GDB commands making debugging Postgres and Greenplum internals easier. Currently this provides a single command 'pgprint' that understand the 'Node' structures used internally, can can print them semi-intelligently.
Forked from PostgreSQL debugging tool found here:
https://github.com/tvondra/gdbpg
Copy the gdbpg.py script somewhere, and load it from gdb:
$ gdb ...
(gdb) source /...path.../gdbpg.py
and then just use pgprint command to print variables from gdb console.
E.g. if plan is pointing to (PlannedStmt *), you may do this:
(gdb) pgprint plan
and you'll get something like this:
PlannedStmt [commandType=CMD_SELECT planGen=PLANGEN_PLANNER queryId=0 hasReturning=false hasModifyingCTE=false canSetTag=true transientPlan=false oneoffPlan=false
simplyUpdatable=false dependsOnRole=false parallelModeNeeded=false numSlices=2 slices=0x56059fb60a08 subplan_sliceIds=0x56059fb61458 rewindPlanIDs=0x0
nParamExec=0 query_mem=0 metricsQueryType=0 '\000']
[planTree]
-> Motion [startup_cost=0 total_cost=2989 plan_rows=96300 plan_width=4 parallel_aware=false plan_node_id=0]
[motionType=MOTIONTYPE_GATHER sendSorted=false motionID=1 collations=0x0]
[targetlist]
TargetEntry [resno=1 resname="a" resorigtbl=16392 resorigcol=1]
Var [varno=OUTER_VAR varattno=1 vartype=23 varnoold=1 varoattno=1]
[flow] Flow [flotype=FLOW_SINGLETON locustype=CdbLocusType_Entry segindex=-1 numsegments=3]
[lefttree]
-> SeqScan [startup_cost=0 total_cost=1063 plan_rows=96300 plan_width=4 parallel_aware=false plan_node_id=1]
[scanrelid=1]
[targetlist]
TargetEntry [resno=1 resname="a" resorigtbl=16392 resorigcol=1]
Var [varno=1 varattno=1 vartype=23 varnoold=1 varoattno=1]
[flow] Flow [flotype=FLOW_PARTITIONED locustype=CdbLocusType_Hashed segindex=0 numsegments=3]
[rtable]
RangeTblEntry [rtekind=RTE_RELATION relid=16392 relkind='r' jointype=JOIN_INNER funcordinality=false ctelevelsup=0 self_reference=false forceDistRandom=false
lateral=false inFromCl=true requiredPerms=2 selectedCols=0x00000400]
[relationOids] OidList: [16392]