BUG #5104: tsearch parser mislead
The following bug has been logged online:
Bug reference: 5104
Logged by: Herve LEFEBVRE
Email address: herve.lefebvre@bull.net
PostgreSQL version: 8.3.8
Operating system: Linux Ubuntu 9.04
Description: tsearch parser mislead
Details:
The french opening and closing quotes ( char iso 00AB and 00BB ) are
considered by the parser as alpha chars rather than punctuation chars (and
so as space symbol).
So, for example, the text " «Bonjour!» is indexed as "«Bonjour" rather
than "bonjour"
"Herve LEFEBVRE" <herve.lefebvre@bull.net> writes:
The french opening and closing quotes ( char iso 00AB and 00BB ) are
considered by the parser as alpha chars rather than punctuation chars (and
so as space symbol).
That would suggest that you're using the wrong locale setting.
You did not say what you're using, though.
Another possibility is that the locale assumes a different encoding
than you're using in the database --- I don't remember if 8.3.x tries
to prevent a mismatch on that.
regards, tom lane