Comment 7 for bug 173757

Btw, I think you made a mistake in your proposed patch in that it now says "gchar *buffer[".
Unless your architecture of choice has single-byte pointers, I expect that will spell an
incorrect value of "buffer + 10" in the places where it's used.

Other than that it looks ok as a stopgap fix. I've been running a locally modified version of
the package with a 511 byte buffer for quite some time without encountering any concrete
problems.