Mercurial > hg > xemacs-beta
changeset 614:42bfaea76395
[xemacs-hg @ 2001-06-11 01:55:26 by martinb]
DECIMAL_PRINT_SIZE sun cc bug workaround
| author | martinb |
|---|---|
| date | Mon, 11 Jun 2001 01:55:28 +0000 |
| parents | 023b83f4e54b |
| children | 4f50f8a33f96 |
| files | src/ChangeLog src/lisp.h |
| diffstat | 2 files changed, 12 insertions(+), 2 deletions(-) [+] |
line wrap: on
line diff
--- a/src/ChangeLog Sun Jun 10 10:42:39 2001 +0000 +++ b/src/ChangeLog Mon Jun 11 01:55:28 2001 +0000 @@ -1,3 +1,9 @@ +2001-06-11 Martin Buchholz <martin@xemacs.org> + + * lisp.h (DECIMAL_PRINT_SIZE): Don't use floating point. + Sun cc (buggily?) cannot use floating point computations to define + a compile-time integral constant. + 2001-06-10 Ben Wing <ben@xemacs.org>
--- a/src/lisp.h Sun Jun 10 10:42:39 2001 +0000 +++ b/src/lisp.h Mon Jun 11 01:55:28 2001 +0000 @@ -2856,9 +2856,13 @@ /* The number of bytes required to store the decimal printed representation of an integral type. Add a few bytes for truncation, optional sign prefix, and null byte terminator. - 2.40824 == log (256) / log (10). */ + 2.40824 == log (256) / log (10). + + We don't use floating point since Sun cc (buggily?) cannot use + floating point computations to define a compile-time integral + constant. */ #define DECIMAL_PRINT_SIZE(integral_type) \ -((size_t) (2.40824 * sizeof (integral_type)) + 3) +(((2410824 * sizeof (integral_type)) / 1000000) + 3) void long_to_string (char *, long); void print_internal (Lisp_Object, Lisp_Object, int);