APPROVE COMMIT I'm committing this without waiting under the "obviously correct" doctrine. We were creating bignums and bigfloats on the stack, then not releasing memory properly before they went out of scope. (We were calling mpz_init() without a corresponding mpz_clear(), and likewise for mpf_init() and mpf_clear(), for those familiar with the GMP interface.) diff -r 2f22818d92d4 src/ChangeLog --- a/src/ChangeLog Sat Sep 06 14:30:15 2014 +0100 +++ b/src/ChangeLog Tue Sep 23 16:44:06 2014 -0600 ＠＠ -1,3 +1,8 ＠＠ +2014-09-23 Jerry James <james(a)xemacs.org&gt; + + * floatfns.c (round_two_bignum_1): Fix memory leak. + (round_two_bigfloat): Ditto. + 2014-09-06 Aidan Kehoe <kehoea(a)parhasard.net&gt; * select-x.c (Fx_store_cutbuffer_internal): diff -r 2f22818d92d4 src/floatfns.c --- a/src/floatfns.c Sat Sep 06 14:30:15 2014 +0100 +++ b/src/floatfns.c Tue Sep 23 16:44:06 2014 -0600 ＠＠ -1681,6 +1681,10 ＠＠ *res = make_bignum_bg (floored); *remain = make_bignum_bg (scratch_bignum); } + bignum_fini (flsecond); + bignum_fini (floored); + bignum_fini (flooring); + bignum_fini (hi2); } static Lisp_Object ＠＠ -1835,6 +1839,7 ＠＠ bigfloat_div (divided, XBIGFLOAT_DATA (number), XBIGFLOAT_DATA (divisor)); res0 = round_one_bigfloat_1 (divided); + bigfloat_fini (divided); bigfloat_set_prec (scratch_bigfloat, prec); bigfloat_set_prec (scratch_bigfloat2, prec); -- Jerry James http://www.jamezone.org/ _______________________________________________ XEmacs-Patches mailing list XEmacs-Patches(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-patches