libzahl

commit e82b5e111223c34deb1e200c9f7cc2ecff9efb01
parent faf2cebb48ec44a14b025bbe858b66efb78577fa
Author: Mattias Andrée <maandree@kth.se>
Date:   Mon,  2 May 2016 01:02:41 +0200

zstr: do not calculate the exact output size, calculate something easier

Signed-off-by: Mattias Andrée <maandree@kth.se>

Diffstat:
M
src/zstr.c
 | 
14
+++++---------

1 file changed, 5 insertions(+), 9 deletions(-)
diff --git a/src/zstr.c b/src/zstr.c
@@ -70,15 +70,11 @@ zstr(z_t a, char *b, size_t n)
 	}
 
 	if (!n) {
-		/* This is not the most efficient way to handle this. It should
-		 * be faster to allocate buffers that sprintint_fix and
-		 * sprintint_min print to, and then allocate `b` and copy the
-		 * buffers in reverse order into `b`. However, that is an overly
-		 * complicated solution. You probably already know the maximum
-		 * length or do not care about performance. Another disadvantage
-		 * with calculating the length before-hand, means that it is not
-		 * possible to reallocate `b` if it is too small. */
-		n = zstr_length(a, 10);
+		/* Calculate a value that is at least the number of
+		 * digits required to store the string. The overshoot
+		 * is not too signicant. */
+		n = (20 * BITS_PER_CHAR / 64 + (BITS_PER_CHAR == 8)) * a->used;
+		/* Note, depends on a ≠ as ensure above. */
 	}
 
 	if (unlikely(!b) && unlikely(!(b = libzahl_temp_allocation = malloc(n + 1))))