On 01/04/2016 02:54 PM, Martin Buchholz wrote:
ZipFile/Assortment already tests empty entry names. Why didn't that
test catch any error? IOW, ;how can ZipFile/Assortment be improved?
The bug only occurs if there is no such empty name entry inside the target zip
file.
The ZipFile/Assortment probably would be the better place to put test. Did not
pay
attention to the old tests.
-Sherman
On Mon, Jan 4, 2016 at 1:02 PM, Xueming Shen<xueming.s...@oracle.com> wrote:
Hi, please help review the change for JDK-8146431
Issue: https://bugs.openjdk.java.net/browse/JDK-8146431
webrev: http://cr.openjdk.java.net/~sherman/8146431
Cause:
The parameter "ulen" in native version zip_util.c.ZIP_GetEntry() serves two
purposes, (1) the length of the name (2) whether or not to try adding a
slash if first try failed. It has an explicit check as
/* If no real length was passed in, we are done */
if (ulen == 0) {
break;
}
The new java implementation uses and checks the flag "addSlash", but it
still needs to check the "name.len" before adding any slash for next try.
Thanks,
Sherman
