Request for approval: Bug 100057 (original) (raw)

Request for approval: Bug 100057 - Potential stack corruption in GetJavaProperties

Andrew Haley aph at redhat.com
Fri May 22 15:44:45 UTC 2009

David Holmes - Sun Microsystems wrote:

If you use malloc then you have to check for a NULL return and deal with the error possibility.

Alternatively use strncpy to make sure it's safe and continue to assume that it will be big enough.

I'm working on fixing this properly, but I just came across a gem of very confusing code:

        if ((encoding = strchr(p, '.')) != NULL) {
            p[encoding++ - p] = '\0';
            p = encoding;
        }

The middle line

            p[encoding++ - p] = '\0';

is unlike anything I have ever seen before. As far as I can tell it is equivalent to

            *encoding++ = '\0';

but expressed in a, er, unique way. I suppose the idea is that it shows that the store is within p.

I just had to share this...

Andrew.

More information about the core-libs-dev mailing list