summaryrefslogtreecommitdiff
path: root/Build/source/libs/mpfr/mpfr-src/INSTALL
diff options
context:
space:
mode:
authorKarl Berry <karl@freefriends.org>2021-02-25 19:22:25 +0000
committerKarl Berry <karl@freefriends.org>2021-02-25 19:22:25 +0000
commitad547a6b5986815fda458221149728d9d9ab1d87 (patch)
tree16296910eb3eca724371474ea9aea3994dc69614 /Build/source/libs/mpfr/mpfr-src/INSTALL
parent947b43de3dd21d58ccc2ffadefc4441ea1c2a813 (diff)
restore Build,TODO from r57911
git-svn-id: svn://tug.org/texlive/trunk@57915 c570f23f-e606-0410-a88d-b1316a301751
Diffstat (limited to 'Build/source/libs/mpfr/mpfr-src/INSTALL')
-rw-r--r--Build/source/libs/mpfr/mpfr-src/INSTALL719
1 files changed, 719 insertions, 0 deletions
diff --git a/Build/source/libs/mpfr/mpfr-src/INSTALL b/Build/source/libs/mpfr/mpfr-src/INSTALL
new file mode 100644
index 00000000000..f1b6a5cccec
--- /dev/null
+++ b/Build/source/libs/mpfr/mpfr-src/INSTALL
@@ -0,0 +1,719 @@
+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).