Changes committed to git://git.pcp.io/kenj/pcp master
Ken McDonell (4):
src/libpcp/{secureconnect.c,secureserver.c} tigher guard on
__pmCloseChannelbyFD
qa/651: better diagnostics
qa/669: improved diagnostics on the failure path
qa/832: tweak filter after latest libpcp change
qa/651 | 96 ++++++++++++++++++++++++++++++++++-------
qa/669 | 4 -
qa/832 | 3 +
qa/832.out | 2
src/libpcp/src/secureconnect.c | 8 ++-
src/libpcp/src/secureserver.c | 8 ++-
6 files changed, 98 insertions(+), 23 deletions(-)
Details ...
commit 665f57c349352eff484aa398328ecbd8b5cfbdee
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Aug 4 17:16:29 2016 +1000
qa/832: tweak filter after latest libpcp change
A little more "close channel" chatter to deal with.
commit 4951e34445ac3c1d74bdc1386b0da2ccdd9bba31
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Aug 4 17:16:02 2016 +1000
qa/669: improved diagnostics on the failure path
commit 71e8052998cd13601144d16b872fb24cff41a1ba
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Aug 4 17:15:46 2016 +1000
qa/651: better diagnostics
commit 5a626e3771473c199eca854e7eb04fd39842aa52
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Thu Aug 4 17:10:54 2016 +1000
src/libpcp/{secureconnect.c,secureserver.c} tigher guard on
__pmCloseChannelbyFD
When expecting a PDU_ATTR (which is the same type as PDU_AUTH, just
to confuse those who are clueless in this part of the code?), it
is apparently OK to receive a PDU of type 0 ... if this happens
do not close the channel to pmcd, and qa/832 is happy.
|