On 09/19/2017 02:47 PM, Andrew Dunstan wrote: > > On 09/19/2017 11:11 AM, Tom Lane wrote: >> Andrew Dunstan <andrew.duns...@2ndquadrant.com> writes: >>> This seems to have upset a number or animals in the buildfarm. >> Actually, after looking closer, my advice is just to drop the new >> test cases involving accented letters. It surprises me not in the >> least that those would have nonportable behavior: probably, some >> locales will case-fold them and some won't. >> >> (This does open up some questions about whether the opclass will >> ever be of use for Alexey's original purpose :-() > Actually, it now looks to me like the problem is we forgot to tell > postgres that this data is in utf8. The problem appears to resolve (at > least on my CentOS test box, where I reproduced the buildfarm error) if > I add > > set client_encoding = 'utf8'; > > to the sql file. > > Of course UTF8 bytes interpreted as LATIN1 will give results that are > ... interesting. > > So let's try that first and see if it make the buildfarm go green. Maybe > there's hope for Alexey's purpose after all.
*sigh* That worked in a couple of instances but crashed and burned elsewhere. I'll just disable the multi-byte tests as Tom suggests. cheers andrew -- Andrew Dunstan https://www.2ndQuadrant.com PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers