The regression tests are known to fail on non-SQL_ASCII encodings for
this and similar reasons.
cheers
andrew
Teodor Sigaev wrote:
> Working on UTF support in tsearch2 I noticed different output of psql:
>
> 1) initdb with SQL_ASCII:
> contrib_regression=# select 'jisa n asd n asld kl ans d
> contrib_regression'# asdfjkl; asdf kl;amsdf';
> ?column?
> -----------------------------------------------------------
> jisa n asd n asld kl ans d
> asdfjkl; asdf kl;amsdf
> (1 row)
>
> 2) initdb with UTF8
> contrib_regression=# select 'jisa n asd n asld kl ans d
> contrib_regression'# asdfjkl; asdf kl;amsdf';
> ?column?
> ------------------------------------
> jisa n asd n asld kl ans d
> asdfjkl; asdf kl;amsdf
> (1 row)
>
>
> Length of '--------' string is different. This difference is a reason
> for installcheck failture of tsearch2 on UTF8 inited db.
>
> Is it supposed behaviour of psql?
>
>
