[ntp:hackers] make check thrashing: looks like test_recv_packet in test-ntp_signd

Hal Murray hmurray at megapathdsl.net
Sat Dec 12 06:38:01 UTC 2015

stenn at ntp.org said:
> Some of the tests for ntp_signd are ... difficult to do properly.  The
> current test code is an initial attempt at some of these and needs
> additional work.

> These tests should all complete quickly - if a test is taking a long time
> that's probably a bug. 

I think this is the cause of the out-of-memory bug I reported a while ago.

I just added a back-trace

I can't figure out what test_recv_packet is trying to do.  There is a comment 
that says:
** HMS: What's going on here?
** Looks like this needs more work.
so maybe I'm not the only one.

It's opening a unix socket, then reading from it.  It looks like maybe it's 
expecting the stuff it wrote in the previous test to be there.  But that test 
opened/wrote/closed a socket with the same name.  I assume the close 
discarded everything.  But then I don't understand where the garbage the read 
is getting is coming from.

A sanity check in recv_packet on the length might be a good idea.

These are my opinions.  I hate spam.

More information about the hackers mailing list