diff options
-rw-r--r-- | INSTALL | 2 | ||||
-rw-r--r-- | NEWS | 1 | ||||
-rw-r--r-- | README.dev | 24 | ||||
-rw-r--r-- | configure.ac | 17 | ||||
-rw-r--r-- | doc/mpc.texi | 6 | ||||
-rw-r--r-- | tests/log.dat | 4 |
6 files changed, 26 insertions, 28 deletions
@@ -6,7 +6,7 @@ in the Texinfo documentation (type 'info mpc.info'). 0. You first need to install GMP and MPFR. See <http://gmplib.org/> and <http://www.mpfr.org>. - MPC requires GMP version 4.2 or later and MPFR version 2.3.1 or later. + MPC requires GMP version 4.3.2 or later and MPFR version 2.4.2 or later. 1. In the directory of the MPC archive, type @@ -8,6 +8,7 @@ Changes in version 0.8.3: - Bug fixes: - trigonometric functions: infinite loop due to overflow for large arguments - exp: close to infinite loop for argument close to 0 + - Minimally required library versions: gmp 4.3.2, mpfr 2.4.2 Changes in version 0.8.2: - Speed-up of mpc_pow_ui through binary exponentiation @@ -14,31 +14,29 @@ Required versions: Creating a new release ---------------------- - 0) Run all tests with valgrind and check no memory leak remains: + 1) Run all tests with valgrind and check no memory leak remains: for f in .libs/lt-t*; do valgrind $f; done - 0b) Disable test in line 35 of log.dat. - 0c) Check compilation with gmp 4.2.2 and mpfr 2.3.1. - 1) Check the version number in configure.ac ("AC_INIT (mpc, _version_...)"), + 2) Check compilation with gmp 4.3.2 and mpfr 2.4.2. + 3) Check the version number in configure.ac ("AC_INIT (mpc, _version_...)"), INSTALL, src/get_version.c, src/mpc.h (remove suffix "-dev" in MPC_VERSION_STRING) and Makefile.vc. - 2) Verify that the API version in src/Makefile.am is correct. - 3) Check that NEWS, BUGS and TODO are up to date, and that the minimal + 4) Verify that the API version in src/Makefile.am is correct. + 5) Check that NEWS, BUGS and TODO are up to date, and that the minimal version numbers of gmp and mpfr in configure.ac, INSTALL and doc/mpc.texi are correct. - 4) Execute "autoreconf; touch doc/mpc.texi; make distcheck". + 6) Execute "autoreconf; touch doc/mpc.texi; make distcheck". This updates the date and version number in doc/version.texi and creates the tarball for distribution. Warning: do this on a machine with a recent libtool (for example 2.2.x instead of 1.5.x which is buggy, cf http://lists.gforge.inria.fr/pipermail/mpc-discuss/2009-March/000106.html) - 5) Create an svn tag from inside the svn root: + 7) Create an svn tag from inside the svn root: svn cp trunk tags/VERSION - 6) Update the web page on the server. + 8) Update the web page on the server. The html documentation is created with makeinfo --html --no-split mpc.texi and the resulting mpc.html is copied to content_html.php. - 7) After the release, update the version number in configure.ac, INSTALL, + 9) After the release, update the version number in configure.ac, INSTALL, src/get_version.c, src/mpc.h and Makefile.vc. - 9) Potentially increase the API version in src/Makefile.am. - 9b) Reenable test in line 35 of log.dat. -10) Commit the changes to svn. +10) Potentially increase the