summaryrefslogtreecommitdiff
path: root/libbb/xrealloc_vector.c
diff options
context:
space:
mode:
authorDenys Vlasenko2009-10-27 10:01:01 +0100
committerDenys Vlasenko2009-10-27 10:01:01 +0100
commit1502a8b4cebe66870d188cda393a620fd48f0f3b (patch)
tree1b3cdb1a2e7a3868fb99278936afc36b6b909256 /libbb/xrealloc_vector.c
parent12450dbeef1afee5ed27f8f3a2395edbdc7a9355 (diff)
downloadbusybox-1502a8b4cebe66870d188cda393a620fd48f0f3b.zip
busybox-1502a8b4cebe66870d188cda393a620fd48f0f3b.tar.gz
libbb/xrealloc_vector.c: better comment
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Diffstat (limited to 'libbb/xrealloc_vector.c')
-rw-r--r--libbb/xrealloc_vector.c5
1 files changed, 3 insertions, 2 deletions
diff --git a/libbb/xrealloc_vector.c b/libbb/xrealloc_vector.c
index bbd5ab8..98fa967 100644
--- a/libbb/xrealloc_vector.c
+++ b/libbb/xrealloc_vector.c
@@ -20,8 +20,9 @@
* idx step, plus one: if idx == 0x20, vector[] is resized to 0x31,
* thus last usable element is vector[0x30].
*
- * In other words: after xrealloc_vector(v, 4, idx) it's ok to use
- * at least v[idx] and v[idx+1], for all idx values.
+ * In other words: after xrealloc_vector(v, 4, idx), with any idx,
+ * it's ok to use at least v[idx] and v[idx+1].
+ * v[idx+2] etc generally are not ok.
*
* New elements are zeroed out, but only if realloc was done
* (not on every call). You can depend on v[idx] and v[idx+1] being