On 05.03.22 09:38, Julien Rouhaud wrote:
I say it works because I did manually check, as far as I can see there isn't
any test that ensures it.

I'm using this naive scenario:

DROP DATABASE IF EXISTS dbicu;
CREATE DATABASE dbicu LOCALE_PROVIDER icu LOCALE 'en_US' ICU_LOCALE 
'en-u-kf-upper' template 'template0';
\c dbicu
CREATE COLLATION upperfirst (provider = icu, locale = 'en-u-kf-upper');
CREATE TABLE icu(def text, en text COLLATE "en_US", upfirst text COLLATE 
upperfirst);
INSERT INTO icu VALUES ('a', 'a', 'a'), ('b','b','b'), ('A','A','A'), 
('B','B','B');
SELECT def AS def FROM icu ORDER BY def;
SELECT def AS en FROM icu ORDER BY en;
SELECT def AS upfirst FROM icu ORDER BY upfirst;
SELECT def AS upfirst_explicit FROM icu ORDER BY en COLLATE upperfirst;
SELECT def AS en_x_explicit FROM icu ORDER BY def COLLATE "en-x-icu";

Maybe there should be some test along those lines included in the patch?

I added something like this to a new test suite under src/test/icu/. (src/test/locale/ was already used for something else.)

Also, it's not even used consistently in the patch.  I can see e.g. "ICU
locale" or "ICU locale setting" being used:

I have fixed those.

Maybe we could point to the ICU documentation for a clear notion of what a
locale ID is, and/or use their own short definition [1]:

The locale object in ICU is an identifier that specifies a particular locale
and has fields for language, country, and an optional code to specify further
variants or subdivisions. These fields also can be represented as a string
with the fields separated by an underscore.

I think the Localization chapter needs to be reorganized a bit, but I'll leave that for a separate patch.

I spent some time looking at the ICU api trying to figure out if using a
posix locale name (e.g. en_US) was actually compatible with an ICU locale name.
It seems that ICU accept any of 'en-us', 'en-US', 'en_us' or 'en_US' as the
same locale, but I might be wrong.  I also didn't find a way to figure out how
to ask ICU if the locale identifier passed is complete garbage or not.  One
sure thing is that the system collation we import are of the form 'en-us', so
it seems weird to have this form in pg_collation and by default another form in
pg_database.

Yeah it seems to be inconsistent about that. The locale ID documentation appears to indicate that "en_US" is the canonical form, but when you ask it to list all the locales it knows about it returns "en-US".

In CREATE DATABASE manual:

+        Specifies the provider to use for the default collation in this
+        database.  Possible values are:
+        <literal>icu</literal>,<indexterm><primary>ICU</primary></indexterm>
+        <literal>libc</literal>.  <literal>libc</literal> is the default.  The
+        available choices depend on the operating system and build options.

That's actually not true as pg_strcasecmp is used in createdb():

+       if (pg_strcasecmp(locproviderstr, "icu") == 0)
+           dblocprovider = COLLPROVIDER_ICU;
+       else if (pg_strcasecmp(locproviderstr, "libc") == 0)
+           dblocprovider = COLLPROVIDER_LIBC;
+       else
+           ereport(ERROR,
+                   (errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+                    errmsg("unrecognized locale provider: %s",
+                           locproviderstr)));

I don't understand what the concern is here.

Unless I'm missing something you entirely removed the warninng about the
performance penalty when using non C/POSIX locale?

Yeah, I think it is not the job of the initdb man page to lecture people about the merits of their locale choice. The same performance warning can also be found in the localization chapter; we don't need to repeat it everywhere a locale choice is mentioned.

initdb has some specific processing to transform the default libc locale to
something more appropriate, but as far as I can see creatdb / CREATE DATABASE
aren't doing that.  It seems inconsistent, and IMHO another reason why
defaulting to the libc locale looks like a bad idea.

This has all been removed. The separate ICU locale option should now be required everywhere (initdb, createdb, CREATE DATABASE).

While on that topic, the doc should probably mention that default ICU
collations can only be deterministic.

Well, there is no option to do otherwise, so I'm not sure where/how to mention that. We usually don't document options that don't exist. ;-)

Also unrelated, but how about raising a warning about possibly hiding
corruption if you use CREATE COLLATION ...  VERSION or CREATE DATABASE ...
COLLATION_VERSION if !IsBinaryUpgrade()?

Hmm, there is a point to that. But then we should just make that an error. Otherwise, we raise the warning but then there is no way to "fix" what was warned about. Seems confusing.

in AlterCollation(), pg_collation_actual_version(), AlterDatabaseRefreshColl()
and pg_database_collation_actual_version():

-   datum = SysCacheGetAttr(COLLOID, tup, Anum_pg_collation_collcollate, 
&isnull);
-   Assert(!isnull);
-   newversion = get_collation_actual_version(collForm->collprovider, 
TextDatumGetCString(datum));
+   datum = SysCacheGetAttr(COLLOID, tup, collForm->collprovider == 
COLLPROVIDER_ICU ? Anum_pg_collation_colliculocale : Anum_pg_collation_collcollate, 
&isnull);
+   if (!isnull)
+       newversion = get_collation_actual_version(collForm->collprovider, 
TextDatumGetCString(datum));
+   else
+       newversion = NULL;

The columns are now nullable, but can you actually end up with a null locale
name in the expected field without manual DML on the catalog, corruption or
similar?  I think it should be a plain error explaining the inconsistency
rather then silently assuming there's no version.  Note that at least
pg_newlocale_from_collation() asserts that the specific libc/icu field it's
interested in isn't null.

This is required because the default collations's fields are null now.

Why not using errorConflictingDefElem for locale_provider?

fixed

in createdb():

+   if (dblocprovider == COLLPROVIDER_ICU)
+   {
+       /*
+        * This would happen if template0 uses the libc provider but the new
+        * database uses icu.
+        */
+       if (!dbiculocale)
+           ereport(ERROR,
+                   (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+                    errmsg("ICU locale must be specified")));
+   }
+
+   if (dblocprovider == COLLPROVIDER_ICU)
+   {
+#ifdef USE_ICU
[...]

Seems like a refactoring leftover, no need for two blocks.

These are two independent blocks in my mind. It's possible that someone might want to insert something in the middle at some point.

Also, I think it
would be better to first error out if there's no support for ICU rather than
complaining about a possibly missing locale that wouldn't be accepted anyway.

Seems better to do general syntax checks first and then deeper checks of the passed values.

+void
+make_icu_collator(const char *iculocstr,
+                 struct pg_locale_struct *resultp)
+{
+#ifdef USE_ICU
+[...]
+   /* We will leak this string if we get an error below :-( */
+   resultp->info.icu.locale = MemoryContextStrdup(TopMemoryContext, iculocstr);
+   resultp->info.icu.ucol = collator;
+#else                          /* not USE_ICU */
+   /* could get here if a collation was created by a build with ICU */
+   ereport(ERROR,
+           (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+            errmsg("ICU is not supported in this build"), \
+            errhint("You need to rebuild PostgreSQL using %s.", 
"--with-icu")));
+#endif                         /* not USE_ICU */
+}

The comment about the leak possibility needs some tweaking since it's been
extracted from a larger function.

fixed

Not material for this patch, but:
@@ -1475,7 +1524,12 @@ pg_newlocale_from_collation(Oid collid)
     Assert(OidIsValid(collid));

     if (collid == DEFAULT_COLLATION_OID)
-       return (pg_locale_t) 0;
+   {
+       if (default_locale.provider == COLLPROVIDER_ICU)
+           return &default_locale;
+       else
+           return (pg_locale_t) 0;
+   }

I'm wondering if we could now always return &default_locale and avoid having to
check if the function returned something in all callers, since CheckMyDatabase
now initialize it.

Maybe that's something to look into, but that would probably require updating a call callers to handle the return values differently, which would require quite a bit of work.

Shouldn't all the code before checking the locale name also be in the #ifdef
USE_ICU?

I like to keep the #ifdef sections as small as possible and limited to the code that really uses the respective library.

@@ -2859,6 +2870,17 @@ dumpDatabase(Archive *fout)
         appendPQExpBufferStr(creaQry, " ENCODING = ");
         appendStringLiteralAH(creaQry, encoding, fout);
     }
+   if (strlen(datlocprovider) > 0)
+   {
+       appendPQExpBufferStr(creaQry, " LOCALE_PROVIDER = ");
+       if (datlocprovider[0] == 'c')
+           appendPQExpBufferStr(creaQry, "libc");
+       else if (datlocprovider[0] == 'i')
+           appendPQExpBufferStr(creaQry, "icu");
+       else
+           fatal("unrecognized locale provider: %s",
+                 datlocprovider);
+   }
     if (strlen(collate) > 0 && strcmp(collate, ctype) == 0)
     {

AFAICS datlocprovider shouldn't be empty.  Maybe raise an error or an assert if
that's the case?

Yeah that was bogus, copied from the earlier encoding handling, which is probably also bogus, but I'm not touching that here.

# CREATE DATABASE db1 LOCALE_PROVIDER icu ICU_LOCALE "fr-x-icu";
ERROR:  22023: new locale provider (i) does not match locale provider of the 
template database (c)
HINT:  Use the same locale provider as in the template database, or use 
template0 as template.

It feels strange to write "LOCALE_PROVIDER icu" and get "provider (i)" in the
error message.  I think it would be better to emit the user-facing value, not
internal one.

Fixed.  I had added a collprovider_name() function but didn't use it here.

Finally, there are some changes in pg_collation, I'm assuming it's done for
consistency since pg_database may need two different locale information, but it
should probably be at least mentioned in the commit message.

done
From 7d9703ae72a0b82168eaf8be3d8274e1101df71a Mon Sep 17 00:00:00 2001
From: Peter Eisentraut <pe...@eisentraut.org>
Date: Mon, 14 Mar 2022 13:46:15 +0100
Subject: [PATCH v6] Add option to use ICU as global locale provider

This adds the option to use ICU as the default locale provider for
either the whole cluster or a database.  New options for initdb,
createdb, and CREATE DATABASE are used to select this.

Since some (legacy) code still uses the libc locale facilities
directly, we still need to set the libc global locale settings even if
ICU is otherwise selected.  So pg_database now has three
locale-related fields: the existing datcollate and datctype, which are
always set, and a new daticulocale, which is only set if ICU is
selected.  A similar change is made in pg_collation for consistency,
but in that case, only the libc-related fields or the ICU-related
field is set, never both.

