Re: Enforcing database encoding and locale match

Tom Lane wrote:
> Alvaro Herrera <alvherre@commandprompt.com> writes:
>> FWIW I tried this program here, and I get
> 
>> C                       ... ANSI_X3.4-1968 - NO MATCH
>> POSIX                   ... ANSI_X3.4-1968 - NO MATCH
> 
>> Note the funny name.  Trying initdb with LC_ALL=C correctly uses
>> SQL_ASCII (I saw the special case in chklocale.c), but I'm wondering if
>> we should list those names explicitely.
> 
> Since we're already special-casing C/POSIX, I don't see a need.
> It looks a bit hopeless to keep up with all the possibilities anyway
> --- by my count we've tested four different platforms so far and
> gotten four different answers for the CODESET name for C :-(
> 
> Linux    ANSI_X3.4-1968
> Darwin    (empty)
> Solaris    646
> HP-UX    roman8

There is some useful link:

http://www.simeji.com/bun/characterencoding_jvm142.txt

with aliases. I also checked all possible locales on Solaris nevada and 
there are two new aliases and probably unsupported TIS620.2533 (thai) 
encoding by postgres.

Patch with new aliases attached.
    Zdenek

===================================================================
RCS file: /zfs_data/cvs_pgsql/cvsroot/pgsql/src/port/chklocale.c,v
retrieving revision 1.4
diff -c -r1.4 chklocale.c
*** src/port/chklocale.c        2007/10/03 17:16:39     1.4
--- src/port/chklocale.c        2007/10/05 17:55:10
***************
*** 127,132 ****
--- 127,133 ----        {PG_WIN874, "???"},  #endif        {PG_WIN1251, "CP1251"},
+       {PG_WIN1251, "ansi-1251"},        {PG_WIN866, "CP866"},
        {PG_ISO_8859_5, "ISO-8859-5"},
***************
*** 152,157 ****
--- 153,159 ----        {PG_BIG5, "BIG5"},        {PG_BIG5, "BIG5HKSCS"},        {PG_BIG5, "CP950"},
+       {PG_BIG5, "Big5-HKSCS"},
        {PG_GBK, "GBK"},        {PG_GBK, "CP936"},