Opened 11 years ago
Closed 11 years ago
#2672 closed defect (fixed)
ST_AsTWKBAgg fails NULL crash test
Reported by: | robe | Owned by: | nicklas |
---|---|---|---|
Priority: | blocker | Milestone: | PostGIS 2.2.0 |
Component: | postgis | Version: | master |
Keywords: | Cc: |
Description
Just ran thru garden 2.2 pot of tests and got a crasher on twkb. This is running on:
POSTGIS="2.2.0dev r12313" GEOS="3.4.2-CAPI-1.8.2 r3924" PROJ="Rel. 4.8.0, 6 March 2012" GDAL="GDAL 1.10.0, released 2013/04/24" LIBXML="2.7.8" LIBJSON="UNKNOWN" RASTER PostgreSQL 9.3.3, compiled by Visual C++ build 1600, 64-bit
SELECT ST_AsTWKBAgg(foo1.the_geom, 3) As result FROM ((SELECT CAST(Null As geometry) As the_geom)) As foo1 LIMIT 3;
errors with:
server closed the connection unexpectedly This probably means the server terminated abnormally before or while processing the request. psql:postgis_gardentest22.sql:272626: connection to server was lost
Change History (7)
comment:1 by , 11 years ago
Owner: | changed from | to
---|
comment:2 by , 11 years ago
Summary: | twkb fails NULL crash test → ST_AsTWKBAgg fails NULL crash test |
---|
comment:3 by , 11 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
comment:4 by , 11 years ago
Resolution: | fixed |
---|---|
Status: | closed → reopened |
strk seems to have run into the issue again -- strk comment on this one if you have confirmed you are running the latest and its still happening to you. Otherwise close.
comment:5 by , 11 years ago
It doesn't crash for me now on 64 bit Debian. But as I commented 4 weeks ago I don't know if the behavior is the most correct.
If it still crashes for you strk, what system are you on?
comment:6 by , 11 years ago
I don't see a testcase in r12318, could you please add one in regress/tickets.sql (and _expect) for this ?
SELECT '#2672', ST_AsTWKBAgg(null::geometry, 3)
I confirm it doesn't happen anymore (forgot to "make install" before running garden)
comment:7 by , 11 years ago
Resolution: | → fixed |
---|---|
Status: | reopened → closed |
Done in r12436
I close it.
Thanks for catching this!
Should be fixed in r12318
It now ignores null inputs and aggregates the other. That behavior can be questioned since that means the corresponding ID will be just lost.
The alternative I guess is putting some empty geometry in the aggregate, but I do not know if that is correct either. No big deal I guess.