diff options
author | zimmerma <zimmerma@280ebfd0-de03-0410-8827-d642c229c3f4> | 2007-10-15 13:14:30 +0000 |
---|---|---|
committer | zimmerma <zimmerma@280ebfd0-de03-0410-8827-d642c229c3f4> | 2007-10-15 13:14:30 +0000 |
commit | 7583bbab6d952e34100ed5208faf9394facffc8b (patch) | |
tree | 34e3653ad4cda4e3cd7f39a8c24cbb83f91473ea | |
parent | 0ddb50a877d313472be9efecab7879207f942ac9 (diff) | |
download | mpfr-7583bbab6d952e34100ed5208faf9394facffc8b.tar.gz |
added table of contents
removed last section (mpf/mpfr integration)
git-svn-id: svn://scm.gforge.inria.fr/svn/mpfr/trunk@4894 280ebfd0-de03-0410-8827-d642c229c3f4
-rw-r--r-- | TODO | 52 |
1 files changed, 15 insertions, 37 deletions
@@ -19,9 +19,17 @@ along with the MPFR Library; see the file COPYING.LIB. If not, write to the Free Software Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. +Table of contents: +1. Documentation +2. Installation +3. Changes in existing functions +4. New functions to implement +5. Efficiency +6. Miscellaneous +7. Portability ############################################################################## -Documentation: +1. Documentation ############################################################################## - add a description of the algorithms used + proof of correctness @@ -31,14 +39,14 @@ Documentation: ############################################################################## -Installation: +2. Installation ############################################################################## - nothing to do currently :-) ############################################################################## -Changes in existing functions: +3. Changes in existing functions ############################################################################## - many functions currently taking into account the precision of the *input* @@ -76,7 +84,7 @@ Changes in existing functions: ############################################################################## -New functions to implement: +4. New functions to implement ############################################################################## - mpfr_fmod (mpfr_t, mpfr_srcptr, mpfr_srcptr, mp_rnd_t) @@ -235,7 +243,7 @@ New functions to implement: ############################################################################## -Efficiency: +5. Efficiency ############################################################################## - fix regression with mpfr_mpz_root (from Keith Briggs, 5 July 2006), for @@ -286,7 +294,7 @@ Efficiency: ############################################################################## -Miscellaneous: +6. Miscellaneous ############################################################################## - [suggested by Tobias Burnus <burnus(at)net-b.de> and @@ -367,7 +375,7 @@ Miscellaneous: ############################################################################## -Portability: +7. Portability ############################################################################## - [Kevin about texp.c long strings] @@ -377,33 +385,3 @@ Portability: couple of places already. Arrays of chars are not much fun. - use http://gcc.gnu.org/viewcvs/trunk/config/stdint.m4 for mpfr-gmp.h - - -############################################################################## -Possible future MPF / MPFR integration: -############################################################################## - -- mpf routines can become "extern inline"s calling mpfr equivalents, - probably just with GMP_RNDZ hard coded, since that's what mpf has - always done. - -- Want to preserve the mpf_t structure size, for binary compatibility. - Breaking compatibility would cause lots of pain and potential subtle - breakage for users. If the fields in mpf_t are not enough then - extra space under _mp_d can be used. - -- mpf_sgn has been a macro directly accessing the _mp_size field, so a - compatible representation would be required. At worst that field - could be maintained for mpf_sgn, but not otherwise used internally. - - mpf_sgn should probably throw an exception if called with NaN, since - there's no useful value it can return, so it might want to become a - function. Inlined copies in existing binaries would hopefully never - see a NaN, if they only do old-style mpf things. - -- mpfr routines replacing mpf routines must be reentrant and thread - safe, since of course that's what has been documented for mpf. - -- mpfr_random will not be wanted since there's no corresponding - mpf_random and new routines should not use the old style global - random state. |