Hi all, I had a look through the svn log to try and construct a list of things which have changed since the last release. Unfortunately, all the log entries for mpir-exp have been lost in the merge to trunk, but I looked through the mpir-exp log so that they will at least be reflected in the Changelog.
Here is what I have found so far: Bugs: * fixed build issue due to incorrect redc_2 header include * proper test for unpatched gcc-4.3.2 * bug in speed/common for dc_bdiv_qr_n * documentation error wrt -ve exponents in mpz_powm_ui * correct mpq_cmp_ui declaration error found by David Cleaver * correct bug in Windows assembler code for karasub Improvements: * Completely new FFT implementation (William Hart) * Support for choice of intmax_t integer types in ui/si functions (Brian Gladman) * Visual Studio 11 support (Brian Gladman) * added cpuid support for more x86_64 CPUs (Jason Moxham) * new redc_1 for sandybridge and bobcat (Jason Moxham) * Python Windows build generator (Brian Gladman) Compatibility: * additional macro support for mpfr-3.1.0 * Add macros for __GNU_MP_RELEASE and __MPIR_RELEASE Is there anything missing? Any corrections? Anything that shouldn't be there? When all is ok, I'll do the paperwork for the new release. I also think we should do a release 2.5.2 based on revision 3897 which fixes the redc_2 build issue in 2.5.1 (it looks like Jason was intending to do a 2.5.2 release for this reason). This will give people a nice stable MPIR to use if they aren't comfortable with switching over to the new FFT immediately (though it has been used pretty extensively in flint for some time now). Bill. On 2 October 2012 09:36, Brian Gladman <b...@gladman.plus.com> wrote: > -----Original Message----- From: Case Van Horsen > Sent: Tuesday, October 02, 2012 7:43 AM > To: mpir-devel@googlegroups.com > Subject: Re: [mpir-devel] Re: MPIR 2.6 release progress > > [snip] > > > Interesting. If I use ABI=64, then "make" and "make check" succeed. > > If I use "configure.bat ABI=32 --cpu=pentium3", then "make" succeeds > but I get the following error for "make check": > > try.c > try.obj : error LNK2001: unresolved external symbol ___gmpn_mod_1c > try.obj : error LNK2001: unresolved external symbol ___gmpn_divrem_1c > try.exe : fatal error LNK1120: 2 unresolved externals > ERROR > > Any ideas where to look? > > ============================================= > > On Windows exported symbols for win32 are prefixed with three '___' whereas > the prefix for x64 symbols is '__'. > > If win32 and x64 versions of MPIR both exist, when try is built it will use > the headers for the last built version of the library. Hence if a link is > done for an x64 version of try but the last built version of MPIR is for > win32, an x64 build will be done but it will use win32 headers. I have > seen build errors of the type you are seeing where this is the cause. > > > Brian > > -- > You received this message because you are subscribed to the Google Groups > "mpir-devel" group. > To post to this group, send email to mpir-devel@googlegroups.com. > To unsubscribe from this group, send email to > mpir-devel+unsubscr...@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/mpir-devel?hl=en. > -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To post to this group, send email to mpir-devel@googlegroups.com. To unsubscribe from this group, send email to mpir-devel+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/mpir-devel?hl=en.
