Re: [HACKERS] nested hstore patch - FailedAssertion("!(value->array.nelems == 1)
On Wed, January 15, 2014 09:46, Oleg Bartunov wrote:
>> On Wed, January 15, 2014 08:01, Oleg Bartunov wrote:
>>> It doesn't crashed in the last version in our repository.
>>>
>>> =# select 'x'::hstore || ('a=>"1"':: hstore) ;
>>>?column?
>>> ---
>>> "x", "a", "1"
>>> (1 row)
>>>
>>
>> OK, shall I use that repository instead of the latest posted patch?
I now installed from:
https://github.com/feodor/postgres
and compiled both a 'fast' and a 'debug' server (=with --enable-cassert see [1])
It turns out that the statement does not crash on a server compiled without
--enable-cassert.
But a compile with --enable-cassert shows that a bug is still lurking:
testdb=# select 'x'::hstore || ('a=>"1"':: hstore) ;
The connection to the server was lost. Attempting reset: Failed.
!>
TRAP: FailedAssertion("!(value->array.nelems == 1)", File: "hstore_support.c",
Line: 896)
Not good.
( please note that the assert is in a different file ('hstore_support.c') from
the earlier assert error that I posted )
Thanks,
Erik Rijkers
[1]
pg_config:
'--prefix=/home/aardvark/pg_stuff/pg_installations/pgsql.nested_hstore_url'
'--bindir=/home/aardvark/pg_stuff/pg_installations/pgsql.nested_hstore_url/bin'
'--libdir=/home/aardvark/pg_stuff/pg_installations/pgsql.nested_hstore_url/lib'
'--with-pgport=46541' '--enable-depend'
'--enable-cassert' '--enable-debug' '--with-openssl' '--with-perl'
'--with-libxml' '--with-libxslt' '--with-zlib'
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Re: [HACKERS] nested hstore patch - FailedAssertion("!(value->array.nelems == 1)
https://github.com/feodor/postgres
On Wed, Jan 15, 2014 at 11:10 AM, Erik Rijkers wrote:
> On Wed, January 15, 2014 08:01, Oleg Bartunov wrote:
>> It doesn't crashed in the last version in our repository.
>>
>> =# select 'x'::hstore || ('a=>"1"':: hstore) ;
>>?column?
>> ---
>> "x", "a", "1"
>> (1 row)
>>
>
> OK, shall I use that repository instead of the latest posted patch?
> No point in testing old code ( I used nested_hstore_and_jsonb-2.patch ).
>
> Could you send a link to where that repository is?
>
> ( btw, your query is not quite the same as the one I used:
> select 'x' || ('a=>"1"':: hstore)
> but your query also crashes my server here so I suppose
> it triggers the same bug )
>
>
> thanks,
>
> Erik Rijkers
>
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Re: [HACKERS] nested hstore patch - FailedAssertion("!(value->array.nelems == 1)
On Wed, January 15, 2014 08:01, Oleg Bartunov wrote:
> It doesn't crashed in the last version in our repository.
>
> =# select 'x'::hstore || ('a=>"1"':: hstore) ;
>?column?
> ---
> "x", "a", "1"
> (1 row)
>
OK, shall I use that repository instead of the latest posted patch?
No point in testing old code ( I used nested_hstore_and_jsonb-2.patch ).
Could you send a link to where that repository is?
( btw, your query is not quite the same as the one I used:
select 'x' || ('a=>"1"':: hstore)
but your query also crashes my server here so I suppose
it triggers the same bug )
thanks,
Erik Rijkers
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
Re: [HACKERS] nested hstore patch - FailedAssertion("!(value->array.nelems == 1)
It doesn't crashed in the last version in our repository.
=# select 'x'::hstore || ('a=>"1"':: hstore) ;
?column?
---
"x", "a", "1"
(1 row)
On Wed, Jan 15, 2014 at 3:53 AM, Erik Rijkers wrote:
> On Mon, January 13, 2014 16:36, Andrew Dunstan wrote:
>> A new version of the patch is attached. It includes all of Erik's docs
>
>> [ nested_hstore_and_jsonb-2.patch ]
>
>
> This crashes the server:
>
> testdb=# select 'x' || ('a=>"1"':: hstore) ;
> The connection to the server was lost. Attempting reset: Failed.
>
> logging:
> TRAP: FailedAssertion("!(value->array.nelems == 1)", File: "jsonb_support.c",
> Line: 904)
> 2014-01-15 00:32:01.854 CET 1206 LOG: server process (PID 3918) was
> terminated by signal 6: Aborted
> 2014-01-15 00:32:01.854 CET 1206 DETAIL: Failed process was running: select
> 'x' || ('a=>"1"':: hstore) ;
>
>
> Btw, I find it strange that:
>
> testdb=# select ('a=>""':: hstore) #%> '{a}' ;
> ?column?
> --
> ""
> (1 row)
>
> so that:
>
> Time: 0.641 ms
> testdb=# select ( ('a=>""':: hstore) #%> '{a}' ) = '' ;
> ?column?
> --
> f
> (1 row)
>
> testdb=# select ( ('a=>""':: hstore) #%> '{a}' ) = '""' ;
> ?column?
> --
> t
> (1 row)
>
> Maybe there is a rationale, but it seems to me that
>('a=>""':: hstore) #%> '{a}'
> should deliver the empty string '', and not two double quotes.
>
>
> Thanks,
>
> Erik Rijkers
>
>
>
>
>
>
>
>
> --
> 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: [HACKERS] nested hstore patch - FailedAssertion("!(value->array.nelems == 1)
On Mon, January 13, 2014 16:36, Andrew Dunstan wrote:
> A new version of the patch is attached. It includes all of Erik's docs
> [ nested_hstore_and_jsonb-2.patch ]
This crashes the server:
testdb=# select 'x' || ('a=>"1"':: hstore) ;
The connection to the server was lost. Attempting reset: Failed.
logging:
TRAP: FailedAssertion("!(value->array.nelems == 1)", File: "jsonb_support.c",
Line: 904)
2014-01-15 00:32:01.854 CET 1206 LOG: server process (PID 3918) was terminated
by signal 6: Aborted
2014-01-15 00:32:01.854 CET 1206 DETAIL: Failed process was running: select
'x' || ('a=>"1"':: hstore) ;
Btw, I find it strange that:
testdb=# select ('a=>""':: hstore) #%> '{a}' ;
?column?
--
""
(1 row)
so that:
Time: 0.641 ms
testdb=# select ( ('a=>""':: hstore) #%> '{a}' ) = '' ;
?column?
--
f
(1 row)
testdb=# select ( ('a=>""':: hstore) #%> '{a}' ) = '""' ;
?column?
--
t
(1 row)
Maybe there is a rationale, but it seems to me that
('a=>""':: hstore) #%> '{a}'
should deliver the empty string '', and not two double quotes.
Thanks,
Erik Rijkers
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
