[postgis-users] Postgis 1.5.3 regression test failures

[Sandro Santilli]

On Thu, Oct 20, 2011 at 10:10:21AM +0200, Sandro Santilli wrote:
> On Thu, Oct 20, 2011 at 10:27:29AM +1030, Stephen Davies wrote:
> > I totally zapped the postgis source tree and started again from the tarball.
> > 
> > Same result.
> > 
> > However, this time I captured the PostgreSQL log. For the period of the make 
> > check run, the log is 5577 lines long but the first few lines are:
> > 
> > LOG:  database system is ready to accept connections
> > LOG:  autovacuum launcher started
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > LOG:  unexpected EOF on client connection
> > ERROR:  parse error - invalid geometry at character 14
> > HINT:  "POINT()" <-- parse error at position 7 within geometry
> > STATEMENT:  select '50', 'POINT()'::GEOMETRY as geom;
> > ERROR:  parse error - invalid geometry at character 14
> > HINT:  "POINT(1)" <-- parse error at position 8 within geometry
> > STATEMENT:  select '51', 'POINT(1)'::GEOMETRY as geom;
> > ERROR:  parse error - invalid geometry at character 14
> > HINT:  "POINT(," <-- parse error at position 7 within geometry
> > STATEMENT:  select '52', 'POINT(,)'::GEOMETRY as geom;
> > ERROR:  parse error - invalid geometry at character 14
> > 
> > Obviously I am missing something crucial but I cannot see what.
> 
> A bug in the code producing those strings, obviously.

Oops, I now see that's actually _expected_ as those bogus strings
are intentionally passed to the parser to test error conditions :)

So now I guess the problem is figuring out why regression test was
failing. Were you not getting the ERROR strings in output ?

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html