summaryrefslogtreecommitdiff
path: root/src/internal
diff options
context:
space:
mode:
authorRich Felker <dalias@aerifal.cx>2012-08-17 16:53:09 -0400
committerRich Felker <dalias@aerifal.cx>2012-08-17 16:53:09 -0400
commit11458e5b098319cf3e2d05c8cbaa74d58db740e3 (patch)
tree027decb697492c454300b4cf7f72bff5c5f55e0e /src/internal
parentdc82ee4e30ec0eae367f7ab076a7325b4d858dd5 (diff)
downloadmusl-11458e5b098319cf3e2d05c8cbaa74d58db740e3.tar.gz
musl-11458e5b098319cf3e2d05c8cbaa74d58db740e3.tar.bz2
musl-11458e5b098319cf3e2d05c8cbaa74d58db740e3.tar.xz
musl-11458e5b098319cf3e2d05c8cbaa74d58db740e3.zip
fix float parsing logic for long decimal expansions
this affects at least the case of very long inputs, but may also affect shorter inputs that become long due to growth while upscaling. basically, the logic for the circular buffer indices of the initial base-10^9 digit and the slot one past the final digit, and for simplicity of the loop logic, assumes an invariant that they're not equal. the upscale loop, which can increase the length of the base-10^9 representation, attempted to preserve this invariant, but was actually only ensuring that the end index did not loop around past the start index, not that the two never become equal. the main (only?) effect of this bug was that subsequent logic treats the excessively long number as having no digits, leading to junk results.
Diffstat (limited to 'src/internal')
-rw-r--r--src/internal/floatscan.c2
1 files changed, 1 insertions, 1 deletions
diff --git a/src/internal/floatscan.c b/src/internal/floatscan.c
index 68f576c9..bba5753b 100644
--- a/src/internal/floatscan.c
+++ b/src/internal/floatscan.c
@@ -199,11 +199,11 @@ static long double decfloat(FILE *f, int c, int bits, int emin, int sign, int po
}
if (carry) {
rp += 9;
+ a = (a-1 & MASK);
if (a == z) {
z = (z-1 & MASK);
x[z-1 & MASK] |= x[z];
}
- a = (a-1 & MASK);
x[a] = carry;
}
}