When i run the SQL queries from the failed test cases on an sqlite shell, all of them seem to have the right opcodes. For Example, SQL queries that are supposed to have OPENEPHEMERAL opcodes when executed on a shell lists the OPENEPHEMERAL opcode. But when a "puts <output of explain cmd>" printed from within the test case, the OPENEPHEMERAL opcode is missing. But for the failed count-* tests (these search for Count opcode), the opcode was missing in both test scenarios i.e. SQLite shell as well as a puts from the test script.
On Sat, May 9, 2009 at 12:39 AM, D. Richard Hipp <d...@hwaci.com> wrote: > > On May 8, 2009, at 2:11 PM, mwnn wrote: > > > Hi, > > Sorry about the previous mail. > > The following test cases have failed when executing "make test" on > > Linux > > running on MIPS platform: > > 28 errors out of 40151 tests > > Failures on these tests: count-2.1 count-2.5 in3-1.6 in3-1.7 in3-1.8 > > in3-1.9 > > in3-1.11 in3-1.12 in3-1.13 in3-1.14 in3-3.5 in3-3.7 in3-4.3 in3-4.4 > > insert-5.3 insert5-2.2 insert5-2.3 insert5-2.4 insert5-2.5 insert5-2.6 > > insert5-2.8 misc3-6.10 nan-4.14 nan-4.15 nan-4.16 nan-4.17 > > trigger9-1.2.3 > > trigger9-1.5.2 > > The nan-* errors are probably because your hardware does not support > 80-bit floating point using "long double" and so the value 9.88e-324 > rounds to 0.0. Not something to worry about. > > The other errors all seem to involve the use of EXPLAIN. It would > appear that EXPLAIN is busted on your build. If you do not use > EXPLAIN, then this is probably nothing to worry about either. > > > D. Richard Hipp > d...@hwaci.com > > > > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