Discussion: 
https://www.postgresql.org/message-id/flat/5e756dd6-0e91-d778-96fd-b1bcb06c161a%402ndquadrant.com

XXX catversion bump
---
 doc/src/sgml/catalogs.sgml                    |   9 +
 doc/src/sgml/ref/create_database.sgml         |  25 +++
 doc/src/sgml/ref/createdb.sgml                |  19 +++
 doc/src/sgml/ref/initdb.sgml                  |  72 ++++++--
 src/backend/catalog/pg_collation.c            |  18 +-
 src/backend/commands/collationcmds.c          |  96 +++++++----
 src/backend/commands/dbcommands.c             | 159 +++++++++++++++---
 src/backend/utils/adt/pg_locale.c             | 144 ++++++++++------
 src/backend/utils/init/postinit.c             |  21 ++-
 src/bin/initdb/Makefile                       |   4 +-
 src/bin/initdb/initdb.c                       |  97 +++++++++--
 src/bin/initdb/t/001_initdb.pl                |  27 +++
 src/bin/pg_dump/pg_dump.c                     |  30 +++-
 src/bin/pg_upgrade/check.c                    |  13 ++
 src/bin/pg_upgrade/info.c                     |  18 +-
 src/bin/pg_upgrade/pg_upgrade.h               |   2 +
 src/bin/psql/describe.c                       |  23 ++-
 src/bin/psql/tab-complete.c                   |   3 +-
 src/bin/scripts/Makefile                      |   2 +
 src/bin/scripts/createdb.c                    |  20 +++
 src/bin/scripts/t/020_createdb.pl             |  28 +++
 src/include/catalog/pg_collation.dat          |   3 +-
 src/include/catalog/pg_collation.h            |  20 ++-
 src/include/catalog/pg_database.dat           |   4 +-
 src/include/catalog/pg_database.h             |   6 +
 src/include/utils/pg_locale.h                 |   5 +
 src/test/Makefile                             |   6 +-
 src/test/icu/.gitignore                       |   2 +
 src/test/icu/Makefile                         |  25 +++
 src/test/icu/README                           |  27 +++
 src/test/icu/t/010_database.pl                |  58 +++++++
 .../regress/expected/collate.icu.utf8.out     |  10 +-
 src/test/regress/sql/collate.icu.utf8.sql     |   8 +-
 33 files changed, 837 insertions(+), 167 deletions(-)
 create mode 100644 src/test/icu/.gitignore
 create mode 100644 src/test/icu/Makefile
 create mode 100644 src/test/icu/README
 create mode 100644 src/test/icu/t/010_database.pl

diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 7777d60514..bcf2b43206 100644
--- a/doc/src/sgml/catalogs.sgml
+++ b/doc/src/sgml/catalogs.sgml
@@ -2384,6 +2384,15 @@ <title><structname>pg_collation</structname> 
Columns</title>
       </para></entry>
      </row>
 
+     <row>
+      <entry role="catalog_table_entry"><para role="column_definition">
+       <structfield>colliculocale</structfield> <type>text</type>
+      </para>
+      <para>
+       ICU locale ID for this collation object
+      </para></entry>
+     </row>
+
      <row>
       <entry role="catalog_table_entry"><para role="column_definition">
        <structfield>collversion</structfield> <type>text</type>
diff --git a/doc/src/sgml/ref/create_database.sgml 
b/doc/src/sgml/ref/create_database.sgml
index f70d0c75b4..33e2d7ce83 100644
--- a/doc/src/sgml/ref/create_database.sgml
+++ b/doc/src/sgml/ref/create_database.sgml
@@ -28,6 +28,8 @@
            [ LOCALE [=] <replaceable class="parameter">locale</replaceable> ]
            [ LC_COLLATE [=] <replaceable 
class="parameter">lc_collate</replaceable> ]
            [ LC_CTYPE [=] <replaceable 
class="parameter">lc_ctype</replaceable> ]
+           [ ICU_LOCALE [=] <replaceable 
class="parameter">icu_locale</replaceable> ]
+           [ LOCALE_PROVIDER [=] <replaceable 
class="parameter">locale_provider</replaceable> ]
            [ COLLATION_VERSION = <replaceable>collation_version</replaceable> ]
            [ TABLESPACE [=] <replaceable 
class="parameter">tablespace_name</replaceable> ]
            [ ALLOW_CONNECTIONS [=] <replaceable 
class="parameter">allowconn</replaceable> ]
@@ -160,6 +162,29 @@ <title>Parameters</title>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><replaceable class="parameter">icu_locale</replaceable></term>
+      <listitem>
+       <para>
+        Specifies the ICU locale ID if the ICU locale provider is used.
+       </para>
+      </listitem>
+     </varlistentry>
+
+     <varlistentry>
+      <term><replaceable>locale_provider</replaceable></term>
+
+      <listitem>
+       <para>
+        Specifies the provider to use for the default collation in this
+        database.  Possible values are:
+        <literal>icu</literal>,<indexterm><primary>ICU</primary></indexterm>
+        <literal>libc</literal>.  <literal>libc</literal> is the default.  The
+        available choices depend on the operating system and build options.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><replaceable>collation_version</replaceable></term>
 
diff --git a/doc/src/sgml/ref/createdb.sgml b/doc/src/sgml/ref/createdb.sgml
index 86473455c9..be42e502d6 100644
--- a/doc/src/sgml/ref/createdb.sgml
+++ b/doc/src/sgml/ref/createdb.sgml
@@ -147,6 +147,25 @@ <title>Options</title>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--icu-locale=<replaceable 
class="parameter">locale</replaceable></option></term>
+      <listitem>
+       <para>
+        Specifies the ICU locale ID to be used in this database, if the
+        ICU locale provider is selected.
+       </para>
+      </listitem>
+     </varlistentry>
+
+     <varlistentry>
+      
<term><option>--locale-provider={<literal>libc</literal>|<literal>icu</literal>}</option></term>
+      <listitem>
+       <para>
+        Specifies the locale provider for the database's default collation.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>-O <replaceable 
class="parameter">owner</replaceable></option></term>
       <term><option>--owner=<replaceable 
class="parameter">owner</replaceable></option></term>
diff --git a/doc/src/sgml/ref/initdb.sgml b/doc/src/sgml/ref/initdb.sgml
index 8f71c7c962..f5d633b0af 100644
--- a/doc/src/sgml/ref/initdb.sgml
+++ b/doc/src/sgml/ref/initdb.sgml
@@ -86,30 +86,45 @@ <title>Description</title>
   </para>
 
   <para>
-   <command>initdb</command> initializes the database cluster's default
-   locale and character set encoding. The character set encoding,
-   collation order (<literal>LC_COLLATE</literal>) and character set classes
-   (<literal>LC_CTYPE</literal>, e.g., upper, lower, digit) can be set 
separately
-   for a database when it is created. <command>initdb</command> determines
-   those settings for the template databases, which will
-   serve as the default for all other databases.
+   <command>initdb</command> initializes the database cluster's default locale
+   and character set encoding. These can also be set separately for each
+   database when it is created. <command>initdb</command> determines those
+   settings for the template databases, which will serve as the default for
+   all other databases.  By default, <command>initdb</command> uses the
+   locale provider <literal>libc</literal>, takes the locale settings from
+   the environment, and determines the encoding from the locale settings.
+   This is almost always sufficient, unless there are special requirements.
   </para>
 
   <para>
-   To alter the default collation order or character set classes, use the
-   <option>--lc-collate</option> and <option>--lc-ctype</option> options.
-   Collation orders other than <literal>C</literal> or 
<literal>POSIX</literal> also have
-   a performance penalty.  For these reasons it is important to choose the
-   right locale when running <command>initdb</command>.
+   To choose a different locale for the cluster, use the option
+   <option>--locale</option>.  There are also individual options
+   <option>--lc-*</option> (see below) to set values for the individual locale
+   categories.  Note that inconsistent settings for different locale
+   categories can give nonsensical results, so this should be used with care.
   </para>
 
   <para>
-   The remaining locale categories can be changed later when the server
-   is started.  You can also use <option>--locale</option> to set the
-   default for all locale categories, including collation order and
-   character set classes. All server locale values (<literal>lc_*</literal>) 
can
-   be displayed via <command>SHOW ALL</command>.
-   More details can be found in <xref linkend="locale"/>.
+   Alternatively, the ICU library can be used to provide locale services.
+   (Again, this only sets the default for subsequently created databases.)  To
+   select this option, specify <literal>--locale-provider=icu</literal>.
+   To chose the specific ICU locale ID to apply, use the option
+   <option>--icu-locale</option>.  Note that
+   for implementation reasons and to support legacy code,
+   <command>initdb</command> will still select and initialize libc locale
+   settings when the ICU locale provider is used.
+  </para>
+
+  <para>
+   When <command>initdb</command> runs, it will print out the locale settings
+   it has chosen.  If you have complex requirements or specified multiple
+   options, it is advisable to check that the result matches what was
+   intended.
+  </para>
+
+  <para>
+   More details about locale settings can be found in <xref
+   linkend="locale"/>.
   </para>
 
   <para>
@@ -210,6 +225,15 @@ <title>Options</title>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      
<term><option>--icu-locale=<replaceable>locale</replaceable></option></term>
+      <listitem>
+       <para>
+        Specifies the ICU locale ID, if the ICU locale provider is used.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry id="app-initdb-data-checksums" xreflabel="data checksums">
       <term><option>-k</option></term>
       <term><option>--data-checksums</option></term>
@@ -264,6 +288,18 @@ <title>Options</title>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      
<term><option>--locale-provider={<literal>libc</literal>|<literal>icu</literal>}</option></term>
+      <listitem>
+       <para>
+        This option sets the locale provider for databases created in the
+        new cluster.  It can be overridden in the <command>CREATE
+        DATABASE</command> command when new databases are subsequently
+        created.  The default is <literal>libc</literal>.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>-N</option></term>
       <term><option>--no-sync</option></term>
diff --git a/src/backend/catalog/pg_collation.c 
b/src/backend/catalog/pg_collation.c
index bfc02d3038..93545786df 100644
--- a/src/backend/catalog/pg_collation.c
+++ b/src/backend/catalog/pg_collation.c
@@ -49,6 +49,7 @@ CollationCreate(const char *collname, Oid collnamespace,
                                bool collisdeterministic,
                                int32 collencoding,
                                const char *collcollate, const char *collctype,
+                               const char *colliculocale,
                                const char *collversion,
                                bool if_not_exists,
                                bool quiet)
@@ -66,8 +67,7 @@ CollationCreate(const char *collname, Oid collnamespace,
        AssertArg(collname);
        AssertArg(collnamespace);
        AssertArg(collowner);
-       AssertArg(collcollate);
-       AssertArg(collctype);
+       AssertArg((collcollate && collctype) || colliculocale);
 
        /*
         * Make sure there is no existing collation of same name & encoding.
@@ -161,8 +161,18 @@ CollationCreate(const char *collname, Oid collnamespace,
        values[Anum_pg_collation_collprovider - 1] = CharGetDatum(collprovider);
        values[Anum_pg_collation_collisdeterministic - 1] = 
BoolGetDatum(collisdeterministic);
        values[Anum_pg_collation_collencoding - 1] = 
Int32GetDatum(collencoding);
-       values[Anum_pg_collation_collcollate - 1] = 
CStringGetTextDatum(collcollate);
-       values[Anum_pg_collation_collctype - 1] = 
CStringGetTextDatum(collctype);
+       if (collcollate)
+               values[Anum_pg_collation_collcollate - 1] = 
CStringGetTextDatum(collcollate);
+       else
+               nulls[Anum_pg_collation_collcollate - 1] = true;
+       if (collctype)
+               values[Anum_pg_collation_collctype - 1] = 
CStringGetTextDatum(collctype);
+       else
+               nulls[Anum_pg_collation_collctype - 1] = true;
+       if (colliculocale)
+               values[Anum_pg_collation_colliculocale - 1] = 
CStringGetTextDatum(colliculocale);
+       else
+               nulls[Anum_pg_collation_colliculocale - 1] = true;
        if (collversion)
                values[Anum_pg_collation_collversion - 1] = 
CStringGetTextDatum(collversion);
        else
diff --git a/src/backend/commands/collationcmds.c 
b/src/backend/commands/collationcmds.c
index 93df1d366c..b86684a833 100644
--- a/src/backend/commands/collationcmds.c
+++ b/src/backend/commands/collationcmds.c
@@ -65,6 +65,7 @@ DefineCollation(ParseState *pstate, List *names, List 
*parameters, bool if_not_e
        DefElem    *versionEl = NULL;
        char       *collcollate;
        char       *collctype;
+       char       *colliculocale;
        bool            collisdeterministic;
        int                     collencoding;
        char            collprovider;
@@ -152,6 +153,12 @@ DefineCollation(ParseState *pstate, List *names, List 
*parameters, bool if_not_e
                else
                        collctype = NULL;
 
+               datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_colliculocale, &isnull);
+               if (!isnull)
+                       colliculocale = TextDatumGetCString(datum);
+               else
+                       colliculocale = NULL;
+
                ReleaseSysCache(tp);
 
                /*
@@ -172,18 +179,7 @@ DefineCollation(ParseState *pstate, List *names, List 
*parameters, bool if_not_e
 
                collcollate = NULL;
                collctype = NULL;
-
-               if (localeEl)
-               {
-                       collcollate = defGetString(localeEl);
-                       collctype = defGetString(localeEl);
-               }
-
-               if (lccollateEl)
-                       collcollate = defGetString(lccollateEl);
-
-               if (lcctypeEl)
-                       collctype = defGetString(lcctypeEl);
+               colliculocale = NULL;
 
                if (providerEl)
                        collproviderstr = defGetString(providerEl);
@@ -211,15 +207,42 @@ DefineCollation(ParseState *pstate, List *names, List 
*parameters, bool if_not_e
                else
                        collprovider = COLLPROVIDER_LIBC;
 
