Re: contrib/cube & contrib/seg regression fixes - Mailing list pgsql-patches

From Tom Lane
Subject Re: contrib/cube & contrib/seg regression fixes
Date
Msg-id 26060.1032062613@sss.pgh.pa.us
Whole thread Raw
In response to Re: contrib/cube & contrib/seg regression fixes  (Joe Conway <mail@joeconway.com>)
List pgsql-patches
Joe Conway <mail@joeconway.com> writes:
> The message seems to be coming from bison:

> #0  elog (lev=20, fmt=0x82c038c "parse error, unexpected $undefined.,
> expecting FLOAT or O_PAREN or O_BRACKET at or before position 1, character
> ('A', \\101), input: 'ABC'\n") at elog.c:114
> #1  0x40019093 in cube_yyerror (msg=0xbfffe300 "parse error, unexpected
> $undefined., expecting FLOAT or O_PAREN or O_BRACKET") at cubeparse.y:204
> #2  0x40018dab in cube_yyparse (result=0xbfffe8b0) at
> /usr/share/bison/bison.simple:799

> I'm using bison-1.35. Recommendations, or should I ignore the regression failures?

AFAICT, bison 1.28 (what I have here) cannot emit any message mentioning
"unexpected"; the word just plain does not appear in bison.simple.
"unexpected $undefined." does not exactly seem like a major advance in
error-message technology anyway :-(.

I suggest we sit tight and hope that the bison guys get their act
together, soon.

            regards, tom lane

pgsql-patches by date:

Previous
From: Joe Conway
Date:
Subject: Re: contrib/cube & contrib/seg regression fixes
Next
From: Tom Lane
Date:
Subject: Re: Patch to return last tid