Re: numeric regression test passes, but why?

[Chapman Flack]

On 01/11/2018 11:23 AM, Alvaro Herrera wrote:
> Dagfinn Ilmari Mannsåker wrote:
>> ilm...@ilmari.org (Dagfinn Ilmari Mannsåker) writes:
>>
>>> The behaviour seems to have changed in 9.6:
>>
>> Indeed, https://www.postgresql.org/docs/current/static/release-9-6.html
>> has the following entry:
>>
>>  * Improve the accuracy of the ln(), log(), exp(), and pow() functions
>>for type numeric (Dean Rasheed)
> 
> Well, the test line was added with that commit (7d9a4737c268), and
> indeed its comment says this used to fail:
> 
> +-- cases that used to error out
> +select 0.12 ^ (-25);
> + ?column?  
> +---
> + 104825960103961013959336.4983657883169110
> +(1 row)

And indeed, my starting message in this thread was that, even in my
recent (e35dba475a440f73dccf9ed1fd61e3abc6ee61db) build, make check
*succeeds*, and for all I can tell, that test *is* executed (it shows
up in the log, and if I re-run it with digits altered, it fails).

And then I can connect interactively to the same server (started
fresh, not the same instance 'make check' was running, haven't tried
to find a way to connect to that), issue the same select, and get
the division by zero.

This is somehow sensitive to some sort of state. But what sort?

-Chap

Re: numeric regression test passes, but why?

[Alvaro Herrera]

Dagfinn Ilmari Mannsåker wrote:
> ilm...@ilmari.org (Dagfinn Ilmari Mannsåker) writes:
> 
> > The behaviour seems to have changed in 9.6:
> 
> Indeed, https://www.postgresql.org/docs/current/static/release-9-6.html
> has the following entry:
> 
>  * Improve the accuracy of the ln(), log(), exp(), and pow() functions
>for type numeric (Dean Rasheed)

Well, the test line was added with that commit (7d9a4737c268), and
indeed its comment says this used to fail:

+-- cases that used to error out
+select 0.12 ^ (-25);
+ ?column?  
+---
+ 104825960103961013959336.4983657883169110
+(1 row)

-- 
Álvaro Herrerahttps://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Re: numeric regression test passes, but why?

[Chapman Flack]

On 01/11/2018 07:44 AM, Sergei Kornilov wrote:
> Hello
> I am surprised, but i can confirm error on versions prior 9.6: on 9.5, 9.4, 
> 9.3 same error. On 9.6 and 10 query works correctly

One of my tests (in fact, the one where I first noticed) was a build
from git a couple days ago at e35dba475a440f73dccf9ed1fd61e3abc6ee61db
though.

-Chap

Re: numeric regression test passes, but why?

[Dagfinn Ilmari Mannsåker]

ilm...@ilmari.org (Dagfinn Ilmari Mannsåker) writes:

> The behaviour seems to have changed in 9.6:

Indeed, https://www.postgresql.org/docs/current/static/release-9-6.html
has the following entry:

 * Improve the accuracy of the ln(), log(), exp(), and pow() functions
   for type numeric (Dean Rasheed)

> $ psql -p5495
> Line style is unicode.
> Border style is 2.
> psql (10.1, server 9.5.10)
> Type "help" for help.
>
> ilmari@[local]:5495 ~=# select 0.12 ^ (-25);
> ERROR:  division by zero
> ilmari@[local]:5495 ~=# \q
>
> $ psql -p5496
> Line style is unicode.
> Border style is 2.
> psql (10.1, server 9.6.6)
> Type "help" for help.
>
> ilmari@[local]:5496 ~=# select 0.12 ^ (-25);
> ┌───┐
> │ ?column?  │
> ├───┤
> │ 104825960103961013959336.4983657883169110 │
> └───┘
> (1 row)
>
> ilmari@[local]:5496 ~=# \q
>
> - ilmari

-- 
"A disappointingly low fraction of the human race is,
 at any given time, on fire." - Stig Sandbeck Mathisen

Re: numeric regression test passes, but why?

[Sergei Kornilov]

Hello
I am surprised, but i can confirm error on versions prior 9.6: on 9.5, 9.4, 9.3 
same error. On 9.6 and 10 query works correctly
All checked postgresql are x86_64 from postgresql debian repository: 
http://apt.postgresql.org/pub/repos/apt/ jessie-pgdg main

postgres=# select 0.5678::numeric operator(pg_catalog.^) (-85::numeric);
ERROR:  division by zero
postgres=# select version();
  version   
   
---
 PostgreSQL 9.5.10 on x86_64-pc-linux-gnu, compiled by gcc (Debian 4.9.2-10) 
4.9.2, 64-bit

Regards, Sergei

11.01.2018, 15:27, "Chapman Flack" :
> On 01/11/18 02:07, Tom Lane wrote:
>
>>  Hm, it won't help your sanity to know that those cases pass fine
>>  for me, interactively, on a couple of different machines:
>>  ...
>>  You sure you're using a stock build of Postgres? No handmade
>>  versions of operator ^ lying around?
>
> Interesting thought
>
> At $work:
> postgresql95-server.x86_64 9.5.10-1PGDG.rhel7 @pgdg95
>
> postgres=> select 0.5678 operator(pg_catalog.^) (-85);
> ERROR: division by zero
> postgres=> select 0.5678::numeric operator(pg_catalog.^) (-85::numeric);
> ERROR: division by zero
>
> Also at home in a gentoo 9.5.1 ebuild...
> Also in a build from e35dba475a440f73dccf9ed1fd61e3abc6ee61db in git.
>
> All x86_64 
>
> -Chap

Re: numeric regression test passes, but why?

