Thread: The default text search configuration will be set to "simple" ?

The default text search configuration will be set to "simple" ?

From
Stefan Schwarzer
Date:
Hi there,

what does this message mean? I didn't find any information on the web,
beside of others having similar messages...

      ------------------------------
The database cluster will be initialized with locales
   COLLATE:  en_US.UTF-8
   CTYPE:    UTF-8
   MESSAGES: en_US.UTF-8
   MONETARY: en_US.UTF-8
   NUMERIC:  en_US.UTF-8
   TIME:     en_US.UTF-8
The default database encoding has accordingly been set to UTF8.
initdb: could not find suitable text search configuration for locale
UTF-8
The default text search configuration will be set to "simple".
      ------------------------------

Thanks for any advice,

Stef

Re: The default text search configuration will be set to "simple" ?

From
Richard Huxton
Date:
Stefan Schwarzer wrote:
> Hi there,
>
> what does this message mean? I didn't find any information on the web,
> beside of others having similar messages...
>
>      ------------------------------
> The database cluster will be initialized with locales
>   COLLATE:  en_US.UTF-8
>   CTYPE:    UTF-8

This one is missing the en_US off the front. UTF-8 isn't a locale, it's
an encoding.

> The default database encoding has accordingly been set to UTF8.

Good.

> initdb: could not find suitable text search configuration for locale UTF-8

Bad. For some reason initdb isn't picking up your CTYPE setting
properly. What operating-system (and version) are you running, and what
version of PG, and did you compile it from source or get it elsewhere?

--
   Richard Huxton
   Archonet Ltd

Re: The default text search configuration will be set to "simple" ?

From
Stefan Schwarzer
Date:
what does this message mean? I didn't find any information on the web, beside of others having similar messages...
    ------------------------------
The database cluster will be initialized with locales
 COLLATE:  en_US.UTF-8
 CTYPE:    UTF-8

This one is missing the en_US off the front. UTF-8 isn't a locale, it's an encoding.

The default database encoding has accordingly been set to UTF8.

Good.

initdb: could not find suitable text search configuration for locale UTF-8

Bad. For some reason initdb isn't picking up your CTYPE setting properly. What operating-system (and version) are you running, and what version of PG, and did you compile it from source or get it elsewhere?

I am running 10.5, on a MacPro. Postgres version 8.3.1 and got it from macports. 

A bit ago, in my many attempts to get postgres & postgis running on 10.5, I added this to the .bash_profile:

adding to .bash_profile:
LANG=en_US.UTF-8; export LANG
export PATH="/opt/local/lib/postgresql83/bin:$PATH"

And then, something which didn't work before, did work finally. But I am not eager to try it out now, because I am happy enough that my postgres&postgis is finally running...

Stef

Re: The default text search configuration will be set to "simple" ?

From
Richard Huxton
Date:
Stefan Schwarzer wrote:
>> properly. What operating-system (and version) are you running, and
>> what version of PG, and did you compile it from source or get it
>> elsewhere?
>
> I am running 10.5, on a MacPro. Postgres version 8.3.1 and got it from
> macports.
>
> A bit ago, in my many attempts to get postgres & postgis running on
> 10.5, I added this to the .bash_profile:
>
> adding to .bash_profile:
> LANG=en_US.UTF-8; export LANG
> export PATH="/opt/local/lib/postgresql83/bin:$PATH"

Hmm - very strange. Your LANG setting seems sensible, and I did
something similar on my Mac to get subversion to work. It could be that
the postgres user-account isn't importing these settings, but it
certainly seemed to get time/money/etc correct.

> And then, something which didn't work before, did work finally. But I am
> not eager to try it out now, because I am happy enough that my
> postgres&postgis is finally running...

If it's nothing to do with the language settings, I'd check you have the
relevant files for tsearch. Look in $SHAREDIR/tsearch_data for a bunch
of <language>.stop files and perhaps .dict or .affix files

You can find out where your sharedir is with "pg_config --sharedir".

--
   Richard Huxton
   Archonet Ltd

Re: The default text search configuration will be set to "simple" ?

From
Alvaro Herrera
Date:
Richard Huxton wrote:
> Stefan Schwarzer wrote:
>>> properly. What operating-system (and version) are you running, and
>>> what version of PG, and did you compile it from source or get it
>>> elsewhere?
>>
>> I am running 10.5, on a MacPro. Postgres version 8.3.1 and got it from
>> macports.
>>
>> A bit ago, in my many attempts to get postgres & postgis running on
>> 10.5, I added this to the .bash_profile:
>>
>> adding to .bash_profile:
>> LANG=en_US.UTF-8; export LANG
>> export PATH="/opt/local/lib/postgresql83/bin:$PATH"
>
> Hmm - very strange. Your LANG setting seems sensible, and I did
> something similar on my Mac to get subversion to work. It could be that
> the postgres user-account isn't importing these settings, but it
> certainly seemed to get time/money/etc correct.

It does make certain sense -- after all, LC_CTYPE is not worried about
country or language issues but only about encoding specifics.  My
locale(7) manpage says

       LC_CTYPE
              This  changes the behavior of the character handling and classifi-
              cation functions, such  as  isupper(3)  and  toupper(3),  and  the
              multi-byte character functions such as mblen(3) or wctomb(3).

On my system, though, locale does report the full locale, not just
encoding, when I change LANG.

--
Alvaro Herrera                                http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

Re: The default text search configuration will be set to "simple" ?

From
Tom Lane
Date:
Alvaro Herrera <alvherre@commandprompt.com> writes:
>> Stefan Schwarzer wrote:
>>> I am running 10.5, on a MacPro. Postgres version 8.3.1 and got it from
>>> macports.

> On my system, though, locale does report the full locale, not just
> encoding, when I change LANG.

Mine too.  On my 10.5.2 machine initdb works as expected when I set
LANG to en_US.UTF-8.  The only way I can reproduce the failure is to
*also* set LC_CTYPE to UTF-8 (without any en_US).  That strikes me
as a configuration error.  PG is looking at LC_CTYPE to decide what
text search config to select, and of course this value tells it
nothing about which language is in use ...

            regards, tom lane