I looked through the "definite" memory loss reported by valgrind and found a few cases corresponding to actual memory leaks. There are currently other leak reports, but those appear to be either false alarms caused by our not fastidiously freeing memory at shutdown or to result from leaks in the GSSAPI code (see #78 for details).
The first few commits here are redundant with a couple other pull requests (#72, #74, #76). Commit 8a8bd52 is the first that is unique to this pull request.
I looked through the "definite" memory loss reported by valgrind and found a few cases corresponding to actual memory leaks. There are currently other leak reports, but those appear to be either false alarms caused by our not fastidiously freeing memory at shutdown or to result from leaks in the GSSAPI code (see #78 for details).
The first few commits here are redundant with a couple other pull requests (#72, #74, #76). Commit 8a8bd52 is the first that is unique to this pull request.