API version in src/Makefile.am. +11) Commit the changes to svn. diff --git a/configure.ac b/configure.ac index 66f66f3..33c1869 100644 --- a/configure.ac +++ b/configure.ac @@ -132,34 +132,33 @@ AC_COMPILE_IFELSE( [AC_LANG_SOURCE( [[ #include "gmp.h" -#if (__GNU_MP_VERSION*100 + __GNU_MP_VERSION_MINOR*10 + __GNU_MP_VERSION_PATCHLEVEL < 420) -# error "Minimal GMP version is 4.2" +#if (__GNU_MP_VERSION*100 + __GNU_MP_VERSION_MINOR*10 + __GNU_MP_VERSION_PATCHLEVEL < 432) +# error "Minimal GMP version is 4.3.2" error #endif ]])], [AC_MSG_RESULT(yes)], [ AC_MSG_RESULT(no) - AC_MSG_ERROR([GMP version >= 4.2 required]) + AC_MSG_ERROR([GMP version >= 4.3.2 required]) ]) -# Check for a recent MPFR: we require MPFR 2.3.1 because of a bug in -# mpfr_div_ui which makes mpc_div_ui return the wrong sign of the zero part -# when applied on a pure real or pure imaginary parameter (see mpfr r4950) +# Check for a recent MPFR: we require MPFR 2.4.2 so that the tests +# in log.dat pass AC_MSG_CHECKING(for recent MPFR) AC_COMPILE_IFELSE( [AC_LANG_SOURCE( [[ #include "mpfr.h" -#if (MPFR_VERSION < MPFR_VERSION_NUM (2,3,1)) -# error "Minimal MPFR version is 2.3.1" +#if (MPFR_VERSION < MPFR_VERSION_NUM (2,4,2)) +# error "Minimal MPFR version is 2.4.2" error #endif ]])], [AC_MSG_RESULT(yes)], [ AC_MSG_RESULT(no) - AC_MSG_ERROR([MPFR version >= 2.3.1 required]) + AC_MSG_ERROR([MPFR version >= 2.4.2 required]) ]) # Checks for header files. diff --git a/doc/mpc.texi b/doc/mpc.texi index 2213767..99b9ea7 100644 --- a/doc/mpc.texi +++ b/doc/mpc.texi @@ -33,6 +33,8 @@ was introduced. @end iftex @include version.texi +@set MINGMP 4.3.2 +@set MINMPFR 2.4.2 @set AUTHORS Andreas Enge, Philippe Th@'eveny, Paul Zimmermann @set COPYRIGHTDATES 2002, 2003, 2004, 2005, 2007, 2008, 2009, 2010 @@ -164,8 +166,8 @@ probably a good idea to skim through it. @chapter Installing MPC @cindex Installation -To build MPC, you first have to install GNU MP (version 4.2 or higher) and -GNU MPFR (version 2.3.1 or higher) on your computer. You need a C compiler, +To build MPC, you first have to install GNU MP (version @value{MINGMP} or higher) and +GNU MPFR (version @value{MINMPFR} or higher) on your computer. You need a C compiler, preferably GCC, but any reasonable compiler should work. And you need a standard Unix @samp{make} program, plus some other standard Unix utility programs. diff --git a/tests/log.dat b/tests/log.dat index e620cfe..9f63027 100644 --- a/tests/log.dat +++ b/tests/log.dat @@ -1,6 +1,6 @@ # Data test file for mpc_log. -# Copyright (C) 2008, 2009 Paul Zimmermann, Andreas Enge +# Copyright (C) 2008, 2009, 2010 Paul Zimmermann, Andreas Enge # This file is part of the MPC Library. # @@ -30,8 +30,6 @@ 0 - 2 -inf 53 +0x3243F6A8885A3p-48 2 -0 2 +0 N N 0 + 2 -inf 53 -0x3243F6A8885A3p-48 2 -0 2 -0 N N 0 - 2 -inf 53 +0x3243F6A8885A3p-48 2 -0 2 +0 N D -# The following test reveals a bug in mpfr-2.4.1 and should be disabled for -# the release as long as mpc does not require a more recent version of mpfr. 0 - 2 -inf 53 -0x3243F6A8885A32p-52 2 -0 2 -0 N D # log(+0 + i*0) = -inf + i*0 |