Opened 12 years ago

Closed 12 years ago

#1496 closed defect (fixed)

Bogus next_{left,right}_edge after ST_ModEdgeSplit or ST_NewEdgesSplit

Reported by: strk Owned by: strk
Priority: critical Milestone: PostGIS 2.0.0
Component: topology Version: master
Keywords: Cc:

Description

When adding an dangling edge with startpoint on isolated node and endpoint on a non-isolated one, the resulting next_left_edge and next_right_edge are bogus!

Change History (6)

comment:1 by strk, 12 years ago

On further analysis the issue is when the _endpoint_ point is isolated, not the _startpoint_.

comment:2 by strk, 12 years ago

Summary: Bogus next_{left,right}_edge on dangling edgeBogus next_{left,right}_edge after ST_

It's even more complex: the bug is on _splitting_ a dangling edge by a point

comment:3 by strk, 12 years ago

Priority: highcritical

Whoa, no tests for ST_ModEdgeSplit !!!

comment:4 by strk, 12 years ago

Summary: Bogus next_{left,right}_edge after ST_Bogus next_{left,right}_edge after ST_ModEdgeSplit

Even non-dangling edges have bogus linking info after ST_ModEdgeSplit…

comment:5 by strk, 12 years ago

Fixed by r8911, which includes regression testing.

ST_NewEdgesSplit is yet to be tested, better keep this open until it is.

comment:6 by strk, 12 years ago

Resolution: fixed
Status: newclosed
Summary: Bogus next_{left,right}_edge after ST_ModEdgeSplitBogus next_{left,right}_edge after ST_ModEdgeSplit or ST_NewEdgesSplit

r8912 fixes ST_NewEdgesSplit function, suffering of the same symptoms. Full test coverage now..

Note: See TracTickets for help on using tickets.