Обсуждение: failed test float8 on mingw

Поиск
Список
Период
Сортировка

failed test float8 on mingw

От
"Jaime Casanova"
Дата:
Hi,

Trying to review some patches on windows... mingw: 5.1.4; msys:
1.0.10; gcc 3.4.5
This is not current head but this morning snapshot tarball... but i
saw this same error with yesterday tarball...

I get this failure on regression tests:

expected output:
 SELECT '10e-400'::float8;
! ERROR:  "10e-400" is out of range for type double precision
! LINE 1: SELECT '10e-400'::float8;
!                ^ SELECT '-10e-400'::float8;
! ERROR:  "-10e-400" is out of range for type double precision
! LINE 1: SELECT '-10e-400'::float8;
!

actual result output:
 SELECT '10e-400'::float8;
!  float8
! --------
!       0
! (1 row)
! SELECT '-10e-400'::float8;
!  float8
! --------
!       0
! (1 row)


--
Atentamente,
Jaime Casanova
Soporte y capacitación de PostgreSQL
Asesoría y desarrollo de sistemas
Guayaquil - Ecuador
Cel. +59387171157


Re: failed test float8 on mingw

От
Magnus Hagander
Дата:
Do earlier versions work if built on this machine? Like 8.3-latest? My
guess is that it's something in mingw or the environment rather than
pg, but let's confirm that before trying to track down what..

/Magnus

On 11 nov 2008, at 06.56, "Jaime Casanova"
<jcasanov@systemguards.com.ec> wrote:

> Hi,
>
> Trying to review some patches on windows... mingw: 5.1.4; msys:
> 1.0.10; gcc 3.4.5
> This is not current head but this morning snapshot tarball... but i
> saw this same error with yesterday tarball...
>
> I get this failure on regression tests:
>
> expected output:
>
>  SELECT '10e-400'::float8;
> ! ERROR:  "10e-400" is out of range for type double precision
> ! LINE 1: SELECT '10e-400'::float8;
> !                ^
>  SELECT '-10e-400'::float8;
> ! ERROR:  "-10e-400" is out of range for type double precision
> ! LINE 1: SELECT '-10e-400'::float8;
> !
>
> actual result output:
>
>  SELECT '10e-400'::float8;
> !  float8
> ! --------
> !       0
> ! (1 row)
> !
>  SELECT '-10e-400'::float8;
> !  float8
> ! --------
> !       0
> ! (1 row)
>
>
> --
> Atentamente,
> Jaime Casanova
> Soporte y capacitación de PostgreSQL
> Asesoría y desarrollo de sistemas
> Guayaquil - Ecuador
> Cel. +59387171157
>
> --
> Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-hackers


Re: failed test float8 on mingw

От
Greg Stark
Дата:
At a guess it's related to the float8-is-pass-by-value patch.

greg

On 11 Nov 2008, at 06:17 PM, Magnus Hagander <magnus@hagander.net>
wrote:

> Do earlier versions work if built on this machine? Like 8.3-latest?
> My guess is that it's something in mingw or the environment rather
> than pg, but let's confirm that before trying to track down what..
>
> /Magnus
>
> On 11 nov 2008, at 06.56, "Jaime Casanova" <jcasanov@systemguards.com.ec
> > wrote:
>
>> Hi,
>>
>> Trying to review some patches on windows... mingw: 5.1.4; msys:
>> 1.0.10; gcc 3.4.5
>> This is not current head but this morning snapshot tarball... but i
>> saw this same error with yesterday tarball...
>>
>> I get this failure on regression tests:
>>
>> expected output:
>>
>> SELECT '10e-400'::float8;
>> ! ERROR:  "10e-400" is out of range for type double precision
>> ! LINE 1: SELECT '10e-400'::float8;
>> !                ^
>> SELECT '-10e-400'::float8;
>> ! ERROR:  "-10e-400" is out of range for type double precision
>> ! LINE 1: SELECT '-10e-400'::float8;
>> !
>>
>> actual result output:
>>
>> SELECT '10e-400'::float8;
>> !  float8
>> ! --------
>> !       0
>> ! (1 row)
>> !
>> SELECT '-10e-400'::float8;
>> !  float8
>> ! --------
>> !       0
>> ! (1 row)
>>
>>
>> --
>> Atentamente,
>> Jaime Casanova
>> Soporte y capacitación de PostgreSQL
>> Asesoría y desarrollo de sistemas
>> Guayaquil - Ecuador
>> Cel. +59387171157
>>
>> --
>> Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
>> To make changes to your subscription:
>> http://www.postgresql.org/mailpref/pgsql-hackers
>
> --
> Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-hackers


Re: failed test float8 on mingw

От
Tom Lane
Дата:
Greg Stark <greg.stark@enterprisedb.com> writes:
> At a guess it's related to the float8-is-pass-by-value patch.

I doubt it.  It looks like underflow-to-zero behavior, which we expect
on some other platforms ... but not mingw.
        regards, tom lane


Re: failed test float8 on mingw

От
"Jaime Casanova"
Дата:
On Tue, Nov 11, 2008 at 12:17 PM, Magnus Hagander <magnus@hagander.net> wrote:
> Do earlier versions work if built on this machine? Like 8.3-latest? My guess
> is that it's something in mingw or the environment rather than pg, but let's
> confirm that before trying to track down what..
>

actually seems there's something wrong with environment... (MinGW:
5.1.4; msys: 1.0.10; gcc: 3.4.5 (mingw-vista special r3))
mmm... that r3 doesn't like me...