-               if (!collcollate)
-                       ereport(ERROR,
-                                       
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
-                                        errmsg("parameter \"lc_collate\" must 
be specified")));
+               if (localeEl)
+               {
+                       if (collprovider == COLLPROVIDER_LIBC)
+                       {
+                               collcollate = defGetString(localeEl);
+                               collctype = defGetString(localeEl);
+                       }
+                       else
+                               colliculocale = defGetString(localeEl);
+               }
 
-               if (!collctype)
-                       ereport(ERROR,
-                                       
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
-                                        errmsg("parameter \"lc_ctype\" must be 
specified")));
+               if (lccollateEl)
+                       collcollate = defGetString(lccollateEl);
+
+               if (lcctypeEl)
+                       collctype = defGetString(lcctypeEl);
+
+               if (collprovider == COLLPROVIDER_LIBC)
+               {
+                       if (!collcollate)
+                               ereport(ERROR,
+                                               
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+                                                errmsg("parameter 
\"lc_collate\" must be specified")));
+
+                       if (!collctype)
+                               ereport(ERROR,
+                                               
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+                                                errmsg("parameter \"lc_ctype\" 
must be specified")));
+               }
+               else if (collprovider == COLLPROVIDER_ICU)
+               {
+                       if (!colliculocale)
+                               ereport(ERROR,
+                                               
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+                                                errmsg("parameter \"locale\" 
must be specified")));
+               }
 
                /*
                 * Nondeterministic collations are currently only supported 
with ICU
@@ -260,7 +283,7 @@ DefineCollation(ParseState *pstate, List *names, List 
*parameters, bool if_not_e
        }
 
        if (!collversion)
-               collversion = get_collation_actual_version(collprovider, 
collcollate);
+               collversion = get_collation_actual_version(collprovider, 
collprovider == COLLPROVIDER_ICU ? colliculocale : collcollate);
 
        newoid = CollationCreate(collName,
                                                         collNamespace,
@@ -270,6 +293,7 @@ DefineCollation(ParseState *pstate, List *names, List 
*parameters, bool if_not_e
                                                         collencoding,
                                                         collcollate,
                                                         collctype,
+                                                        colliculocale,
                                                         collversion,
                                                         if_not_exists,
                                                         false);        /* not 
quiet */
@@ -352,9 +376,11 @@ AlterCollation(AlterCollationStmt *stmt)
        datum = SysCacheGetAttr(COLLOID, tup, Anum_pg_collation_collversion, 
&isnull);
        oldversion = isnull ? NULL : TextDatumGetCString(datum);
 
-       datum = SysCacheGetAttr(COLLOID, tup, Anum_pg_collation_collcollate, 
&isnull);
-       Assert(!isnull);
-       newversion = get_collation_actual_version(collForm->collprovider, 
TextDatumGetCString(datum));
+       datum = SysCacheGetAttr(COLLOID, tup, collForm->collprovider == 
COLLPROVIDER_ICU ? Anum_pg_collation_colliculocale : 
Anum_pg_collation_collcollate, &isnull);
+       if (!isnull)
+               newversion = 
get_collation_actual_version(collForm->collprovider, 
TextDatumGetCString(datum));
+       else
+               newversion = NULL;
 
        /* cannot change from NULL to non-NULL or vice versa */
        if ((!oldversion && newversion) || (oldversion && !newversion))
@@ -414,9 +440,11 @@ pg_collation_actual_version(PG_FUNCTION_ARGS)
 
        collprovider = ((Form_pg_collation) GETSTRUCT(tp))->collprovider;
 
-       datum = SysCacheGetAttr(COLLOID, tp, Anum_pg_collation_collcollate, 
&isnull);
-       Assert(!isnull);
-       version = get_collation_actual_version(collprovider, 
TextDatumGetCString(datum));
+       datum = SysCacheGetAttr(COLLOID, tp, collprovider == COLLPROVIDER_ICU ? 
Anum_pg_collation_colliculocale : Anum_pg_collation_collcollate, &isnull);
+       if (!isnull)
+               version = get_collation_actual_version(collprovider, 
TextDatumGetCString(datum));
+       else
+               version = NULL;
 
        ReleaseSysCache(tp);
 
@@ -643,7 +671,7 @@ pg_import_system_collations(PG_FUNCTION_ARGS)
                         */
                        collid = CollationCreate(localebuf, nspid, GetUserId(),
                                                                         
COLLPROVIDER_LIBC, true, enc,
-                                                                        
localebuf, localebuf,
+                                                                        
localebuf, localebuf, NULL,
                                                                         
get_collation_actual_version(COLLPROVIDER_LIBC, localebuf),
                                                                         true, 
true);
                        if (OidIsValid(collid))
@@ -704,7 +732,7 @@ pg_import_system_collations(PG_FUNCTION_ARGS)
 
                        collid = CollationCreate(alias, nspid, GetUserId(),
                                                                         
COLLPROVIDER_LIBC, true, enc,
-                                                                        
locale, locale,
+                                                                        
locale, locale, NULL,
                                                                         
get_collation_actual_version(COLLPROVIDER_LIBC, locale),
                                                                         true, 
true);
                        if (OidIsValid(collid))
@@ -745,7 +773,7 @@ pg_import_system_collations(PG_FUNCTION_ARGS)
                        const char *name;
                        char       *langtag;
                        char       *icucomment;
-                       const char *collcollate;
+                       const char *iculocstr;
                        Oid                     collid;
 
                        if (i == -1)
@@ -754,20 +782,20 @@ pg_import_system_collations(PG_FUNCTION_ARGS)
                                name = uloc_getAvailable(i);
 
                        langtag = get_icu_language_tag(name);
-                       collcollate = U_ICU_VERSION_MAJOR_NUM >= 54 ? langtag : 
name;
+                       iculocstr = U_ICU_VERSION_MAJOR_NUM >= 54 ? langtag : 
name;
 
                        /*
                         * Be paranoid about not allowing any non-ASCII strings 
into
                         * pg_collation
                         */
-                       if (!pg_is_ascii(langtag) || !pg_is_ascii(collcollate))
+                       if (!pg_is_ascii(langtag) || !pg_is_ascii(iculocstr))
                                continue;
 
                        collid = CollationCreate(psprintf("%s-x-icu", langtag),
                                                                         nspid, 
GetUserId(),
                                                                         
COLLPROVIDER_ICU, true, -1,
-                                                                        
collcollate, collcollate,
-                                                                        
get_collation_actual_version(COLLPROVIDER_ICU, collcollate),
+                                                                        NULL, 
NULL, iculocstr,
+                                                                        
get_collation_actual_version(COLLPROVIDER_ICU, iculocstr),
                                                                         true, 
true);
                        if (OidIsValid(collid))
                        {
diff --git a/src/backend/commands/dbcommands.c 
b/src/backend/commands/dbcommands.c
index c37e3c9a9a..f0b57fd69a 100644
--- a/src/backend/commands/dbcommands.c
+++ b/src/backend/commands/dbcommands.c
@@ -86,7 +86,8 @@ static bool get_db_info(const char *name, LOCKMODE lockmode,
                                                Oid *dbIdP, Oid *ownerIdP,
                                                int *encodingP, bool 
*dbIsTemplateP, bool *dbAllowConnP,
                                                TransactionId *dbFrozenXidP, 
MultiXactId *dbMinMultiP,
-                                               Oid *dbTablespace, char 
**dbCollate, char **dbCtype,
+                                               Oid *dbTablespace, char 
**dbCollate, char **dbCtype, char **dbIculocale,
+                                               char *dbLocProvider,
                                                char **dbCollversion);
 static bool have_createdb_privilege(void);
 static void remove_dbtablespaces(Oid db_id);
@@ -107,6 +108,8 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
        int                     src_encoding = -1;
        char       *src_collate = NULL;
        char       *src_ctype = NULL;
+       char       *src_iculocale = NULL;
+       char            src_locprovider;
        char       *src_collversion = NULL;
        bool            src_istemplate;
        bool            src_allowconn;
@@ -128,6 +131,8 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
        DefElem    *dlocale = NULL;
        DefElem    *dcollate = NULL;
        DefElem    *dctype = NULL;
+       DefElem    *diculocale = NULL;
+       DefElem    *dlocprovider = NULL;
        DefElem    *distemplate = NULL;
        DefElem    *dallowconnections = NULL;
        DefElem    *dconnlimit = NULL;
@@ -137,6 +142,8 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
        const char *dbtemplate = NULL;
        char       *dbcollate = NULL;
        char       *dbctype = NULL;
+       char       *dbiculocale = NULL;
+       char            dblocprovider = '\0';
        char       *canonname;
        int                     encoding = -1;
        bool            dbistemplate = false;
@@ -194,6 +201,18 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                                errorConflictingDefElem(defel, pstate);
                        dctype = defel;
                }
+               else if (strcmp(defel->defname, "icu_locale") == 0)
+               {
+                       if (diculocale)
+                               errorConflictingDefElem(defel, pstate);
+                       diculocale = defel;
+               }
+               else if (strcmp(defel->defname, "locale_provider") == 0)
+               {
+                       if (dlocprovider)
+                               errorConflictingDefElem(defel, pstate);
+                       dlocprovider = defel;
+               }
                else if (strcmp(defel->defname, "is_template") == 0)
                {
                        if (distemplate)
@@ -257,12 +276,6 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                                         parser_errposition(pstate, 
defel->location)));
        }
 
-       if (dlocale && (dcollate || dctype))
-               ereport(ERROR,
-                               (errcode(ERRCODE_SYNTAX_ERROR),
-                                errmsg("conflicting or redundant options"),
-                                errdetail("LOCALE cannot be specified together 
with LC_COLLATE or LC_CTYPE.")));
-
        if (downer && downer->arg)
                dbowner = defGetString(downer);
        if (dtemplate && dtemplate->arg)
@@ -304,6 +317,31 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                dbcollate = defGetString(dcollate);
        if (dctype && dctype->arg)
                dbctype = defGetString(dctype);
+       if (diculocale && diculocale->arg)
+               dbiculocale = defGetString(diculocale);
+       if (dlocprovider && dlocprovider->arg)
+       {
+               char       *locproviderstr = defGetString(dlocprovider);
+
+               if (pg_strcasecmp(locproviderstr, "icu") == 0)
+                       dblocprovider = COLLPROVIDER_ICU;
+               else if (pg_strcasecmp(locproviderstr, "libc") == 0)
+                       dblocprovider = COLLPROVIDER_LIBC;
+               else
+                       ereport(ERROR,
+                                       
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+                                        errmsg("unrecognized locale provider: 
%s",
+                                                       locproviderstr)));
+       }
+       if (diculocale && dblocprovider != COLLPROVIDER_ICU)
+               ereport(ERROR,
+                               (errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+                                errmsg("ICU locale cannot be specified unless 
locale provider is ICU")));
+       if (dblocprovider == COLLPROVIDER_ICU && !dbiculocale)
+       {
+               if (dlocale && dlocale->arg)
+                       dbiculocale = defGetString(dlocale);
+       }
        if (distemplate && distemplate->arg)
                dbistemplate = defGetBoolean(distemplate);
        if (dallowconnections && dallowconnections->arg)
@@ -355,7 +393,8 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                                         &src_dboid, &src_owner, &src_encoding,
                                         &src_istemplate, &src_allowconn,
                                         &src_frozenxid, &src_minmxid, 
&src_deftablespace,
-                                        &src_collate, &src_ctype, 
&src_collversion))
+                                        &src_collate, &src_ctype, 
&src_iculocale, &src_locprovider,
+                                        &src_collversion))
                ereport(ERROR,
                                (errcode(ERRCODE_UNDEFINED_DATABASE),
                                 errmsg("template database \"%s\" does not 
exist",
@@ -381,6 +420,10 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                dbcollate = src_collate;
        if (dbctype == NULL)
                dbctype = src_ctype;
+       if (dbiculocale == NULL)
+               dbiculocale = src_iculocale;
+       if (dblocprovider == '\0')
+               dblocprovider = src_locprovider;
 
        /* Some encodings are client only */
        if (!PG_VALID_BE_ENCODING(encoding))
@@ -402,6 +445,37 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
 
        check_encoding_locale_matches(encoding, dbcollate, dbctype);
 
+       if (dblocprovider == COLLPROVIDER_ICU)
+       {
+               /*
+                * This would happen if template0 uses the libc provider but 
the new
+                * database uses icu.
+                */
+               if (!dbiculocale)
+                       ereport(ERROR,
+                                       
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+                                        errmsg("ICU locale must be 
specified")));
+       }
+
+       if (dblocprovider == COLLPROVIDER_ICU)
+       {
+#ifdef USE_ICU
+               UErrorCode  status;
+
+               status = U_ZERO_ERROR;
+               ucol_open(dbiculocale, &status);
+               if (U_FAILURE(status))
+                       ereport(ERROR,
+                                       (errmsg("could not open collator for 
locale \"%s\": %s",
+                                                       dbiculocale, 
u_errorName(status))));
+#else
+               ereport(ERROR,
+                               (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+                                errmsg("ICU is not supported in this build"), \
+                                errhint("You need to rebuild PostgreSQL using 
%s.", "--with-icu")));
+#endif
+       }
+
        /*
         * Check that the new encoding and locale settings match the source
         * database.  We insist on this because we simply copy the source data 
---
@@ -435,6 +509,25 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                                         errmsg("new LC_CTYPE (%s) is 
incompatible with the LC_CTYPE of the template database (%s)",
                                                        dbctype, src_ctype),
                                         errhint("Use the same LC_CTYPE as in 
the template database, or use template0 as template.")));
+
+               if (dblocprovider != src_locprovider)
+                       ereport(ERROR,
+                                       
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+                                        errmsg("new locale provider (%s) does 
not match locale provider of the template database (%s)",
+                                                       
collprovider_name(dblocprovider), collprovider_name(src_locprovider)),
+                                        errhint("Use the same locale provider 
as in the template database, or use template0 as template.")));
+
+               if (dblocprovider == COLLPROVIDER_ICU)
+               {
+                       Assert(dbiculocale);
+                       Assert(src_iculocale);
+                       if (strcmp(dbiculocale, src_iculocale) != 0)
+                               ereport(ERROR,
+                                               
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+                                                errmsg("new ICU locale (%s) is 
incompatible with the ICU locale of the template database (%s)",
+                                                               dbiculocale, 
src_iculocale),
+                                                errhint("Use the same ICU 
locale as in the template database, or use template0 as template.")));
+               }
        }
 
        /*
@@ -453,7 +546,7 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
        {
                char       *actual_versionstr;
 
-               actual_versionstr = 
get_collation_actual_version(COLLPROVIDER_LIBC, dbcollate);
+               actual_versionstr = get_collation_actual_version(dblocprovider, 
dblocprovider == COLLPROVIDER_ICU ? dbiculocale : dbcollate);
                if (!actual_versionstr)
                        ereport(ERROR,
                                        (errmsg("template database \"%s\" has a 
collation version, but no actual collation version could be determined",
@@ -481,7 +574,7 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
         * collation version, which is normally only the case for template0.
         */
        if (dbcollversion == NULL)
-               dbcollversion = get_collation_actual_version(COLLPROVIDER_LIBC, 
dbcollate);
+               dbcollversion = get_collation_actual_version(dblocprovider, 
dblocprovider == COLLPROVIDER_ICU ? dbiculocale : dbcollate);
 
        /* Resolve default tablespace for new database */
        if (dtablespacename && dtablespacename->arg)
@@ -620,6 +713,9 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
         * block on the unique index, and fail after we commit).
         */
 
+       Assert((dblocprovider == COLLPROVIDER_ICU && dbiculocale) ||
+                  (dblocprovider != COLLPROVIDER_ICU && !dbiculocale));
+
        /* Form tuple */
        MemSet(new_record, 0, sizeof(new_record));
        MemSet(new_record_nulls, false, sizeof(new_record_nulls));
@@ -629,6 +725,7 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
                DirectFunctionCall1(namein, CStringGetDatum(dbname));
        new_record[Anum_pg_database_datdba - 1] = ObjectIdGetDatum(datdba);
        new_record[Anum_pg_database_encoding - 1] = Int32GetDatum(encoding);
+       new_record[Anum_pg_database_datlocprovider - 1] = 
CharGetDatum(dblocprovider);
        new_record[Anum_pg_database_datistemplate - 1] = 
BoolGetDatum(dbistemplate);
        new_record[Anum_pg_database_datallowconn - 1] = 
BoolGetDatum(dballowconnections);
        new_record[Anum_pg_database_datconnlimit - 1] = 
Int32GetDatum(dbconnlimit);
@@ -637,6 +734,10 @@ createdb(ParseState *pstate, const CreatedbStmt *stmt)
        new_record[Anum_pg_database_dattablespace - 1] = 
ObjectIdGetDatum(dst_deftablespace);
        new_record[Anum_pg_database_datcollate - 1] = 
CStringGetTextDatum(dbcollate);
        new_record[Anum_pg_database_datctype - 1] = 
CStringGetTextDatum(dbctype);
+       if (dbiculocale)
+               new_record[Anum_pg_database_daticulocale - 1] = 
CStringGetTextDatum(dbiculocale);
+       else
+               new_record_nulls[Anum_pg_database_daticulocale - 1] = true;
        if (dbcollversion)
                new_record[Anum_pg_database_datcollversion - 1] = 
CStringGetTextDatum(dbcollversion);
        else
@@ -907,7 +1008,7 @@ dropdb(const char *dbname, bool missing_ok, bool force)
        pgdbrel = table_open(DatabaseRelationId, RowExclusiveLock);
 
        if (!get_db_info(dbname, AccessExclusiveLock, &db_id, NULL, NULL,
-                                        &db_istemplate, NULL, NULL, NULL, 
NULL, NULL, NULL, NULL))
+                                        &db_istemplate, NULL, NULL, NULL, 
NULL, NULL, NULL, NULL, NULL, NULL))
        {
                if (!missing_ok)
                {
@@ -1109,7 +1210,7 @@ RenameDatabase(const char *oldname, const char *newname)
        rel = table_open(DatabaseRelationId, RowExclusiveLock);
 
        if (!get_db_info(oldname, AccessExclusiveLock, &db_id, NULL, NULL,
-                                        NULL, NULL, NULL, NULL, NULL, NULL, 
NULL, NULL))
+                                        NULL, NULL, NULL, NULL, NULL, NULL, 
NULL, NULL, NULL, NULL))
                ereport(ERROR,
                                (errcode(ERRCODE_UNDEFINED_DATABASE),
                                 errmsg("database \"%s\" does not exist", 
oldname)));
@@ -1222,7 +1323,7 @@ movedb(const char *dbname, const char *tblspcname)
        pgdbrel = table_open(DatabaseRelationId, RowExclusiveLock);
 
        if (!get_db_info(dbname, AccessExclusiveLock, &db_id, NULL, NULL,
-                                        NULL, NULL, NULL, NULL, 
&src_tblspcoid, NULL, NULL, NULL))
+                                        NULL, NULL, NULL, NULL, 
&src_tblspcoid, NULL, NULL, NULL, NULL, NULL))
                ereport(ERROR,
                                (errcode(ERRCODE_UNDEFINED_DATABASE),
                                 errmsg("database \"%s\" does not exist", 
dbname)));
@@ -1755,9 +1856,11 @@ AlterDatabaseRefreshColl(AlterDatabaseRefreshCollStmt 
*stmt)
        datum = heap_getattr(tuple, Anum_pg_database_datcollversion, 
RelationGetDescr(rel), &isnull);
        oldversion = isnull ? NULL : TextDatumGetCString(datum);
 
-       datum = heap_getattr(tuple, Anum_pg_database_datcollate, 
RelationGetDescr(rel), &isnull);
-       Assert(!isnull);
-       newversion = get_collation_actual_version(COLLPROVIDER_LIBC, 
TextDatumGetCString(datum));
+       datum = heap_getattr(tuple, datForm->datlocprovider == COLLPROVIDER_ICU 
? Anum_pg_database_daticulocale : Anum_pg_database_datcollate, 
RelationGetDescr(rel), &isnull);
+       if (!isnull)
+               newversion = 
get_collation_actual_version(datForm->datlocprovider, 
TextDatumGetCString(datum));
+       else
+               newversion = NULL;
 
        /* cannot change from NULL to non-NULL or vice versa */
        if ((!oldversion && newversion) || (oldversion && !newversion))
@@ -1943,6 +2046,7 @@ pg_database_collation_actual_version(PG_FUNCTION_ARGS)
 {
        Oid                     dbid = PG_GETARG_OID(0);
        HeapTuple       tp;
+       char            datlocprovider;
        Datum           datum;
        bool            isnull;
        char       *version;
@@ -1953,9 +2057,13 @@ pg_database_collation_actual_version(PG_FUNCTION_ARGS)
                                (errcode(ERRCODE_UNDEFINED_OBJECT),
                                 errmsg("database with OID %u does not exist", 
dbid)));
 
-       datum = SysCacheGetAttr(DATABASEOID, tp, Anum_pg_database_datcollate, 
&isnull);
-       Assert(!isnull);
-       version = get_collation_actual_version(COLLPROVIDER_LIBC, 
TextDatumGetCString(datum));
+       datlocprovider = ((Form_pg_database) GETSTRUCT(tp))->datlocprovider;
+
+       datum = SysCacheGetAttr(DATABASEOID, tp, datlocprovider == 
COLLPROVIDER_ICU ? Anum_pg_database_daticulocale : Anum_pg_database_datcollate, 
&isnull);
+       if (!isnull)
+               version = get_collation_actual_version(datlocprovider, 
TextDatumGetCString(datum));
+       else
+               version = NULL;
 
        ReleaseSysCache(tp);
 
@@ -1981,7 +2089,8 @@ get_db_info(const char *name, LOCKMODE lockmode,
                        Oid *dbIdP, Oid *ownerIdP,
                        int *encodingP, bool *dbIsTemplateP, bool *dbAllowConnP,
                        TransactionId *dbFrozenXidP, MultiXactId *dbMinMultiP,
-                       Oid *dbTablespace, char **dbCollate, char **dbCtype,
+                       Oid *dbTablespace, char **dbCollate, char **dbCtype, 
char **dbIculocale,
+                       char *dbLocProvider,
                        char **dbCollversion)
 {
        bool            result = false;
@@ -2075,6 +2184,8 @@ get_db_info(const char *name, LOCKMODE lockmode,
                                if (dbTablespace)
                                        *dbTablespace = dbform->dattablespace;
                                /* default locale settings for this database */
+                               if (dbLocProvider)
+                                       *dbLocProvider = dbform->datlocprovider;
                                if (dbCollate)
                                {
                                        datum = SysCacheGetAttr(DATABASEOID, 
tuple, Anum_pg_database_datcollate, &isnull);
@@ -2087,6 +2198,14 @@ get_db_info(const char *name, LOCKMODE lockmode,
                                        Assert(!isnull);
                                        *dbCtype = TextDatumGetCString(datum);
                                }
+                               if (dbIculocale)
+                               {
+                                       datum = SysCacheGetAttr(DATABASEOID, 
tuple, Anum_pg_database_daticulocale, &isnull);
+                                       if (isnull)
+                                               *dbIculocale = NULL;
+                                       else
+                                               *dbIculocale = 
TextDatumGetCString(datum);
+                               }
                                if (dbCollversion)
                                {
                                        datum = SysCacheGetAttr(DATABASEOID, 
tuple, Anum_pg_database_datcollversion, &isnull);
diff --git a/src/backend/utils/adt/pg_locale.c 
b/src/backend/utils/adt/pg_locale.c
index 871a710967..4019255f8e 100644
--- a/src/backend/utils/adt/pg_locale.c
+++ b/src/backend/utils/adt/pg_locale.c
@@ -1288,26 +1288,37 @@ lookup_collation_cache(Oid collation, bool set_flags)
        {
                /* Attempt to set the flags */
                HeapTuple       tp;
-               Datum           datum;
-               bool            isnull;
-               const char *collcollate;
-               const char *collctype;
+               Form_pg_collation collform;
 
                tp = SearchSysCache1(COLLOID, ObjectIdGetDatum(collation));
                if (!HeapTupleIsValid(tp))
                        elog(ERROR, "cache lookup failed for collation %u", 
collation);
+               collform = (Form_pg_collation) GETSTRUCT(tp);
 
