logb now works correctly on large integers
commit74f87fd111904e5156727c72590d6fc4f67e8366
authorPaul Eggert <eggert@cs.ucla.edu>
Fri, 3 Mar 2017 17:17:51 +0000 (3 09:17 -0800)
committerPaul Eggert <eggert@cs.ucla.edu>
Fri, 3 Mar 2017 17:19:08 +0000 (3 09:19 -0800)
treef6802878c5105def6d6889d5b8f71e4fe9285b79
parentf1fe3fcfc568c1527714ff3a95e678816e2787d4
logb now works correctly on large integers

* admin/merge-gnulib (GNULIB_MODULES): Add count-leading-zeros.
* etc/NEWS: Document the change.
* lib/count-leading-zeros.c, lib/count-leading-zeros.h:
* m4/count-leading-zeros.m4: New files, copied from Gnulib.
* lib/gnulib.mk, m4/gnulib-comp.m4: Regenerate.
* src/floatfns.c: Include count-leading-zeros.h.
(Flogb): Do not convert fixnum to float before taking the log,
as the rounding error can cause the answer to be off by 1.
* src/lisp.h (EMACS_UINT_WIDTH): New constant.
* test/src/floatfns-tests.el (logb-extreme-fixnum): New test.
admin/merge-gnulib
etc/NEWS
lib/count-leading-zeros.c [new file with mode: 0644]
lib/count-leading-zeros.h [new file with mode: 0644]
lib/gnulib.mk
m4/count-leading-zeros.m4 [new file with mode: 0644]
m4/gnulib-comp.m4
src/floatfns.c
src/lisp.h
test/src/floatfns-tests.el