Fujii Masao
masao****@gmail*****
2013年 10月 3日 (木) 19:15:48 JST
On Thu, Oct 3, 2013 at 5:47 PM, Beena Emerson <memis****@gmail*****> wrote: > On Wed, Oct 2, 2013 at 8:14 PM, Fujii Masao <masao****@gmail*****> wrote: >> >> On Tue, Oct 1, 2013 at 8:55 PM, Beena Emerson <memis****@gmail*****> >> wrote: >> > On Mon, Sep 30, 2013 at 10:29 PM, Fujii Masao <masao****@gmail*****> >> > wrote: >> >> >> >> Hmm... I was not able to reproduce this problem. I renamed cnt_bigm_sml >> >> to >> >> cnt_sml in the patch you had submitted, applied it to HEAD of pg_bigm, >> >> and >> >> executed similarity('wow', 'wow'). Its result was 1. >> > >> > >> > Even I tried after changing cnt_bigm_cml to cnt_sml in my patch but I >> > got >> > wrong result: >> > >> > test_db=# CREATE EXTENSION pg_trgm ; >> > CREATE EXTENSION >> > test_db=# CREATE EXTENSION pg_bigm; >> > CREATE EXTENSION >> > test_db=# SELECT similarity('wow','wow'); >> > similarity >> > ------------ >> > 0 >> > (1 row) >> > >> > I wonder if I am doing something wrong. I will look into this again and >> > let >> > you know soon. >> >> I did something wrong. I could reproduce the problem.... :( > > > Ok. > > The pg_trgm.sql file was not included because an update would cause problems > in previous versions. But I feel we might be able to handle this by having > special pg_trgm test files for different versions so that new trgm functions > are only tested in the newer versions of PostgreSQL. > What do you think? ISTM that this is not just a problem of only pg_trgm. pg_bigm might accidentally define a global function which other modules use. So I'm not sure whether it's really worth adding only pg_trgm regression test into pg_bigm. Regards, -- Fujii Masao