-               datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collcollate, &isnull);
-               Assert(!isnull);
-               collcollate = TextDatumGetCString(datum);
-               datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collctype, &isnull);
-               Assert(!isnull);
-               collctype = TextDatumGetCString(datum);
-
-               cache_entry->collate_is_c = ((strcmp(collcollate, "C") == 0) ||
-                                                                        
(strcmp(collcollate, "POSIX") == 0));
-               cache_entry->ctype_is_c = ((strcmp(collctype, "C") == 0) ||
-                                                                  
(strcmp(collctype, "POSIX") == 0));
+               if (collform->collprovider == COLLPROVIDER_LIBC)
+               {
+                       Datum           datum;
+                       bool            isnull;
+                       const char *collcollate;
+                       const char *collctype;
+
+                       datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collcollate, &isnull);
+                       Assert(!isnull);
+                       collcollate = TextDatumGetCString(datum);
+                       datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collctype, &isnull);
+                       Assert(!isnull);
+                       collctype = TextDatumGetCString(datum);
+
+                       cache_entry->collate_is_c = ((strcmp(collcollate, "C") 
== 0) ||
+                                                                               
 (strcmp(collcollate, "POSIX") == 0));
+                       cache_entry->ctype_is_c = ((strcmp(collctype, "C") == 
0) ||
+                                                                          
(strcmp(collctype, "POSIX") == 0));
+               }
+               else
+               {
+                       cache_entry->collate_is_c = false;
+                       cache_entry->ctype_is_c = false;
+               }
 
                cache_entry->flags_valid = true;
 
@@ -1340,6 +1351,9 @@ lc_collate_is_c(Oid collation)
                static int      result = -1;
                char       *localeptr;
 
+               if (default_locale.provider == COLLPROVIDER_ICU)
+                       return false;
+
                if (result >= 0)
                        return (bool) result;
                localeptr = setlocale(LC_COLLATE, NULL);
@@ -1390,6 +1404,9 @@ lc_ctype_is_c(Oid collation)
                static int      result = -1;
                char       *localeptr;
 
+               if (default_locale.provider == COLLPROVIDER_ICU)
+                       return false;
+
                if (result >= 0)
                        return (bool) result;
                localeptr = setlocale(LC_CTYPE, NULL);
@@ -1418,6 +1435,38 @@ lc_ctype_is_c(Oid collation)
        return (lookup_collation_cache(collation, true))->ctype_is_c;
 }
 
+struct pg_locale_struct default_locale;
+
+void
+make_icu_collator(const char *iculocstr,
+                                 struct pg_locale_struct *resultp)
+{
+#ifdef USE_ICU
+       UCollator  *collator;
+       UErrorCode      status;
+
+       status = U_ZERO_ERROR;
+       collator = ucol_open(iculocstr, &status);
+       if (U_FAILURE(status))
+               ereport(ERROR,
+                               (errmsg("could not open collator for locale 
\"%s\": %s",
+                                               iculocstr, 
u_errorName(status))));
+
+       if (U_ICU_VERSION_MAJOR_NUM < 54)
+               icu_set_collation_attributes(collator, iculocstr);
+
+       /* We will leak this string if the caller errors later :-( */
+       resultp->info.icu.locale = MemoryContextStrdup(TopMemoryContext, 
iculocstr);
+       resultp->info.icu.ucol = collator;
+#else                                                  /* not USE_ICU */
+       /* could get here if a collation was created by a build with ICU */
+       ereport(ERROR,
+                       (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+                        errmsg("ICU is not supported in this build"), \
+                        errhint("You need to rebuild PostgreSQL using %s.", 
"--with-icu")));
+#endif                                                 /* not USE_ICU */
+}
+
 
 /* simple subroutine for reporting errors from newlocale() */
 #ifdef HAVE_LOCALE_T
@@ -1475,7 +1524,12 @@ pg_newlocale_from_collation(Oid collid)
        Assert(OidIsValid(collid));
 
        if (collid == DEFAULT_COLLATION_OID)
-               return (pg_locale_t) 0;
+       {
+               if (default_locale.provider == COLLPROVIDER_ICU)
+                       return &default_locale;
+               else
+                       return (pg_locale_t) 0;
+       }
 
        cache_entry = lookup_collation_cache(collid, false);
 
@@ -1484,8 +1538,6 @@ pg_newlocale_from_collation(Oid collid)
                /* We haven't computed this yet in this session, so do it */
                HeapTuple       tp;
                Form_pg_collation collform;
-               const char *collcollate;
-               const char *collctype pg_attribute_unused();
                struct pg_locale_struct result;
                pg_locale_t resultp;
                Datum           datum;
@@ -1496,13 +1548,6 @@ pg_newlocale_from_collation(Oid collid)
                        elog(ERROR, "cache lookup failed for collation %u", 
collid);
                collform = (Form_pg_collation) GETSTRUCT(tp);
 
-               datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collcollate, &isnull);
-               Assert(!isnull);
-               collcollate = TextDatumGetCString(datum);
-               datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collctype, &isnull);
-               Assert(!isnull);
-               collctype = TextDatumGetCString(datum);
-
                /* We'll fill in the result struct locally before allocating 
memory */
                memset(&result, 0, sizeof(result));
                result.provider = collform->collprovider;
@@ -1511,8 +1556,17 @@ pg_newlocale_from_collation(Oid collid)
                if (collform->collprovider == COLLPROVIDER_LIBC)
                {
 #ifdef HAVE_LOCALE_T
+                       const char *collcollate;
+                       const char *collctype pg_attribute_unused();
                        locale_t        loc;
 
+                       datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collcollate, &isnull);
+                       Assert(!isnull);
+                       collcollate = TextDatumGetCString(datum);
+                       datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collctype, &isnull);
+                       Assert(!isnull);
+                       collctype = TextDatumGetCString(datum);
+
                        if (strcmp(collcollate, collctype) == 0)
                        {
                                /* Normal case where they're the same */
@@ -1563,36 +1617,12 @@ pg_newlocale_from_collation(Oid collid)
                }
                else if (collform->collprovider == COLLPROVIDER_ICU)
                {
-#ifdef USE_ICU
-                       UCollator  *collator;
-                       UErrorCode      status;
-
-                       if (strcmp(collcollate, collctype) != 0)
-                               ereport(ERROR,
-                                               
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
-                                                errmsg("collations with 
different collate and ctype values are not supported by ICU")));
-
-                       status = U_ZERO_ERROR;
-                       collator = ucol_open(collcollate, &status);
-                       if (U_FAILURE(status))
-                               ereport(ERROR,
-                                               (errmsg("could not open 
collator for locale \"%s\": %s",
-                                                               collcollate, 
u_errorName(status))));
+                       const char *iculocstr;
 
-                       if (U_ICU_VERSION_MAJOR_NUM < 54)
-                               icu_set_collation_attributes(collator, 
collcollate);
-
-                       /* We will leak this string if we get an error below 
:-( */
-                       result.info.icu.locale = 
MemoryContextStrdup(TopMemoryContext,
-                                                                               
                                 collcollate);
-                       result.info.icu.ucol = collator;
-#else                                                  /* not USE_ICU */
-                       /* could get here if a collation was created by a build 
with ICU */
-                       ereport(ERROR,
-                                       (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
-                                        errmsg("ICU is not supported in this 
build"), \
-                                        errhint("You need to rebuild 
PostgreSQL using %s.", "--with-icu")));
-#endif                                                 /* not USE_ICU */
+                       datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_colliculocale, &isnull);
+                       Assert(!isnull);
+                       iculocstr = TextDatumGetCString(datum);
+                       make_icu_collator(iculocstr, &result);
                }
 
                datum = SysCacheGetAttr(COLLOID, tp, 
Anum_pg_collation_collversion,
@@ -1604,7 +1634,11 @@ pg_newlocale_from_collation(Oid collid)
 
                        collversionstr = TextDatumGetCString(datum);
 
-                       actual_versionstr = 
get_collation_actual_version(collform->collprovider, collcollate);
+                       datum = SysCacheGetAttr(COLLOID, tp, 
collform->collprovider == COLLPROVIDER_ICU ? Anum_pg_collation_colliculocale : 
Anum_pg_collation_collcollate, &isnull);
+                       Assert(!isnull);
+
+                       actual_versionstr = 
get_collation_actual_version(collform->collprovider,
+                                                                               
                                         TextDatumGetCString(datum));
                        if (!actual_versionstr)
                        {
                                /*
diff --git a/src/backend/utils/init/postinit.c 
b/src/backend/utils/init/postinit.c
index 86d193c89f..6452b42dbf 100644
--- a/src/backend/utils/init/postinit.c
+++ b/src/backend/utils/init/postinit.c
@@ -318,6 +318,7 @@ CheckMyDatabase(const char *name, bool am_superuser, bool 
override_allow_connect
        bool            isnull;
        char       *collate;
        char       *ctype;
+       char       *iculocale;
 
        /* Fetch our pg_database row normally, via syscache */
        tup = SearchSysCache1(DATABASEOID, ObjectIdGetDatum(MyDatabaseId));
@@ -420,6 +421,24 @@ CheckMyDatabase(const char *name, bool am_superuser, bool 
override_allow_connect
                                                   " which is not recognized by 
setlocale().", ctype),
                                 errhint("Recreate the database with another 
locale or install the missing locale.")));
 
+       if (dbform->datlocprovider == COLLPROVIDER_ICU)
+       {
+               datum = SysCacheGetAttr(DATABASEOID, tup, 
Anum_pg_database_daticulocale, &isnull);
+               Assert(!isnull);
+               iculocale = TextDatumGetCString(datum);
+               make_icu_collator(iculocale, &default_locale);
+       }
+       else
+               iculocale = NULL;
+
+       default_locale.provider = dbform->datlocprovider;
+       /*
+        * Default locale is currently always deterministic.  Nondeterministic
+        * locales currently don't support pattern matching, which would break a
+        * lot of things if applied globally.
+        */
+       default_locale.deterministic = true;
+
        /*
         * Check collation version.  See similar code in
         * pg_newlocale_from_collation().  Note that here we warn instead of 
error
@@ -434,7 +453,7 @@ CheckMyDatabase(const char *name, bool am_superuser, bool 
override_allow_connect
 
                collversionstr = TextDatumGetCString(datum);
 
-               actual_versionstr = 
get_collation_actual_version(COLLPROVIDER_LIBC, collate);
+               actual_versionstr = 
get_collation_actual_version(dbform->datlocprovider, dbform->datlocprovider == 
COLLPROVIDER_ICU ? iculocale : collate);
                if (!actual_versionstr)
                        ereport(WARNING,
                                        (errmsg("database \"%s\" has no actual 
collation version, but a version was recorded",
diff --git a/src/bin/initdb/Makefile b/src/bin/initdb/Makefile
index eba282267a..8dd25e7afc 100644
--- a/src/bin/initdb/Makefile
+++ b/src/bin/initdb/Makefile
@@ -40,7 +40,7 @@ OBJS = \
 all: initdb
 
 initdb: $(OBJS) | submake-libpq submake-libpgport submake-libpgfeutils
-       $(CC) $(CFLAGS) $(OBJS) $(LDFLAGS) $(LDFLAGS_EX) $(LIBS) -o $@$(X)
+       $(CC) $(CFLAGS) $(OBJS) $(LDFLAGS) $(LDFLAGS_EX) $(LIBS) $(ICU_LIBS) -o 
$@$(X)
 
 # We must pull in localtime.c from src/timezones
 localtime.c: % : $(top_srcdir)/src/timezone/%
@@ -62,6 +62,8 @@ clean distclean maintainer-clean:
 # ensure that changes in datadir propagate into object file
 initdb.o: initdb.c $(top_builddir)/src/Makefile.global
 
+export with_icu
+
 check:
        $(prove_check)
 
diff --git a/src/bin/initdb/initdb.c b/src/bin/initdb/initdb.c
index 97f15971e2..cbcd55288f 100644
--- a/src/bin/initdb/initdb.c
+++ b/src/bin/initdb/initdb.c
@@ -55,6 +55,10 @@
 #include <signal.h>
 #include <time.h>
 
