It is possible that DistributedEngineComm can get the Stats object from
an MQE object and the MQE object freed before it's stats object is
passed to InetStreamSocket.
This patch makes sure that DistributedEngineComm gets a reference to MQE
instead of the pointer to the Stats object in another reference.
Therefore making sure that the Stats object still exists in
InetStreamSocket.
It is possible that DistributedEngineComm can get the Stats object from
an MQE object and the MQE object freed before it's stats object is
passed to InetStreamSocket.
This patch makes sure that DistributedEngineComm gets a reference to MQE
instead of the pointer to the Stats object in another reference.
Therefore making sure that the Stats object still exists in
InetStreamSocket.
Whilst very rare we can hit a case where we attempt to unlock objLock
when it is already unlocked. With the Boost version in Ubuntu 16.04 this
triggers an abort() effectively crashing PrimProc.
In this patch we switch to a pthread mutex instead which does not have
this limitation. At a later date we can look into refactoring how BPP
and this mutex works.
Whilst very rare we can hit a case where we attempt to unlock objLock
when it is already unlocked. With the Boost version in Ubuntu 16.04 this
triggers an abort() effectively crashing PrimProc.
In this patch we switch to a pthread mutex instead which does not have
this limitation. At a later date we can look into refactoring how BPP
and this mutex works.
The fix for MCOL-838 broke VARBINARY as it truncated on the first NUL on
StringStore deserialize. This fix uses append() to force a copy instead
whilst preserving length.
This fixes test012
When subqueries and group by are used in CrossEngine the first row group
is either corrupted or ignored. This is related to MCOL-430 which fixed
the case for FE1 mode.
When subqueries and group by are used in CrossEngine the first row group
is either corrupted or ignored. This is related to MCOL-430 which fixed
the case for FE1 mode.
