Обсуждение: Locale support broken in latest snapshots
Hello! A month ago I tested a snapshot (21 or 24 Jan), and locale/multibyte support worked. In the snapshot of 21 Feb locale support have been broken. "SELECT * FROM table ORDER BY name" returns rows in wrong order; "SELECT * FROM table WHERE name ~* re" performs case-sensitive match. I remember Tatsuo once fixed case-sensitiveness by replacing char with "unsigned char". Should I teach my gcc to always use unsigned chars? Oleg. ---- Oleg Broytmann http://members.xoom.com/phd2/ phd2@earthling.net Programmers don't die, they justGOSUB without RETURN.
Oleg Broytmann <phd@phd.russ.ru> writes:
> In the snapshot of 21 Feb locale support have been broken. "SELECT *
> FROM table ORDER BY name" returns rows in wrong order; "SELECT * FROM table
> WHERE name ~* re" performs case-sensitive match.
Ugh.
> I remember Tatsuo once fixed case-sensitiveness by replacing char with
> "unsigned char". Should I teach my gcc to always use unsigned chars?
No, that's not a solution (since it's not available to non-gcc users).
We need to find the coding error and fix it.
However, if you can try that as a test to see if it fixes the behavior,
please do; knowing whether it does will help narrow down what the bug
could be.
regards, tom lane
Oleg Bartunov gave me the advice to rerun initdb -E KOI8... And - oops - this helped. Locale test for koi8 locale passed. It is strange - I did fresh and clean install into empty directory, and of course I did initdb -E. Locale tests failed. But stopping postmaster, removing data directory, running initdb -E and restarting postmaster helped... On Mon, 28 Feb 2000, Tom Lane wrote: > Oleg Broytmann <phd@phd.russ.ru> writes: > > In the snapshot of 21 Feb locale support have been broken. "SELECT * > > FROM table ORDER BY name" returns rows in wrong order; "SELECT * FROM table > > WHERE name ~* re" performs case-sensitive match. > > Ugh. > > > I remember Tatsuo once fixed case-sensitiveness by replacing char with > > "unsigned char". Should I teach my gcc to always use unsigned chars? > > No, that's not a solution (since it's not available to non-gcc users). > We need to find the coding error and fix it. > > However, if you can try that as a test to see if it fixes the behavior, > please do; knowing whether it does will help narrow down what the bug > could be. Oleg. ---- Oleg Broytmann http://members.xoom.com/phd2/ phd2@earthling.net Programmers don't die, they justGOSUB without RETURN.