+#ifdef USE_ICU
+#include <unicode/ucol.h>
+#endif
+
 #ifdef HAVE_SHM_OPEN
 #include "sys/mman.h"
 #endif
@@ -132,6 +136,8 @@ static char *lc_monetary = NULL;
 static char *lc_numeric = NULL;
 static char *lc_time = NULL;
 static char *lc_messages = NULL;
+static char locale_provider = COLLPROVIDER_LIBC;
+static char *icu_locale = NULL;
 static const char *default_text_search_config = NULL;
 static char *username = NULL;
 static bool pwprompt = false;
@@ -1405,6 +1411,12 @@ bootstrap_template1(void)
        bki_lines = replace_token(bki_lines, "LC_CTYPE",
                                                          
escape_quotes_bki(lc_ctype));
 
+       bki_lines = replace_token(bki_lines, "ICU_LOCALE",
+                                                         locale_provider == 
COLLPROVIDER_ICU ? escape_quotes_bki(icu_locale) : "_null_");
+
+       sprintf(buf, "%c", locale_provider);
+       bki_lines = replace_token(bki_lines, "LOCALE_PROVIDER", buf);
+
        /* Also ensure backend isn't confused by this environment var: */
        unsetenv("PGCLIENTENCODING");
 
@@ -2165,7 +2177,6 @@ setlocales(void)
         * canonicalize locale names, and obtain any missing values from our
         * current environment
         */
-
        check_locale_name(LC_CTYPE, lc_ctype, &canonname);
        lc_ctype = canonname;
        check_locale_name(LC_COLLATE, lc_collate, &canonname);
@@ -2184,6 +2195,37 @@ setlocales(void)
        check_locale_name(LC_CTYPE, lc_messages, &canonname);
        lc_messages = canonname;
 #endif
+
+       if (locale_provider == COLLPROVIDER_ICU)
+       {
+               if (!icu_locale)
+               {
+                       pg_log_error("ICU locale must be specified");
+                       exit(1);
+               }
+
+               /*
+                * Check ICU locale ID
+                */
+#ifdef USE_ICU
+               {
+                       UErrorCode      status;
+
+                       status = U_ZERO_ERROR;
+                       ucol_open(icu_locale, &status);
+                       if (U_FAILURE(status))
+                       {
+                               pg_log_error("could not open collator for 
locale \"%s\": %s",
+                                                        icu_locale, 
u_errorName(status));
+                               exit(1);
+                       }
+               }
+#else
+               pg_log_error("ICU is not supported in this build");
+               fprintf(stderr, _("You need to rebuild PostgreSQL using 
%s.\n"), "--with-icu");
+               exit(1);
+#endif
+       }
 }
 
 /*
@@ -2202,6 +2244,7 @@ usage(const char *progname)
        printf(_(" [-D, --pgdata=]DATADIR     location for this database 
cluster\n"));
        printf(_("  -E, --encoding=ENCODING   set default encoding for new 
databases\n"));
        printf(_("  -g, --allow-group-access  allow group read/execute on data 
directory\n"));
+       printf(_("      --icu-locale=LOCALE   set ICU locale ID for new 
databases\n"));
        printf(_("  -k, --data-checksums      use data page checksums\n"));
        printf(_("      --locale=LOCALE       set default locale for new 
databases\n"));
        printf(_("      --lc-collate=, --lc-ctype=, --lc-messages=LOCALE\n"
@@ -2209,6 +2252,8 @@ usage(const char *progname)
                         "                            set default locale in the 
respective category for\n"
                         "                            new databases (default 
taken from environment)\n"));
        printf(_("      --no-locale           equivalent to --locale=C\n"));
+       printf(_("      --locale-provider={libc|icu}\n"
+                        "                            set default locale 
provider for new databases\n"));
        printf(_("      --pwfile=FILE         read password for the new 
superuser from file\n"));
        printf(_("  -T, --text-search-config=CFG\n"
                         "                            default text search 
configuration\n"));
@@ -2372,21 +2417,26 @@ setup_locale_encoding(void)
 {
        setlocales();
 
-       if (strcmp(lc_ctype, lc_collate) == 0 &&
+       if (locale_provider == COLLPROVIDER_LIBC &&
+               strcmp(lc_ctype, lc_collate) == 0 &&
                strcmp(lc_ctype, lc_time) == 0 &&
                strcmp(lc_ctype, lc_numeric) == 0 &&
                strcmp(lc_ctype, lc_monetary) == 0 &&
-               strcmp(lc_ctype, lc_messages) == 0)
+               strcmp(lc_ctype, lc_messages) == 0 &&
+               (!icu_locale || strcmp(lc_ctype, icu_locale) == 0))
                printf(_("The database cluster will be initialized with locale 
\"%s\".\n"), lc_ctype);
        else
        {
-               printf(_("The database cluster will be initialized with 
locales\n"
-                                "  COLLATE:  %s\n"
-                                "  CTYPE:    %s\n"
-                                "  MESSAGES: %s\n"
-                                "  MONETARY: %s\n"
-                                "  NUMERIC:  %s\n"
-                                "  TIME:     %s\n"),
+               printf(_("The database cluster will be initialized with this 
locale configuration:\n"));
+               printf(_("  provider:    %s\n"), 
collprovider_name(locale_provider));
+               if (icu_locale)
+                       printf(_("  ICU locale:  %s\n"), icu_locale);
+               printf(_("  LC_COLLATE:  %s\n"
+                                "  LC_CTYPE:    %s\n"
+                                "  LC_MESSAGES: %s\n"
+                                "  LC_MONETARY: %s\n"
+                                "  LC_NUMERIC:  %s\n"
+                                "  LC_TIME:     %s\n"),
                           lc_collate,
                           lc_ctype,
                           lc_messages,
@@ -2395,7 +2445,9 @@ setup_locale_encoding(void)
                           lc_time);
        }
 
-       if (!encoding)
+       if (!encoding && locale_provider == COLLPROVIDER_ICU)
+               encodingid = PG_UTF8;
+       else if (!encoding)
        {
                int                     ctype_enc;
 
@@ -2899,6 +2951,8 @@ main(int argc, char *argv[])
                {"data-checksums", no_argument, NULL, 'k'},
                {"allow-group-access", no_argument, NULL, 'g'},
                {"discard-caches", no_argument, NULL, 14},
+               {"locale-provider", required_argument, NULL, 15},
+               {"icu-locale", required_argument, NULL, 16},
                {NULL, 0, NULL, 0}
        };
 
@@ -3045,6 +3099,20 @@ main(int argc, char *argv[])
                                                                                
 extra_options,
                                                                                
 "-c debug_discard_caches=1");
                                break;
+                       case 15:
+                               if (strcmp(optarg, "icu") == 0)
+                                       locale_provider = COLLPROVIDER_ICU;
+                               else if (strcmp(optarg, "libc") == 0)
+                                       locale_provider = COLLPROVIDER_LIBC;
+                               else
+                               {
+                                       pg_log_error("unrecognized locale 
provider: %s", optarg);
+                                       exit(1);
+                               }
+                               break;
+                       case 16:
+                               icu_locale = pg_strdup(optarg);
+                               break;
                        default:
                                /* getopt_long already emitted a complaint */
                                fprintf(stderr, _("Try \"%s --help\" for more 
information.\n"),
@@ -3073,6 +3141,13 @@ main(int argc, char *argv[])
                exit(1);
        }
 
+       if (icu_locale && locale_provider != COLLPROVIDER_ICU)
+       {
+               pg_log_error("%s cannot be specified unless locale provider 
\"%s\" is chosen",
+                                        "--icu-locale", "icu");
+               exit(1);
+       }
+
        atexit(cleanup_directories_atexit);
 
        /* If we only need to fsync, just do it and exit */
diff --git a/src/bin/initdb/t/001_initdb.pl b/src/bin/initdb/t/001_initdb.pl
index 7dc8cdd855..c636bf3ab2 100644
--- a/src/bin/initdb/t/001_initdb.pl
+++ b/src/bin/initdb/t/001_initdb.pl
@@ -93,4 +93,31 @@
                'check PGDATA permissions');
 }
 
+# Locale provider tests
+
+if ($ENV{with_icu} eq 'yes')
+{
+       command_fails_like(['initdb', '--no-sync', '--locale-provider=icu', 
"$tempdir/data2"],
+               qr/initdb: error: ICU locale must be specified/,
+               'locale provider ICU requires --icu-locale');
+
+       command_ok(['initdb', '--no-sync', '--locale-provider=icu', 
'--icu-locale=en', "$tempdir/data3"],
+               'option --icu-locale');
+
+       command_fails_like(['initdb', '--no-sync', '--locale-provider=icu', 
'--icu-locale=@colNumeric=lower', "$tempdir/dataX"],
+               qr/initdb: error: could not open collator for locale/,
+               'fails for invalid ICU locale');
+}
+else
+{
+       command_fails(['initdb', '--no-sync', '--locale-provider=icu', 
"$tempdir/data2"],
+                                 'locale provider ICU fails since no ICU 
support');
+}
+
+command_fails(['initdb', '--no-sync', '--locale-provider=xyz', 
"$tempdir/dataX"],
+                         'fails for invalid locale provider');
+
+command_fails(['initdb', '--no-sync', '--locale-provider=libc', 
'--icu-locale=en', "$tempdir/dataX"],
+                         'fails for invalid option combination');
+
 done_testing();
diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c
index 4dd24b8c89..725cd2e4eb 100644
--- a/src/bin/pg_dump/pg_dump.c
+++ b/src/bin/pg_dump/pg_dump.c
@@ -2753,8 +2753,10 @@ dumpDatabase(Archive *fout)
                                i_datname,
                                i_datdba,
                                i_encoding,
+                               i_datlocprovider,
                                i_collate,
                                i_ctype,
+                               i_daticulocale,
                                i_frozenxid,
                                i_minmxid,
                                i_datacl,
@@ -2769,8 +2771,10 @@ dumpDatabase(Archive *fout)
        const char *datname,
                           *dba,
                           *encoding,
+                          *datlocprovider,
                           *collate,
                           *ctype,
+                          *iculocale,
                           *datistemplate,
                           *datconnlimit,
                           *tablespace;
@@ -2794,9 +2798,9 @@ dumpDatabase(Archive *fout)
        else
                appendPQExpBuffer(dbQry, "0 AS datminmxid, ");
        if (fout->remoteVersion >= 150000)
-               appendPQExpBuffer(dbQry, "datcollversion, ");
+               appendPQExpBuffer(dbQry, "datlocprovider, daticulocale, 
datcollversion, ");
        else
-               appendPQExpBuffer(dbQry, "NULL AS datcollversion, ");
+               appendPQExpBuffer(dbQry, "'c' AS datlocprovider, NULL AS 
daticulocale, NULL AS datcollversion, ");
        appendPQExpBuffer(dbQry,
                                          "(SELECT spcname FROM pg_tablespace t 
WHERE t.oid = dattablespace) AS tablespace, "
                                          "shobj_description(oid, 
'pg_database') AS description "
@@ -2810,8 +2814,10 @@ dumpDatabase(Archive *fout)
        i_datname = PQfnumber(res, "datname");
        i_datdba = PQfnumber(res, "datdba");
        i_encoding = PQfnumber(res, "encoding");
+       i_datlocprovider = PQfnumber(res, "datlocprovider");
        i_collate = PQfnumber(res, "datcollate");
        i_ctype = PQfnumber(res, "datctype");
+       i_daticulocale = PQfnumber(res, "daticulocale");
        i_frozenxid = PQfnumber(res, "datfrozenxid");
        i_minmxid = PQfnumber(res, "datminmxid");
        i_datacl = PQfnumber(res, "datacl");
@@ -2826,8 +2832,13 @@ dumpDatabase(Archive *fout)
        datname = PQgetvalue(res, 0, i_datname);
        dba = getRoleName(PQgetvalue(res, 0, i_datdba));
        encoding = PQgetvalue(res, 0, i_encoding);
+       datlocprovider = PQgetvalue(res, 0, i_datlocprovider);
        collate = PQgetvalue(res, 0, i_collate);
        ctype = PQgetvalue(res, 0, i_ctype);
+       if (!PQgetisnull(res, 0, i_daticulocale))
+               iculocale = PQgetvalue(res, 0, i_daticulocale);
+       else
+               iculocale = NULL;
        frozenxid = atooid(PQgetvalue(res, 0, i_frozenxid));
        minmxid = atooid(PQgetvalue(res, 0, i_minmxid));
        dbdacl.acl = PQgetvalue(res, 0, i_datacl);
@@ -2859,6 +2870,16 @@ dumpDatabase(Archive *fout)
                appendPQExpBufferStr(creaQry, " ENCODING = ");
                appendStringLiteralAH(creaQry, encoding, fout);
        }
+
+       appendPQExpBufferStr(creaQry, " LOCALE_PROVIDER = ");
+       if (datlocprovider[0] == 'c')
+               appendPQExpBufferStr(creaQry, "libc");
+       else if (datlocprovider[0] == 'i')
+               appendPQExpBufferStr(creaQry, "icu");
+       else
+               fatal("unrecognized locale provider: %s",
+                         datlocprovider);
+
        if (strlen(collate) > 0 && strcmp(collate, ctype) == 0)
        {
                appendPQExpBufferStr(creaQry, " LOCALE = ");
@@ -2877,6 +2898,11 @@ dumpDatabase(Archive *fout)
                        appendStringLiteralAH(creaQry, ctype, fout);
                }
        }
+       if (iculocale)
+       {
+               appendPQExpBufferStr(creaQry, " ICU_LOCALE = ");
+               appendStringLiteralAH(creaQry, iculocale, fout);
+       }
 
