Thread: initdb fails: problem with array?

The current snapshot won't initdb, because running 
information_schema.sql fails.
The two occurences of "WHERE u.usesysid = ANY( g.grolist )" are the 
problem. Replacing the ANY clause with some dummy will let everything 
run ok.

select usename from pg_user, pg_group where usesysid = ANY (grolist)
will yield ERROR: syntaxt error at or near "grolist", so I believe that 
the ANY(ARRAY) support is broken.

Regards,
Andreas

Andreas Pflug wrote:
> The current snapshot won't initdb, because running 
> information_schema.sql fails.
> The two occurences of "WHERE u.usesysid = ANY( g.grolist )" are the 
> problem. Replacing the ANY clause with some dummy will let everything 
> run ok.
> 
> select usename from pg_user, pg_group where usesysid = ANY (grolist)
> will yield ERROR: syntaxt error at or near "grolist", so I believe that 
> the ANY(ARRAY) support is broken.
> 

I'm not seeing any problems with a fresh copy of cvs from 15 minutes 
after you posted. Are you still having the problem?

Joe

Joe Conway wrote:

> Andreas Pflug wrote:
>
>> The current snapshot won't initdb, because running 
>> information_schema.sql fails.
>> The two occurences of "WHERE u.usesysid = ANY( g.grolist )" are the 
>> problem. Replacing the ANY clause with some dummy will let everything 
>> run ok.
>>
>> select usename from pg_user, pg_group where usesysid = ANY (grolist)
>> will yield ERROR: syntaxt error at or near "grolist", so I believe 
>> that the ANY(ARRAY) support is broken.
>>
>
> I'm not seeing any problems with a fresh copy of cvs from 15 minutes 
> after you posted. Are you still having the problem?
>
I just checked out (at the moment hba.c changed, so I had to redo it), 
make clean, make, pgsql stop, make install, pgsql start and it's still 
there.

Regards,
Andreas

Andreas Pflug <pgadmin@pse-consulting.de> writes:
> I just checked out (at the moment hba.c changed, so I had to redo it), 
> make clean, make, pgsql stop, make install, pgsql start and it's still 
> there.

It works fine for me too.  Try removing 
src/backend/parser/gram.c and src/backend/parser/parse.h
to force bison to get re-run when you build.
        regards, tom lane

Tom Lane wrote:

>Andreas Pflug <pgadmin@pse-consulting.de> writes:
>  
>
>>I just checked out (at the moment hba.c changed, so I had to redo it), 
>>make clean, make, pgsql stop, make install, pgsql start and it's still 
>>there.
>>    
>>
>
>It works fine for me too.  Try removing 
>src/backend/parser/gram.c and src/backend/parser/parse.h
>to force bison to get re-run when you build.
>
>  
>

Thanks Tom,

rerunning bison did the job.

Regards,
Andreas