Hello!
On Sun, 13 Dec 1998, Thomas G. Lockhart wrote:
> OK, I'm not arguing strongly against it, but imho the current code is
> not broken on any major *mature* platform :) The glibc2 stuff is fairly
> new, and everyone is still working out the details. We apparently
> understand the problem, but we haven't seen Oleg's patches to know what
> he actually did, whether the patches we are planning fix his problem, or
> what the effects are on machines (even of the same type) outside of his
> own.
I haven't worked out a real patch yet. I did a little patch for
configure.in/config.h.in, but you already committed the same patch. :) I did a patch for datetime regression test
(attached).The patch
triggers an error on glibc2. Works right on Solaris. I am not sure whether
we should include the patch into regression test...
Oleg.
---- Oleg Broytmann http://members.tripod.com/~phd2/ phd2@earthling.net Programmers don't die, they
justGOSUB without RETURN.
*** ./src/test/regress/expected/datetime.out.orig Mon Jan 5 06:35:27 1998
--- ./src/test/regress/expected/datetime.out Wed Dec 9 15:47:20 1998
***************
*** 28,33 ****
--- 28,40 ---- @ 0 (1 row)
+ QUERY: SET DateStyle = 'Postgres,noneuropean';
+ QUERY: SELECT datetime('1991-10-01', '11:00') AS "Oct_01_11am_1991";
+ Oct_01_11am_1991
+ ----------------------------
+ Tue Oct 01 11:00:00 1991 PDT
+ (1 row)
+ QUERY: CREATE TABLE DATETIME_TBL( d1 datetime); QUERY: INSERT INTO DATETIME_TBL VALUES ('current'); QUERY: INSERT
INTODATETIME_TBL VALUES ('today');
*** ./src/test/regress/sql/datetime.sql.orig Sat Nov 15 05:55:57 1997
--- ./src/test/regress/sql/datetime.sql Wed Dec 9 15:43:55 1998
***************
*** 10,15 ****
--- 10,18 ---- SELECT ('current'::datetime = 'now'::datetime) as "True"; SELECT ('now'::datetime - 'current'::datetime)
AS"ZeroSecs";
+ SET DateStyle = 'Postgres,noneuropean';
+ SELECT datetime('1991-10-01', '11:00') AS "Oct_01_11am_1991";
+ CREATE TABLE DATETIME_TBL( d1 datetime); INSERT INTO DATETIME_TBL VALUES ('current');