        /*
         * For binary upgrade, carry over the collation version.  For normal
diff --git a/src/bin/pg_upgrade/check.c b/src/bin/pg_upgrade/check.c
index 019bcb6c7b..cf3b398d9e 100644
--- a/src/bin/pg_upgrade/check.c
+++ b/src/bin/pg_upgrade/check.c
@@ -10,6 +10,7 @@
 #include "postgres_fe.h"
 
 #include "catalog/pg_authid_d.h"
+#include "catalog/pg_collation.h"
 #include "fe_utils/string_utils.h"
 #include "mb/pg_wchar.h"
 #include "pg_upgrade.h"
@@ -349,6 +350,18 @@ check_locale_and_encoding(DbInfo *olddb, DbInfo *newdb)
        if (!equivalent_locale(LC_CTYPE, olddb->db_ctype, newdb->db_ctype))
                pg_fatal("lc_ctype values for database \"%s\" do not match:  
old \"%s\", new \"%s\"\n",
                                 olddb->db_name, olddb->db_ctype, 
newdb->db_ctype);
+       if (olddb->db_collprovider != newdb->db_collprovider)
+               pg_fatal("locale providers for database \"%s\" do not match:  
old \"%s\", new \"%s\"\n",
+                                olddb->db_name,
+                                collprovider_name(olddb->db_collprovider),
+                                collprovider_name(newdb->db_collprovider));
+       if ((olddb->db_iculocale == NULL && newdb->db_iculocale != NULL) ||
+               (olddb->db_iculocale != NULL && newdb->db_iculocale == NULL) ||
+               (olddb->db_iculocale != NULL && newdb->db_iculocale != NULL && 
strcmp(olddb->db_iculocale, newdb->db_iculocale) != 0))
+               pg_fatal("ICU locale values for database \"%s\" do not match:  
old \"%s\", new \"%s\"\n",
+                                olddb->db_name,
+                                olddb->db_iculocale ? olddb->db_iculocale : 
"(null)",
+                                newdb->db_iculocale ? newdb->db_iculocale : 
"(null)");
 }
 
 /*
diff --git a/src/bin/pg_upgrade/info.c b/src/bin/pg_upgrade/info.c
index 69ef23119f..5c3968e0ea 100644
--- a/src/bin/pg_upgrade/info.c
+++ b/src/bin/pg_upgrade/info.c
@@ -312,11 +312,20 @@ get_db_infos(ClusterInfo *cluster)
                                i_encoding,
                                i_datcollate,
                                i_datctype,
+                               i_datlocprovider,
+                               i_daticulocale,
                                i_spclocation;
        char            query[QUERY_ALLOC];
 
        snprintf(query, sizeof(query),
-                        "SELECT d.oid, d.datname, d.encoding, d.datcollate, 
d.datctype, "
+                        "SELECT d.oid, d.datname, d.encoding, d.datcollate, 
d.datctype, ");
+       if (GET_MAJOR_VERSION(old_cluster.major_version) <= 1500)
+               snprintf(query + strlen(query), sizeof(query) - strlen(query),
+                                "'c' AS datlocprovider, NULL AS daticulocale, 
");
+       else
+               snprintf(query + strlen(query), sizeof(query) - strlen(query),
+                                "datlocprovider, daticulocale, ");
+       snprintf(query + strlen(query), sizeof(query) - strlen(query),
                         "pg_catalog.pg_tablespace_location(t.oid) AS 
spclocation "
                         "FROM pg_catalog.pg_database d "
                         " LEFT OUTER JOIN pg_catalog.pg_tablespace t "
@@ -331,6 +340,8 @@ get_db_infos(ClusterInfo *cluster)
        i_encoding = PQfnumber(res, "encoding");
        i_datcollate = PQfnumber(res, "datcollate");
        i_datctype = PQfnumber(res, "datctype");
+       i_datlocprovider = PQfnumber(res, "datlocprovider");
+       i_daticulocale = PQfnumber(res, "daticulocale");
        i_spclocation = PQfnumber(res, "spclocation");
 
        ntups = PQntuples(res);
@@ -343,6 +354,11 @@ get_db_infos(ClusterInfo *cluster)
                dbinfos[tupnum].db_encoding = atoi(PQgetvalue(res, tupnum, 
i_encoding));
                dbinfos[tupnum].db_collate = pg_strdup(PQgetvalue(res, tupnum, 
i_datcollate));
                dbinfos[tupnum].db_ctype = pg_strdup(PQgetvalue(res, tupnum, 
i_datctype));
+               dbinfos[tupnum].db_collprovider = PQgetvalue(res, tupnum, 
i_datlocprovider)[0];
+               if (PQgetisnull(res, tupnum, i_daticulocale))
+                       dbinfos[tupnum].db_iculocale = NULL;
+               else
+                       dbinfos[tupnum].db_iculocale = 
pg_strdup(PQgetvalue(res, tupnum, i_daticulocale));
                snprintf(dbinfos[tupnum].db_tablespace, 
sizeof(dbinfos[tupnum].db_tablespace), "%s",
                                 PQgetvalue(res, tupnum, i_spclocation));
        }
diff --git a/src/bin/pg_upgrade/pg_upgrade.h b/src/bin/pg_upgrade/pg_upgrade.h
index b9b3ac81b2..6d7fd88c0c 100644
--- a/src/bin/pg_upgrade/pg_upgrade.h
+++ b/src/bin/pg_upgrade/pg_upgrade.h
@@ -171,6 +171,8 @@ typedef struct
                                                                                
         * path */
        char       *db_collate;
        char       *db_ctype;
+       char            db_collprovider;
+       char       *db_iculocale;
        int                     db_encoding;
        RelInfoArr      rel_arr;                /* array of all user relinfos */
 } DbInfo;
diff --git a/src/bin/psql/describe.c b/src/bin/psql/describe.c
index 9229eacb6d..991bfc1546 100644
--- a/src/bin/psql/describe.c
+++ b/src/bin/psql/describe.c
@@ -896,6 +896,18 @@ listAllDbs(const char *pattern, bool verbose)
                                          gettext_noop("Encoding"),
                                          gettext_noop("Collate"),
                                          gettext_noop("Ctype"));
+       if (pset.sversion >= 150000)
+               appendPQExpBuffer(&buf,
+                                                 "       d.daticulocale as 
\"%s\",\n"
+                                                 "       CASE d.datlocprovider 
WHEN 'c' THEN 'libc' WHEN 'i' THEN 'icu' END AS \"%s\",\n",
+                                                 gettext_noop("ICU Locale"),
+                                                 gettext_noop("Locale 
Provider"));
+       else
+               appendPQExpBuffer(&buf,
+                                                 "       d.datcollate as 
\"%s\",\n"
+                                                 "       'libc' AS \"%s\",\n",
+                                                 gettext_noop("ICU Locale"),
+                                                 gettext_noop("Locale 
Provider"));
        appendPQExpBufferStr(&buf, "       ");
        printACLColumn(&buf, "d.datacl");
        if (verbose)
@@ -4625,7 +4637,7 @@ listCollations(const char *pattern, bool verbose, bool 
showSystem)
        PQExpBufferData buf;
        PGresult   *res;
        printQueryOpt myopt = pset.popt;
-       static const bool translate_columns[] = {false, false, false, false, 
false, true, false};
+       static const bool translate_columns[] = {false, false, false, false, 
false, false, true, false};
 
        initPQExpBuffer(&buf);
 
@@ -4639,6 +4651,15 @@ listCollations(const char *pattern, bool verbose, bool 
showSystem)
                                          gettext_noop("Collate"),
                                          gettext_noop("Ctype"));
 
+       if (pset.sversion >= 150000)
+               appendPQExpBuffer(&buf,
+                                                 ",\n       c.colliculocale AS 
\"%s\"",
+                                                 gettext_noop("ICU Locale"));
+       else
+               appendPQExpBuffer(&buf,
+                                                 ",\n       c.collcollate AS 
\"%s\"",
+                                                 gettext_noop("ICU Locale"));
+
        if (pset.sversion >= 100000)
                appendPQExpBuffer(&buf,
                                                  ",\n       CASE 
c.collprovider WHEN 'd' THEN 'default' WHEN 'c' THEN 'libc' WHEN 'i' THEN 'icu' 
END AS \"%s\"",
diff --git a/src/bin/psql/tab-complete.c b/src/bin/psql/tab-complete.c
index 17172827a9..380cbc0b1f 100644
--- a/src/bin/psql/tab-complete.c
+++ b/src/bin/psql/tab-complete.c
@@ -2738,7 +2738,8 @@ psql_completion(const char *text, int start, int end)
                COMPLETE_WITH("OWNER", "TEMPLATE", "ENCODING", "TABLESPACE",
                                          "IS_TEMPLATE",
                                          "ALLOW_CONNECTIONS", "CONNECTION 
LIMIT",
-                                         "LC_COLLATE", "LC_CTYPE", "LOCALE", 
"OID");
+                                         "LC_COLLATE", "LC_CTYPE", "LOCALE", 
"OID",
+                                         "LOCALE_PROVIDER", "ICU_LOCALE");
 
        else if (Matches("CREATE", "DATABASE", MatchAny, "TEMPLATE"))
                COMPLETE_WITH_QUERY(Query_for_list_of_template_databases);
diff --git a/src/bin/scripts/Makefile b/src/bin/scripts/Makefile
index b833109da6..25e7da3d3f 100644
--- a/src/bin/scripts/Makefile
+++ b/src/bin/scripts/Makefile
@@ -53,6 +53,8 @@ clean distclean maintainer-clean:
        rm -f common.o $(WIN32RES)
        rm -rf tmp_check
 
+export with_icu
+
 check:
        $(prove_check)
 
diff --git a/src/bin/scripts/createdb.c b/src/bin/scripts/createdb.c
index b0c6805bc9..6f612abf7c 100644
--- a/src/bin/scripts/createdb.c
+++ b/src/bin/scripts/createdb.c
@@ -38,6 +38,8 @@ main(int argc, char *argv[])
                {"lc-ctype", required_argument, NULL, 2},
                {"locale", required_argument, NULL, 'l'},
                {"maintenance-db", required_argument, NULL, 3},
+               {"locale-provider", required_argument, NULL, 4},
+               {"icu-locale", required_argument, NULL, 5},
                {NULL, 0, NULL, 0}
        };
 
@@ -61,6 +63,8 @@ main(int argc, char *argv[])
        char       *lc_collate = NULL;
        char       *lc_ctype = NULL;
        char       *locale = NULL;
+       char       *locale_provider = NULL;
+       char       *icu_locale = NULL;
 
        PQExpBufferData sql;
 