8.3, FAILED same test
8.2, FAILED too

8.1 doesn't even compile and give this error

gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Winline
-Wdeclaration-after-statement -Wendif-labels -fno-strict-aliasing
-fwrapv -g -I../../src/port -DFRONTEND -I../../src/include
-I./src/include/port/win32 -DEXEC_BACKEND
"-I../../src/include/port/win32"  -c -o getrusage.o getrusage.c -MMD
In file included from ../../src/include/rusagestub.h:17,                from getrusage.c:18:
c:/mingw/bin/../lib/gcc/mingw32/3.4.5/../../../../include/sys/time.h:27:
error: redefinition of `struct timezone'
c:/mingw/bin/../lib/gcc/mingw32/3.4.5/../../../../include/sys/time.h:40:
error: conflicting types for 'gettimeofday'
../../src/include/port.h:266: error: previous declaration of
'gettimeofday' was here
c:/mingw/bin/../lib/gcc/mingw32/3.4.5/../../../../include/sys/time.h:40:
error: conflicting types for 'gettimeofday'
../../src/include/port.h:266: error: previous declaration of
'gettimeofday' was here


All of this, is with
./configure --prefix=/home/Administrador/pgsql-8.1 --enable-cassert
--enable-debug --enable-depend --with-pgport=5434 --without-zlib

Something i forgot to mention, because it doesn't seems related, is
that i had to manually include ntsecapi.h in security.h mingw's
library

--
Atentamente,
Jaime Casanova
Soporte y capacitación de PostgreSQL
Asesoría y desarrollo de sistemas
Guayaquil - Ecuador
Cel. +59387171157


Re: failed test float8 on mingw

От
Magnus Hagander
Дата:
Jaime Casanova wrote:
> On Tue, Nov 11, 2008 at 12:17 PM, Magnus Hagander <magnus@hagander.net> wrote:
>> Do earlier versions work if built on this machine? Like 8.3-latest? My guess
>> is that it's something in mingw or the environment rather than pg, but let's
>> confirm that before trying to track down what..
>>
> 
> actually seems there's something wrong with environment... (MinGW:
> 5.1.4; msys: 1.0.10; gcc: 3.4.5 (mingw-vista special r3))
> mmm... that r3 doesn't like me...
> 
> 8.3, FAILED same test
> 8.2, FAILED too

Hmm. So which parts of that differ from the buildfarm boxes that *do*
work, and why :-( I wonder if we're back to trolling mingw-cvs for
incompatible changes yet again...


> 8.1 doesn't even compile and give this error

This is known. It was a fix applied after 8.1 became unsupported on
Windows. Since mingw has a tendency to change their header files
incompatibly every now and then...

> Something i forgot to mention, because it doesn't seems related, is
> that i had to manually include ntsecapi.h in security.h mingw's
> library

There's a patch floating around for that somewhere as well. Once again,
they changed their includes files incompatibly...

//Magnus


Re: failed test float8 on mingw

От
"Jaime Casanova"
Дата:
On Tue, Nov 11, 2008 at 4:32 PM, Magnus Hagander <magnus@hagander.net> wrote:
> Jaime Casanova wrote:
>> On Tue, Nov 11, 2008 at 12:17 PM, Magnus Hagander <magnus@hagander.net> wrote:
>>> Do earlier versions work if built on this machine? Like 8.3-latest? My guess
>>> is that it's something in mingw or the environment rather than pg, but let's
>>> confirm that before trying to track down what..
>>>
>>
>> actually seems there's something wrong with environment... (MinGW:
>> 5.1.4; msys: 1.0.10; gcc: 3.4.5 (mingw-vista special r3))
>> mmm... that r3 doesn't like me...
>>
>> 8.3, FAILED same test
>> 8.2, FAILED too
>
> Hmm. So which parts of that differ from the buildfarm boxes that *do*
> work, and why :-( I wonder if we're back to trolling mingw-cvs for
> incompatible changes yet again...
>

well. i'm on a XP SP3 and using gcc 3.4.5... no animal in buildfarm
says that... maybe the gcc version? i'll try with 3.4.2

--
Atentamente,
Jaime Casanova
Soporte y capacitación de PostgreSQL
Asesoría y desarrollo de sistemas
Guayaquil - Ecuador
Cel. +59387171157


Re: failed test float8 on mingw

От
"Jaime Casanova"
Дата:
On Tue, Nov 11, 2008 at 4:48 PM, Jaime Casanova
<jcasanov@systemguards.com.ec> wrote:
> On Tue, Nov 11, 2008 at 4:32 PM, Magnus Hagander <magnus@hagander.net> wrote:
>> Jaime Casanova wrote:
>>> On Tue, Nov 11, 2008 at 12:17 PM, Magnus Hagander <magnus@hagander.net> wrote:
>>>> Do earlier versions work if built on this machine? Like 8.3-latest? My guess
>>>> is that it's something in mingw or the environment rather than pg, but let's
>>>> confirm that before trying to track down what..
>>>>
>>>
>>> actually seems there's something wrong with environment... (MinGW:
>>> 5.1.4; msys: 1.0.10; gcc: 3.4.5 (mingw-vista special r3))
>>> mmm... that r3 doesn't like me...
>>>
>>> 8.3, FAILED same test
>>> 8.2, FAILED too
>>
>> Hmm. So which parts of that differ from the buildfarm boxes that *do*
>> work, and why :-( I wonder if we're back to trolling mingw-cvs for
>> incompatible changes yet again...
>>
> maybe the gcc version? i'll try with 3.4.2
>

yeah... it was the gcc version... 3.4.2 works good, and passes all
regressions...

--
Atentamente,
Jaime Casanova
Soporte y capacitación de PostgreSQL
Asesoría y desarrollo de sistemas
Guayaquil - Ecuador
Cel. +59387171157