the Microsoft runtime (which might not be conform to ISO C),
|
|
you can use the MinGW runtime package (which is an integral part of MinGW).
|
|
|
|
With old MinGW versions, you can get an ISO-compliant printf()
|
|
if you compile your application with either '-ansi', '-posix' or
|
|
'-D__USE_MINGW_ANSI_STDIO'. But note that this latter option,
|
|
which was useful in the past (see below) should no longer be used.
|
|
|
|
The following applies to old MinGW versions, and may be discouraged
|
|
with recent MinGW versions.
|
|
|
|
In order to have the MPFR formatted output functions based on an
|
|
ISO-compliant printf(), you need to compile GMP (not MPFR) with
|
|
CC="gcc -D__USE_MINGW_ANSI_STDIO" (since the standard printf modifiers
|
|
%e, %Ld and %td are passed to GMP).
|
|
|
|
To avoid failures of some of the printf-related tests, MPFR needs to
|
|
be compiled with the same __USE_MINGW_ANSI_STDIO as with GMP, i.e.
|
|
this macro should be defined for both or undefined for both; this
|
|
should be the case by default, unless CC or CFLAGS has been redefined.
|
|
For instance, if __USE_MINGW_ANSI_STDIO is defined in GMP but not in
|
|
MPFR, the following error may occur:
|
|
|
|
Error in mpfr_vsprintf (s, "%e", ...);
|
|
expected: "-1.250000e+000"
|
|
got: "-1.250000e+00"
|
|
FAIL tsprintf.exe (exit status: 1)
|
|
|
|
and in the opposite case:
|
|
|
|
Error in mpfr_vsprintf (s, "%e", ...);
|
|
expected: "-1.250000e+00"
|
|
got: "-1.250000e+000"
|
|
FAIL tsprintf.exe (exit status: 1)
|
|
|
|
Note with old GMP versions: Other issues could arise due to the fact
|
|
that the C functions vsnprintf and vsprintf both used internally in
|
|
old GMP versions do not produce the same output:
|
|
|
|
https://sympa.inria.fr/sympa/arc/mpfr/2016-03/msg00045.html
|
|
https://sympa.inria.fr/sympa/arc/mpfr/2016-03/msg00051.html
|
|
https://sympa.inria.fr/sympa/arc/mpfr/2016-03/msg00053.html
|
|
|
|
If __USE_MINGW_ANSI_STDIO has not been defined when building GMP,
|
|
then the length modifiers j, L and t are not supported with the GMP
|
|
formatted output functions, and as a consequence, also with MPFR.
|
|
This is automatically detected by the configure script, except when
|
|
cross-compiling (e.g. under Linux), in which case some macros need
|
|
to be defined explicitly, e.g. with
|
|
|
|
"CPPFLAGS=-DNPRINTF_J -DNPRINTF_L -DNPRINTF_T"
|
|
|
|
in order to avoid potential issues with the MPFR library and failures
|
|
in the test suite (the corresponding tests are disabled explicitly).
|
|
|
|
b. Using Cygwin
|
|
===============
|
|
|
|
This build should be similar to that for MinGW except that the resulting
|
|
library depends on the Cygwin DLL and cannot therefore be used with
|
|
Visual Studio as with MinGW. Indeed, the binaries compiled with Cygwin
|
|
require a dynamic library (cygwin.dll) to work; there is a Cygwin option
|
|
-mno-cygwin to build native code, but it may require some non-portable tricks.
|
|
|
|
In case of failure, you may need to pass LDFLAGS='-shared-libgcc' at the
|
|
end of the configure line due to a bug in GCC. Otherwise, if threading
|
|
support is not needed, you can configure MPFR with --disable-thread-safe.
|
|
|
|
c. Using Microsoft Visual C++ and Intel C++ Compilers
|
|
=====================================================
|
|
|
|
Static and dynamic MPFR libraries for the 32- and 64-bit versions of
|
|
Windows can be built with Microsoft Visual Studio 2015 using the
|
|
Microsoft Visual C++ compiler, see:
|
|
|
|
https://visualstudio.microsoft.com/
|
|
|
|
The Intel C++ compiler provided as a part of Intel Parallel Studio XE
|
|
can also be used:
|
|
|
|
https://software.intel.com/en-us/parallel-studio-xe
|
|
|
|
The relevant build projects are available here:
|
|
|
|
https://github.com/BrianGladman
|
|
|
|
d. Using the CompCert compiler
|
|
==============================
|
|
|
|
[this was tested with CompCert 3.7 and MPFR revision 13851 on x86_64-linux]
|
|
|
|
CompCert (http://compcert.inria.fr/) is a formally verified compiler.
|
|
To compile MPFR with CompCert:
|
|
|
|
$ ./configure CC=ccomp CFLAGS="-fbitfields -flongdouble -fstruct-passing" --disable-shared
|
|
|
|
You also need to unset LD_LIBRARY_PATH, and/or you might need to change
|
|
wl="" into wl="-Wl," in the libtool file.
|
|
|
|
All tests (make check) should pass (tget_set_d64, tget_set_d128 and
|
|
tset_float128 are skipped, since CompCert does not support _Decimal64,
|
|
_Decimal128 nor _Float128).
|