I got some repeatable segfaults with Gentoo build of nget, but after manually rebuilding from the source tarball, with --enable-debug, I no longer get those. The sigsegv was caused by a read of 4 bytes from address 0x148 (or so). This usually means reference by p[0x148], where p is a null pointer. It happened inside nget itself, but I didn't have debug enabled, so I don't know in more detail. The bug may be related to handling big files. For the record: the machine this was tested on is rock-solid and doesn't have any memory faults or anything. It has been in production use with 60 days of uptime now (running 2.6.4-cko4).
I have also gotten segs with nget 0.26, I usually just clear the cache and it ends up not being an issue any longer. I don't know, but maybe #54387 will help/fix.
0.27 in cvs. did it fix the problem?