Re: [GENERAL] Test for char errors

2015-06-18 Thread Michael Paquier
On Fri, Jun 19, 2015 at 8:29 AM, 夏高 wrote: > Thanks Michael! Could you tell me which option determines what expected > output is used? Have a look at results_differ() in pg_regress.c ;) The file selected as expected output is the one with less lines of diffs. -- Michael -- Sent via pgsql-gene

Re: [GENERAL] Test for char errors

2015-06-18 Thread 夏高
On Wed, Jun 17, 2015 at 5:15 PM, Gao wrote: >> I don't know why the files are not the same but tests all passed. Helps are >> appreciated, thanks! >Some tests have multiple expected outputs. In the case of char, there >is not only char.out, but as well char_1.out and char_2.out. In your >case cha

Re: [GENERAL] Test for char errors

2015-06-17 Thread Michael Paquier
On Wed, Jun 17, 2015 at 5:15 PM, 夏高 wrote: > I don't know why the files are not the same but tests all passed. Helps are > appreciated, thanks! Some tests have multiple expected outputs. In the case of char, there is not only char.out, but as well char_1.out and char_2.out. In your case char_1.out

[GENERAL] Test for char errors

2015-06-17 Thread 夏高
I downloaded psotgresql-9.4.4 source code and build it on Centos 6.5 x64 edition. Then I run 'make test' and it reported that 'All 145 tests passed'. But the expected output and actual output of test 'char' are not same. The expected output of in 'src/test/regress/expected/char.out' is: SELECT ''