Re: [COMMITTERS] pgsql: ICU support

On 3/24/17 22:28, Tom Lane wrote: > David Rowley writes: >> I see this commit changed the definition of pg_locale_t >> but forgot to update varstr_cmp() completely. > > It's pretty disturbing that we don't seem to have caught that > in testing. Or during

Re: [COMMITTERS] pgsql: ICU support

On 3/24/17 20:32, David Rowley wrote: > but forgot to update varstr_cmp() completely. > > result = wcscoll_l((LPWSTR) a1p, (LPWSTR) a2p, mylocale); > > should be: > > result = wcscoll_l((LPWSTR) a1p, (LPWSTR) a2p, mylocale->info.lt); > > Patch attached. Fixed. -- Peter Eisentraut

Re: [COMMITTERS] pgsql: ICU support

On 3/23/17 23:53, Tom Lane wrote: > Ah, scratch that, I was overthinking it. The problem is pretty > clear from the error message: > > + ERROR: collation "en-x-icu" for encoding "SQL_ASCII" does not exist > > I can reproduce this with vanilla configure options if I'm running > with --with-icu

Re: [COMMITTERS] pgsql: ICU support

David Rowley writes: > I see this commit changed the definition of pg_locale_t > but forgot to update varstr_cmp() completely. It's pretty disturbing that we don't seem to have caught that in testing. Some work on code coverage seems indicated.

Re: [COMMITTERS] pgsql: ICU support

On 24 March 2017 at 06:33, Peter Eisentraut wrote: > http://git.postgresql.org/pg/commitdiff/eccfef81e1f73ee41f1d8bfe4fa4e80576945048 [...] > src/include/utils/pg_locale.h| 32 +- I see this commit changed the definition of pg_locale_t +struct pg_locale_t

Re: [COMMITTERS] pgsql: ICU support

On 03/24/2017 09:26 AM, Tom Lane wrote: > Andrew Dunstan writes: >> On 03/23/2017 11:53 PM, Tom Lane wrote: >>> + ERROR: collation "en-x-icu" for encoding "SQL_ASCII" does not exist >>> >>> I can reproduce this with vanilla configure options if I'm running >>>

Re: [COMMITTERS] pgsql: ICU support

Andrew Dunstan writes: > On 03/23/2017 11:53 PM, Tom Lane wrote: >> + ERROR: collation "en-x-icu" for encoding "SQL_ASCII" does not exist >> >> I can reproduce this with vanilla configure options if I'm running >> with --with-icu and LANG=C. In short, this

Re: [COMMITTERS] pgsql: ICU support

On 03/23/2017 11:53 PM, Tom Lane wrote: > I wrote: >> The buildfarm's mostly happy, but "prion" still isn't, which >> suggests there's someplace in the new code that references an >> already-closed relcache entry. > Ah, scratch that, I was overthinking it. The problem is pretty > clear from the

Re: [COMMITTERS] pgsql: ICU support

I wrote: > The buildfarm's mostly happy, but "prion" still isn't, which > suggests there's someplace in the new code that references an > already-closed relcache entry. Ah, scratch that, I was overthinking it. The problem is pretty clear from the error message: + ERROR: collation "en-x-icu"

Re: [COMMITTERS] pgsql: ICU support

The buildfarm's mostly happy, but "prion" still isn't, which suggests there's someplace in the new code that references an already-closed relcache entry. regards, tom lane -- Sent via pgsql-committers mailing list (pgsql-committers@postgresql.org) To make changes to

Re: [COMMITTERS] pgsql: ICU support

On 03/23/2017 04:07 PM, Peter Eisentraut wrote: > On 3/23/17 15:33, Peter Eisentraut wrote: >> ICU support > Seeing some crashes on the build farm, investigating. If someone can > reproduce locally and get a backtrace, let me know. > The buildfarm actually does that :-) See for example

Re: [COMMITTERS] pgsql: ICU support

On 3/23/17 15:33, Peter Eisentraut wrote: > ICU support Seeing some crashes on the build farm, investigating. If someone can reproduce locally and get a backtrace, let me know. -- Peter Eisentraut http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA,

[COMMITTERS] pgsql: ICU support

ICU support Add a column collprovider to pg_collation that determines which library provides the collation data. The existing choices are default and libc, and this adds an icu choice, which uses the ICU4C library. The pg_locale_t type is changed to a union that contains the provider-specific