>> > I can agree, so current behave can be useful in some cases, but still it >> is >> > bug (inconsistency) between PostgreSQL parser and PostgreSQL escaping >> > functions. >> > >> > Currently, any multibyte char can be unescaped identifier (only >> apostrophes >> > are tested). We should to test white chars too. >> >> Really? I thought we do that test. >> > > what you are expecting from this test? UTF single quotes are tested only in > quote functions probably.
I just wanted to demonstrate multibyte chars including ASCII white spaces can be an identifier.
I understand now.
> We should to test white chars too.
What do you exactly propose regarding white chars and multibyte chars here? Maybe you propose to consider non ASCII white spaces (treate them as ASCII white spaces)?
I propose the work with UTF white chars should be same like ASCII white chars. The current design is too simple - with possible pretty bad issues. Daniel's example is good - there is big gap in design.
