The following bug has been logged online:
Bug reference: 5532
Logged by: Michael Lewis
Email address: mikelikespie@gmail.com
PostgreSQL version: 9.0 trunk
Operating system: OS X
Description: Valid UTF8 sequence errors as invalid
Details:
I'm using Python to sanitize my logs from invalid UTF8 characters before
COPYing them into postgres. I came across this one sequence that seems to
be valid UTF8 (in the extended range I believe).
It goes through both pythons encoding as well as iconv without an error and
is valid as far as my understanding of UTF8 goes so I am assuming it is a
bug.
Test case:
create table t (v varchar);
insert into t values (E'\xed\xbc\xad');
In bash you can do:
echo -e "\xed\xbc\xad" | iconv -f UTF-8 ; echo $?
to validate it
Thanks,
Mike