Fix --remove-files in update/append mode.

[chaz/tar] / src / arith.h

diff --git a/src/arith.h b/src/arith.h
index e8c2ace395734d76b0a710529d806105b674f385..a2abfe826c1e1f37b3a4c6b310753f004a7fa004 100644 (file)
--- a/src/arith.h
+++ b/src/arith.h
@@ -1,9 +1,9 @@
-/* Simple arithmetic for numbers greater than a long int, for GNU tar.
-   Copyright (C) 1996, 1997 Free Software Foundation, Inc.
+/* Long integers, for GNU tar.
+   Copyright 1999, 2007 Free Software Foundation, Inc.

 
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
 
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by

-   the Free Software Foundation; either version 2, or (at your option)
+   the Free Software Foundation; either version 3, or (at your option)

    any later version.
 
    This program is distributed in the hope that it will be useful,
    any later version.
 
    This program is distributed in the hope that it will be useful,


@@ -13,111 +13,15 @@
 
    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software Foundation,
 
    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software Foundation,

-   Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
-
-/* Also, see comments at beginning of arith.c.  */
-
-#define BITS_PER_BYTE 8                /* number of bits in each sizeof unit */
-#define BITS_PER_TARLONG 42    /* wanted number of bits in each tarlong */
-
-/* In all cases, tarlong is the proper type for a big number.
-
-   For simulated arithmetic, SUPERDIGIT is the base, TARLONG_FORMAT is the
-   format to print a single super-digit filled with zeroes to the left, and
-   BITS_PER_SUPERDIGIT is the smallest number of bits required to fully
-   represent each super-digit.  LONGS_PER_TARLONG says how many longs are
-   required for a full tarlong, and SIZEOF_TARLONG is the size of a tarlong
-   in bytes.
-
-   For straight compiler arithmetic, SUPERDIGIT is zero and TARLONG_FORMAT
-   is the format to directly print a tarlong (without zero-filling).
-
-   The values of SIZEOF_LONG_LONG and SIZEOF_UNSIGNED_LONG, below, are
-   obtained through the configuration process.  */
-
-#if BITS_PER_BYTE * SIZEOF_UNSIGNED_LONG >= BITS_PER_TARLONG
-# define SUPERDIGIT 0
-# define TARLONG_FORMAT "%uld"
-typedef unsigned long tarlong;
-#else
-# if BITS_PER_BYTE * SIZEOF_LONG_LONG >= BITS_PER_TARLONG + 1
-#  define SUPERDIGIT 0
-#  define TARLONG_FORMAT "%lld"
-typedef long long tarlong;
-# else
-#  if BITS_PER_BYTE * SIZEOF_UNSIGNED_LONG >= 64
-#   define SUPERDIGIT 1000000000L
-#   define BITS_PER_SUPERDIGIT 29
-#   define TARLONG_FORMAT "%09uld"
-#  else
-#   if BITS_PER_BYTE * SIZEOF_UNSIGNED_LONG >= 32
-#    define SUPERDIGIT 10000L
-#    define BITS_PER_SUPERDIGIT 14
-#    define TARLONG_FORMAT "%04uld"
-#   endif
-#  endif
-# endif
-#endif
-
-#if SUPERDIGIT
-
-# define LONGS_PER_TARLONG \
-    ((BITS_PER_TARLONG + BITS_PER_SUPERDIGIT - 1) / BITS_PER_SUPERDIGIT)
-# define SIZEOF_TARLONG (LONGS_PER_TARLONG * sizeof (unsigned long))
-
-/* The NEC EWS 4.2 C compiler gets confused by a pointer to a typedef that
-   is an array.  So we wrap the array into a struct.  (Pouah!)  */
-
-struct tarlong
-{
-  unsigned long digit[LONGS_PER_TARLONG];
-};
-
-typedef struct tarlong tarlong;
-
-int zerop_tarlong_helper PARAMS ((unsigned long *));
-int lessp_tarlong_helper PARAMS ((unsigned long *, unsigned long *));
-void clear_tarlong_helper PARAMS ((unsigned long *));
-void add_to_tarlong_helper PARAMS ((unsigned long *, int));
-void mult_tarlong_helper PARAMS ((unsigned long *, int));
-void print_tarlong_helper PARAMS ((unsigned long *, FILE *));
-
-# define zerop_tarlong(Accumulator) \
-   zerop_tarlong_helper (&(Accumulator).digit[0])
-
-# define lessp_tarlong(First, Second) \
-   lessp_tarlong_helper (&(First).digit[0], &(Second).digit[0])
-
-# define clear_tarlong(Accumulator) \
-   clear_tarlong_helper (&(Accumulator).digit[0])
-
-# define add_to_tarlong(Accumulator, Value) \
-   add_to_tarlong_helper (&(Accumulator).digit[0], (Value))
-
-# define mult_tarlong(Accumulator, Value) \
-   mult_tarlong_helper (&(Accumulator).digit[0], (Value))
-
-# define print_tarlong(Accumulator, File) \
-   print_tarlong_helper (&(Accumulator).digit[0], (File))
-
-#else /* not SUPERDIGIT */
-
-# define zerop_tarlong(Accumulator) \
-   ((Accumulator) == 0)
-
-# define lessp_tarlong(First, Second) \
-   ((First) < (Second))
-
-# define clear_tarlong(Accumulator) \
-   ((Accumulator) = 0)
-
-# define add_to_tarlong(Accumulator, Value) \
-   ((Accumulator) += (Value))
-
-# define mult_tarlong(Accumulator, Value) \
-   ((Accumulator) *= (Value))
-
-# define print_tarlong(Accumulator, File) \
-   (fprintf ((File), TARLONG_FORMAT, (Accumulator)))
-
-#endif /* not SUPERDIGIT */
+   Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
+
+/* Handle large integers for calculating big tape lengths and the
+   like.  In practice, double precision does for now.  On the vast
+   majority of machines, it counts up to 2**52 bytes without any loss
+   of information, and counts up to 2**62 bytes if data are always
+   blocked in 1 kB boundaries.  We'll need arbitrary precision
+   arithmetic anyway once we get into the 2**64 range, so there's no
+   point doing anything fancy before then.  */
+
+#define TARLONG_FORMAT "%.0f"
+typedef double tarlong;