[postgis-devel] [PostGIS] #179: ST_MakeLine and ST_MakeLine_Garry crash server with null arrays

[PostGIS]

#179: ST_MakeLine and ST_MakeLine_Garry crash server with null arrays
----------------------+-----------------------------------------------------
  Reporter:  robe     |       Owner:  pramsey      
      Type:  defect   |      Status:  reopened     
  Priority:  medium   |   Milestone:  postgis 1.3.7
 Component:  postgis  |     Version:  1.3.X        
Resolution:           |    Keywords:               
----------------------+-----------------------------------------------------
Comment (by robe):

 Nicklas,
 Ah indeed you are right.  Its still broken in 1.4 and 1.5.

 Well the funny thing is my 1.4.1SVN I can't get it to break, but I was
 able to get it to break on 1.4.0 RTM and my 1.5 compiled from 8/5. So
 maybe Mark fixed for the aggs (or  forgot about ST_MakeLine array version)

 Can you get it to break against 1.4 branch?

 I presume I must have missed this because it doesn't always break.  Just
 sometimes. and probably the times I was testing it it was behaving.

 Regarding the note about STRICT.  I don't think its the aggregate we mark
 as strict, but the state function.  pgis_geometry_accum_transfn (for 1.4+)
 and for (1.3+) -- st_geom_accum and that should unless I am missing
 something fix the aggregate issue.

 To fix the ST_Unite... array version -- we would mark that as STRICT as
 well.  That however I think would just prevent people from using it with
 NULL arrays (or hmm arrays that contain NULL).  Not sure which.  Haven't
 tested to tel.  It wouldn't be ideal, but I presume if 1.3 is a hard fix
 its better to get an error and the correct answer than to have a crashing
 server.  For 1.3 those array functions are not documented.  So would just
 be abusive people like me looking behind the hood using it.  The world
 needs to be protected from people like me of course :).

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/179#comment:18>
PostGIS <http://trac.osgeo.org/postgis/>
PostGIS