Thread: Regression test failures
E250 single processor - Solaris 2.6 - gcc2.8.1 - snapshot from 050699 Several things in the regress testing concern me : int2 .. failed int4 .. failed int8 .. failed # geometry .. failed abstime .. failed tinterval .. failed # horology .. failed triggers .. failed plpgsql .. Bus Error - core dumped failed Any/all opinions on what these failures mean are greatly appreciated. Jonny Hinojosa TCA Internet Jonny@TCA.net 409.693.8885 x114
At 23:19 +0300 on 07/05/1999, Jonny Hinojosa wrote: > > E250 single processor - Solaris 2.6 - gcc2.8.1 - snapshot from 050699 Install bison, install flex, empty the config cache, config again. Herouth -- Herouth Maoz, Internet developer. Open University of Israel - Telem project http://telem.openu.ac.il/~herutma
[root@BigData /root]# bison --version GNU Bison version 1.27 [root@BigData /root]# flex --version flex version 2.5.4 I do not believe that is the problem. Thanks for the suggestion. I'm just flappen in the breeze now. Jonny -----Original Message----- From: owner-pgsql-general@postgreSQL.org [mailto:owner-pgsql-general@postgreSQL.org]On Behalf Of Herouth Maoz Sent: Sunday, May 09, 1999 7:12 AM To: Jonny Hinojosa; PostgreSQL General Subject: Re: [GENERAL] Regression test failures At 23:19 +0300 on 07/05/1999, Jonny Hinojosa wrote: > > E250 single processor - Solaris 2.6 - gcc2.8.1 - snapshot from 050699 Install bison, install flex, empty the config cache, config again. Herouth -- Herouth Maoz, Internet developer. Open University of Israel - Telem project http://telem.openu.ac.il/~herutma
At 17:49 +0300 on 09/05/1999, Jonny Hinojosa wrote: > [root@BigData /root]# bison --version > GNU Bison version 1.27 > [root@BigData /root]# flex --version > flex version 2.5.4 > > I do not believe that is the problem. Thanks for the suggestion. I'm just > flappen in the breeze now. Look in the config log and see if the config found them and used them. Also, check what the particular differences in the regression tests are. This may give you a clue as to the nature of the problem. Herouth -- Herouth Maoz, Internet developer. Open University of Israel - Telem project http://telem.openu.ac.il/~herutma
*** expected/int2.out Thu Apr 15 02:02:44 1999 --- results/int2.out Sun May 9 10:52:36 1999 *************** *** 7,13 **** QUERY: INSERT INTO INT2_TBL(f1) VALUES ('32767'); QUERY: INSERT INTO INT2_TBL(f1) VALUES ('-32767'); QUERY: INSERT INTO INT2_TBL(f1) VALUES ('100000'); ! ERROR: pg_atoi: error reading "100000": Numerical result out of range QUERY: INSERT INTO INT2_TBL(f1) VALUES ('asdf'); ERROR: pg_atoi: error in "asdf": can't parse "asdf" QUERY: SELECT '' AS five, INT2_TBL.*; --- 7,13 ---- QUERY: INSERT INTO INT2_TBL(f1) VALUES ('32767'); QUERY: INSERT INTO INT2_TBL(f1) VALUES ('-32767'); QUERY: INSERT INTO INT2_TBL(f1) VALUES ('100000'); ! ERROR: pg_atoi: error reading "100000": Result too large QUERY: INSERT INTO INT2_TBL(f1) VALUES ('asdf'); ERROR: pg_atoi: error in "asdf": can't parse "asdf" QUERY: SELECT '' AS five, INT2_TBL.*; ---------------------- So, this suggests that the only reason the test failed is that the words used in the error message are different from what was expected. Jonny -----Original Message----- From: owner-pgsql-general@postgreSQL.org [mailto:owner-pgsql-general@postgreSQL.org]On Behalf Of Herouth Maoz Sent: Sunday, May 09, 1999 10:30 AM To: Jonny Hinojosa; PostgreSQL General Subject: RE: [GENERAL] Regression test failures ~~~~~~ Look in the config log and see if the config found them and used them. Also, check what the particular differences in the regression tests are. This may give you a clue as to the nature of the problem. Herouth -- Herouth Maoz, Internet developer. Open University of Israel - Telem project http://telem.openu.ac.il/~herutma
At 19:01 +0300 on 09/05/1999, Jonny Hinojosa wrote: > > So, this suggests that the only reason the test failed is that the words > used in the error message are different from what was expected. Is that case, you can relax. Whenever you have regression troubles, you must check them. In the case of abstime, I found that the regression test tends to fail for the old dates because it doesn't consider daylight saving time before 1970 or so. So if that's what happens in your regression test, again, relax - unless you intend to use old dates in your applications including the time portion. Note that this is a system thing, not a postgres thing. Horology is derived from the same problem. Another acceptable discrepancy is when float comparisons fail over the eighth or so digit after the decimal point. You can dismiss it. Failures over parse errors are serious and usually relate to gram.c and its fathers bison and flex. A core dump is a serious matter, though. Herouth -- Herouth Maoz, Internet developer. Open University of Israel - Telem project http://telem.openu.ac.il/~herutma