[Dagfinn Ilmari Mannsåker]

Tom Lane  writes:

> Chapman Flack  writes:
>> I see there are some tests in src/test/regress:
>> [ that don't work for me ]
>> Is there some special GUC setting in effect during the make check
>> that would be different in my ordinary session? What else could
>> be different? This is making me question my sanity.
>
> Hm, it won't help your sanity to know that those cases pass fine
> for me, interactively, on a couple of different machines:
>
> $ psql postgres
> psql (11devel)
> Type "help" for help.
>
> postgres=#  select 0.12 ^ (-25);
>  ?column?  
> ---
>  104825960103961013959336.4983657883169110
> (1 row)
>
> postgres=# select 0.5678 ^ (-85);
> ?column?
> 
>  782333637740774446257.7719390061997396
> (1 row)
>
> You sure you're using a stock build of Postgres?  No handmade
> versions of operator ^ lying around?

The behaviour seems to have changed in 9.6:

$ psql -p5495
Line style is unicode.
Border style is 2.
psql (10.1, server 9.5.10)
Type "help" for help.

ilmari@[local]:5495 ~=# select 0.12 ^ (-25);
ERROR:  division by zero
ilmari@[local]:5495 ~=# \q

$ psql -p5496
Line style is unicode.
Border style is 2.
psql (10.1, server 9.6.6)
Type "help" for help.

ilmari@[local]:5496 ~=# select 0.12 ^ (-25);
┌───┐
│ ?column?  │
├───┤
│ 104825960103961013959336.4983657883169110 │
└───┘
(1 row)

ilmari@[local]:5496 ~=# \q

- ilmari
-- 
"A disappointingly low fraction of the human race is,
 at any given time, on fire." - Stig Sandbeck Mathisen

Re: numeric regression test passes, but why?

[Chapman Flack]

On 01/11/18 02:07, Tom Lane wrote:

> Hm, it won't help your sanity to know that those cases pass fine
> for me, interactively, on a couple of different machines:
> ...
> You sure you're using a stock build of Postgres?  No handmade
> versions of operator ^ lying around?

Interesting thought

At $work:
postgresql95-server.x86_64   9.5.10-1PGDG.rhel7  @pgdg95

postgres=> select 0.5678 operator(pg_catalog.^) (-85);
ERROR:  division by zero
postgres=> select 0.5678::numeric operator(pg_catalog.^) (-85::numeric);
ERROR:  division by zero

Also at home in a gentoo 9.5.1 ebuild...
Also in a build from e35dba475a440f73dccf9ed1fd61e3abc6ee61db in git.

All x86_64 

-Chap

Re: numeric regression test passes, but why?

[Tom Lane]

Chapman Flack  writes:
> I see there are some tests in src/test/regress:
> [ that don't work for me ]
> Is there some special GUC setting in effect during the make check
> that would be different in my ordinary session? What else could
> be different? This is making me question my sanity.

Hm, it won't help your sanity to know that those cases pass fine
for me, interactively, on a couple of different machines:

$ psql postgres
psql (11devel)
Type "help" for help.

postgres=#  select 0.12 ^ (-25);
 ?column?  
---
 104825960103961013959336.4983657883169110
(1 row)

postgres=# select 0.5678 ^ (-85);
?column?

 782333637740774446257.7719390061997396
(1 row)

You sure you're using a stock build of Postgres?  No handmade
versions of operator ^ lying around?

regards, tom lane

numeric regression test passes, but why?

[Chapman Flack]

I see there are some tests in src/test/regress:

sql/numeric.sql
expected/numeric.out

They pass. I see "numeric ... ok" in a make check.

I do not doubt they are being run, because if I edit numeric.sql
and fudge some digits, say around

-- cases that used to error out
select 0.12 ^ (-25);
select 0.5678 ^ (-85);

and I change the -25 to -26, numeric then fails in my next make check.

All that seems just as it should be. Why, then, if I try to duplicate
those exact tests in an interactive session, would this happen:

postgres=# select 0.12 ^ (-25);
ERROR:  division by zero
STATEMENT:  select 0.12 ^ (-25);

postgres=# select 0.5678 ^ (-85);
ERROR:  division by zero
STATEMENT:  select 0.5678 ^ (-85);

... they error out, that is, do exactly the thing the tests are there
to make sure they do not.

Is there some special GUC setting in effect during the make check
that would be different in my ordinary session? What else could
be different? This is making me question my sanity.

-Chap