Thread: initdb failed due to syntax error

initdb failed due to syntax error

From
jsco47@yahoo.com (jeff)
Date:
I recently downloaded and built Postgresql 7.2.2.  In trying to run
the "gmake check" after the build I get an error indicating that
initdb failed.  I turned on the debug option so the text below
contains some of the debug statements output from "gmake check".  Note
the error at the bottom "ERROR:  syntax error at line 2658: unexpected
token parse error".  It seems like this is complaining about the
syntax of line 2658 in the postgres.bki file, but I don't see any
problem with the syntax here.

Does anyone have any ideas about the problem here?

Thanks,
Jeff Stevens





DEBUG:  start transaction
DEBUG:  relation created with oid 16406
DEBUG:  commit transaction
DEBUG:  start transaction
DEBUG:  open relation pg_aggregate, attrsize 66
DEBUG:  create attribute 0 name aggname len 32 num 1 type 19
DEBUG:  create attribute 1 name aggowner len 4 num 2 type 23
DEBUG:  create attribute 2 name aggtransfn len 4 num 3 type 24
DEBUG:  create attribute 3 name aggfinalfn len 4 num 4 type 24
DEBUG:  create attribute 4 name aggbasetype len 4 num 5 type 26
DEBUG:  create attribute 5 name aggtranstype len 4 num 6 type 26
DEBUG:  create attribute 6 name aggfinaltype len 4 num 7 type 26
DEBUG:  create attribute 7 name agginitval len -1 num 8 type 25
DEBUG:  commit transaction
ERROR:  syntax error at line 2658: unexpected token parse error

initdb failed.
Data directory /array/web/src/postgresql-7.2.2/src/test/regress/./tmp_check/datawill not be removed at user's request.


Re: initdb failed due to syntax error

From
Tom Lane
Date:
jsco47@yahoo.com (jeff) writes:
> I recently downloaded and built Postgresql 7.2.2.  In trying to run
> the "gmake check" after the build I get an error indicating that
> initdb failed.  I turned on the debug option so the text below
> contains some of the debug statements output from "gmake check".  Note
> the error at the bottom "ERROR:  syntax error at line 2658: unexpected
> token parse error".  It seems like this is complaining about the
> syntax of line 2658 in the postgres.bki file, but I don't see any
> problem with the syntax here.

Why not show us what's in postgres.bki around that line, rather than
assuming you know what's correct?
        regards, tom lane