Обсуждение: [HACKERS] Problems with LSEGs

============================================================================
                        POSTGRES BUG REPORT TEMPLATE
============================================================================


Your name               :       David Smith
Your email address      :       dasmith@perseus.tufts.edu


System Configuration
- ---------------------
  Architecture (example: Intel Pentium)         :   DEC Alpha

  Operating System (example: Linux 1.3.42 ELF)  :   Digital UNIX 4.0

  PostgreSQL version (example: Postgres95-1.01) :   PostgreSQL-6.1 (2 June)

  Compiler used (example:  gcc 2.7.0)           :   DEC cc


Please enter a FULL description of your problem:
- ------------------------------------------------

The LSEGs are worse in today's distribution than they were last week.
Simply trying to select rows with LSEGs dumps core on me.

I can't really send you more information. I was trying to track down a bug
in dist_ps, but my sense from last week's snapshot is that the problems
with dist_ps were *not* related to the slope of the segment.

Please describe a way to repeat the problem.   Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------

foo=> \i segtest.sql
create table a (name text,p point);
CREATE
create table b (name text,s lseg);
CREATE
insert into a values ('point1', '(5,5)'::point );
INSERT 147742
insert into b values ('seg1', '[(1,1),(9,5)]' );
INSERT 147743
select * from b;
FATAL:  unexpected results from the backend, it probably dumped core.FATAL:  une
xpected results from the backend, it probably dumped core.EOF
foo=>



If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

------------------------------

David Smith wrote:
>   Architecture (example: Intel Pentium)         :   DEC Alpha
>   Operating System (example: Linux 1.3.42 ELF)  :   Digital UNIX 4.0
>   PostgreSQL version (example: Postgres95-1.01) :   PostgreSQL-6.1 (2 June)
>   Compiler used (example:  gcc 2.7.0)           :   DEC cc
...
> The LSEGs are worse in today's distribution than they were last week.
> Simply trying to select rows with LSEGs dumps core on me.
...
Well, as far as I know lsegs haven't changed recently, but with lots of
small changes recently something might have or...

I tried your test case (which is a _lot_ simpler than the regression
test) and did not find a problem. On the assumption that you are seeing
an alignment problem, I checked pg_type.h for the alignment definition
for lseg, but it is specified as "d" which should be 8-byte aligned and
_should_ be OK with the Alpha.

I don't have any more visibility into your problem, but if you want to
pursue it you can try embedding some printf() statements into geo_ops.c
to isolate the misbehaving code.

            - Tom

------------------------------

End of hackers-digest V1 #374
*****************************