Fujii Masao
masao****@gmail*****
2013年 10月 23日 (水) 01:53:36 JST
On Tue, Oct 22, 2013 at 8:47 PM, Beena Emerson <memis****@gmail*****> wrote: > > On Tue, Oct 22, 2013 at 9:32 AM, Fujii Masao <masao****@gmail*****> wrote: >> >> On Tue, Oct 22, 2013 at 12:54 PM, Beena Emerson <memis****@gmail*****> >> wrote: >> > On Mon, Oct 21, 2013 at 11:52 PM, Fujii Masao <masao****@gmail*****> >> > wrote: >> >> >> >> >> >> +<li><p>Remove USE_WIDE_UPPER_LOWER dependency (Beena Emerson)</p></li> >> >> >> >> I'm afraid that it's not easy for a user to understand what this change >> >> is. >> >> Could you change this description so that a user can more easily >> >> understand? >> >> Since this change came from the change of pg_trgm, I think that you can >> >> consult the release note of PostgreSQL about that change of pg_trgm. > > > There is no mention of this change in the release note. > > I think we can add the following as the explanation: > Previously, if USE_WIDE_UPPER_LOWER was not defined, the multibyte character > boundaries are ignored and bigrams was made from bytes instead. Since > formation of bigrams does not have to depend on presence of to_lower and > friends, it has been removed. > >> >> I'm wondering if users may need to rebuild the index when upgrading >> pg_bigm >> from version 1.0 to 1.1 if its platform that pg_bigm is running depends on >> USE_WIDE_UPPER_LOWER. That is, in that platform, removing that flag may >> mean the change of index key layout. > > > As the commit information on pg_trgm states: > "If there is anybody still using Postgres on machines without wcstombs() or > towlower(), and they have non-ASCII data indexed by pg_trgm, they'll need > to REINDEX those indexes after pg_upgrade to 9.3, else searches may fail > incorrectly. It seems likely that there are no such installations, though." > > So I feel, even in bigm this would be the case and we might have to add > similar information in the release note. I found the following in the release note of PostgreSQL9.3. Isn't this related to the change log in question? ------------------------ Improve pg_trgm's handling of multibyte characters (Tom Lane) On a platform that does not have the wcstombs() or towlower() library functions, this could result in an incompatible change in the contents of pg_trgm indexes for non-ASCII data. In such cases, REINDEX those indexes to ensure correct search results. ------------------------ Regards, -- Fujii Masao
Fork