On Tue, 1 Aug 2017 08:16:54 -0400
Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
> On 8/1/17 02:12, Victor Wagner wrote:
> >> We are only calling uloc_toLanguageTag() with keyword/value
> >> combinations that ICU itself previously told us were supported. So
> >> just ignoring errors doesn't seem proper in this case.
> >>
> > We know that this version of ICU is broken. But what choice we
> > have?
>
> I don't know that we can already reach that conclusion. Maybe the
Because it was fixed in subsequent versions.
And 4.2 is first version where this function appeared.
So, we still have problems with SLES11 which use even older version and
have to be supported at least two more years.
> APIs have changed or we are not using them correctly. Are there any
> bug reports or changelog entries related to this? I don't think we
> should just give up and ignore errors.
We also can provide configuration option or command-line option for
initdb which would restrict list of languages for which collation
sequences would be created. This would help for all but two languages.