@@ -119,6 +123,12 @@ main(int argc, char *argv[])
                        case 3:
                                maintenance_db = pg_strdup(optarg);
                                break;
+                       case 4:
+                               locale_provider = pg_strdup(optarg);
+                               break;
+                       case 5:
+                               icu_locale = pg_strdup(optarg);
+                               break;
                        default:
                                fprintf(stderr, _("Try \"%s --help\" for more 
information.\n"), progname);
                                exit(1);
@@ -217,6 +227,13 @@ main(int argc, char *argv[])
                appendPQExpBufferStr(&sql, " LC_CTYPE ");
                appendStringLiteralConn(&sql, lc_ctype, conn);
        }
+       if (locale_provider)
+               appendPQExpBuffer(&sql, " LOCALE_PROVIDER %s", locale_provider);
+       if (icu_locale)
+       {
+               appendPQExpBufferStr(&sql, " ICU_LOCALE ");
+               appendStringLiteralConn(&sql, icu_locale, conn);
+       }
 
        appendPQExpBufferChar(&sql, ';');
 
@@ -273,6 +290,9 @@ help(const char *progname)
        printf(_("  -l, --locale=LOCALE          locale settings for the 
database\n"));
        printf(_("      --lc-collate=LOCALE      LC_COLLATE setting for the 
database\n"));
        printf(_("      --lc-ctype=LOCALE        LC_CTYPE setting for the 
database\n"));
+       printf(_("      --icu-locale=LOCALE      ICU locale setting for the 
database\n"));
+       printf(_("      --locale-provider={libc|icu}\n"
+                        "                               locale provider for 
the database's default collation\n"));
        printf(_("  -O, --owner=OWNER            database user to own the new 
database\n"));
        printf(_("  -T, --template=TEMPLATE      template database to copy\n"));
        printf(_("  -V, --version                output version information, 
then exit\n"));
diff --git a/src/bin/scripts/t/020_createdb.pl 
b/src/bin/scripts/t/020_createdb.pl
index 639245466e..35deec9a92 100644
--- a/src/bin/scripts/t/020_createdb.pl
+++ b/src/bin/scripts/t/020_createdb.pl
@@ -25,9 +25,37 @@
        qr/statement: CREATE DATABASE foobar2 ENCODING 'LATIN1'/,
        'create database with encoding');
 
+if ($ENV{with_icu} eq 'yes')
+{
+       # This fails because template0 uses libc provider and has no ICU
+       # locale set.  It would succeed if template0 used the icu
+       # provider.  XXX Maybe split into multiple tests?
+       $node->command_fails(
+               [ 'createdb', '-T', 'template0', '--locale-provider=icu', 
'foobar4' ],
+               'create database with ICU fails without ICU locale specified');
+
+       $node->issues_sql_like(
+               [ 'createdb', '-T', 'template0', '--locale-provider=icu', 
'--icu-locale=en', 'foobar5' ],
+               qr/statement: CREATE DATABASE foobar5 .* LOCALE_PROVIDER icu 
ICU_LOCALE 'en'/,
+               'create database with ICU locale specified');
+
+       $node->command_fails(
+               [ 'createdb', '-T', 'template0', '--locale-provider=icu', 
'--icu-locale=@colNumeric=lower', 'foobarX' ],
+               'fails for invalid ICU locale');
+}
+else
+{
+       $node->command_fails(
+               [ 'createdb', '-T', 'template0', '--locale-provider=icu', 
'foobar4' ],
+               'create database with ICU fails since no ICU support');
+}
+
 $node->command_fails([ 'createdb', 'foobar1' ],
        'fails if database already exists');
 
+$node->command_fails([ 'createdb', '-T', 'template0', '--locale-provider=xyz', 
'foobarX' ],
+       'fails for invalid locale provider');
+
 # Check use of templates with shared dependencies copied from the template.
 my ($ret, $stdout, $stderr) = $node->psql(
        'foobar2',
diff --git a/src/include/catalog/pg_collation.dat 
b/src/include/catalog/pg_collation.dat
index 4b56825d82..f7470ead49 100644
--- a/src/include/catalog/pg_collation.dat
+++ b/src/include/catalog/pg_collation.dat
@@ -14,8 +14,7 @@
 
 { oid => '100', oid_symbol => 'DEFAULT_COLLATION_OID',
   descr => 'database\'s default collation',
-  collname => 'default', collprovider => 'd', collencoding => '-1',
-  collcollate => '', collctype => '' },
+  collname => 'default', collprovider => 'd', collencoding => '-1' },
 { oid => '950', oid_symbol => 'C_COLLATION_OID',
   descr => 'standard C collation',
   collname => 'C', collprovider => 'c', collencoding => '-1',
diff --git a/src/include/catalog/pg_collation.h 
b/src/include/catalog/pg_collation.h
index 8763dd4080..c642c3bb95 100644
--- a/src/include/catalog/pg_collation.h
+++ b/src/include/catalog/pg_collation.h
@@ -40,8 +40,9 @@ CATALOG(pg_collation,3456,CollationRelationId)
        bool            collisdeterministic BKI_DEFAULT(t);
        int32           collencoding;   /* encoding for this collation; -1 = 
"all" */
 #ifdef CATALOG_VARLEN                  /* variable-length fields start here */
-       text            collcollate BKI_FORCE_NOT_NULL;         /* LC_COLLATE 
setting */
-       text            collctype BKI_FORCE_NOT_NULL;           /* LC_CTYPE 
setting */
+       text            collcollate BKI_DEFAULT(_null_);        /* LC_COLLATE 
setting */
+       text            collctype BKI_DEFAULT(_null_);          /* LC_CTYPE 
setting */
+       text            colliculocale BKI_DEFAULT(_null_);      /* ICU locale 
ID */
        text            collversion BKI_DEFAULT(_null_);        /* 
provider-dependent
                                                                                
                         * version of collation
                                                                                
                         * data */
@@ -66,6 +67,20 @@ DECLARE_UNIQUE_INDEX_PKEY(pg_collation_oid_index, 3085, 
CollationOidIndexId, on
 #define COLLPROVIDER_ICU               'i'
 #define COLLPROVIDER_LIBC              'c'
 
+static inline const char *
+collprovider_name(char c)
+{
+       switch (c)
+       {
+               case COLLPROVIDER_ICU:
+                       return "icu";
+               case COLLPROVIDER_LIBC:
+                       return "libc";
+               default:
+                       return "???";
+       }
+}
+
 #endif                                                 /* 
EXPOSE_TO_CLIENT_CODE */
 
 
@@ -75,6 +90,7 @@ extern Oid    CollationCreate(const char *collname, Oid 
collnamespace,
                                                        bool 
collisdeterministic,
                                                        int32 collencoding,
                                                        const char 
*collcollate, const char *collctype,
+                                                       const char 
*colliculocale,
                                                        const char *collversion,
                                                        bool if_not_exists,
                                                        bool quiet);
diff --git a/src/include/catalog/pg_database.dat 
b/src/include/catalog/pg_database.dat
index e7e42d6023..5feedff7bf 100644
--- a/src/include/catalog/pg_database.dat
+++ b/src/include/catalog/pg_database.dat
@@ -14,9 +14,9 @@
 
 { oid => '1', oid_symbol => 'TemplateDbOid',
   descr => 'default template for new databases',
-  datname => 'template1', encoding => 'ENCODING', datistemplate => 't',
+  datname => 'template1', encoding => 'ENCODING', datlocprovider => 
'LOCALE_PROVIDER', datistemplate => 't',
   datallowconn => 't', datconnlimit => '-1', datfrozenxid => '0',
   datminmxid => '1', dattablespace => 'pg_default', datcollate => 'LC_COLLATE',
-  datctype => 'LC_CTYPE', datacl => '_null_' },
+  datctype => 'LC_CTYPE', daticulocale => 'ICU_LOCALE', datacl => '_null_' },
 
 ]
diff --git a/src/include/catalog/pg_database.h 
b/src/include/catalog/pg_database.h
index 76adbd4aad..a9f4a8071f 100644
--- a/src/include/catalog/pg_database.h
+++ b/src/include/catalog/pg_database.h
@@ -40,6 +40,9 @@ CATALOG(pg_database,1262,DatabaseRelationId) 
BKI_SHARED_RELATION BKI_ROWTYPE_OID
        /* character encoding */
        int32           encoding;
 
+       /* locale provider, see pg_collation.collprovider */
+       char            datlocprovider;
+
        /* allowed as CREATE DATABASE template? */
        bool            datistemplate;
 
@@ -65,6 +68,9 @@ CATALOG(pg_database,1262,DatabaseRelationId) 
BKI_SHARED_RELATION BKI_ROWTYPE_OID
        /* LC_CTYPE setting */
        text            datctype BKI_FORCE_NOT_NULL;
 
+       /* ICU locale ID */
+       text            daticulocale;
+
        /* provider-dependent version of collation data */
        text            datcollversion BKI_DEFAULT(_null_);
 
diff --git a/src/include/utils/pg_locale.h b/src/include/utils/pg_locale.h
index 30e423af0e..9b158f24a0 100644
--- a/src/include/utils/pg_locale.h
+++ b/src/include/utils/pg_locale.h
@@ -103,6 +103,11 @@ struct pg_locale_struct
 
 typedef struct pg_locale_struct *pg_locale_t;
 
+extern struct pg_locale_struct default_locale;
+
+extern void make_icu_collator(const char *iculocstr,
+                                                         struct 
pg_locale_struct *resultp);
+
 extern pg_locale_t pg_newlocale_from_collation(Oid collid);
 
 extern char *get_collation_actual_version(char collprovider, const char 
*collcollate);
diff --git a/src/test/Makefile b/src/test/Makefile
index 46275915ff..69ef074d75 100644
--- a/src/test/Makefile
+++ b/src/test/Makefile
@@ -14,6 +14,10 @@ include $(top_builddir)/src/Makefile.global
 
 SUBDIRS = perl regress isolation modules authentication recovery subscription
 
+ifeq ($(with_icu),yes)
+SUBDIRS += icu
+endif
+
 # Test suites that are not safe by default but can be run if selected
 # by the user via the whitespace-separated list in variable
 # PG_TEST_EXTRA:
@@ -37,7 +41,7 @@ endif
 # clean" etc to recurse into them.  (We must filter out those that we
 # have conditionally included into SUBDIRS above, else there will be
 # make confusion.)
-ALWAYS_SUBDIRS = $(filter-out $(SUBDIRS),examples kerberos ldap ssl)
+ALWAYS_SUBDIRS = $(filter-out $(SUBDIRS),examples kerberos icu ldap ssl)
 
 # We want to recurse to all subdirs for all standard targets, except that
 # installcheck and install should not recurse into the subdirectory "modules".
diff --git a/src/test/icu/.gitignore b/src/test/icu/.gitignore
new file mode 100644
index 0000000000..871e943d50
--- /dev/null
+++ b/src/test/icu/.gitignore
@@ -0,0 +1,2 @@
+# Generated by test suite
+/tmp_check/
diff --git a/src/test/icu/Makefile b/src/test/icu/Makefile
new file mode 100644
index 0000000000..e30f5e9524
--- /dev/null
+++ b/src/test/icu/Makefile
@@ -0,0 +1,25 @@
+#-------------------------------------------------------------------------
+#
+# Makefile for src/test/icu
+#
+# Portions Copyright (c) 1996-2022, PostgreSQL Global Development Group
+# Portions Copyright (c) 1994, Regents of the University of California
+#
+# src/test/icu/Makefile
+#
+#-------------------------------------------------------------------------
+
+subdir = src/test/icu
+top_builddir = ../../..
+include $(top_builddir)/src/Makefile.global
+
+export with_icu
+
+check:
+       $(prove_check)
+
+installcheck:
+       $(prove_installcheck)
+
+clean distclean maintainer-clean:
+       rm -rf tmp_check
diff --git a/src/test/icu/README b/src/test/icu/README
new file mode 100644
index 0000000000..cfc9353dff
--- /dev/null
+++ b/src/test/icu/README
@@ -0,0 +1,27 @@
+src/test/icu/README
+
+Regression tests for ICU functionality
+======================================
+
+This directory contains a test suite for ICU functionality.
+
+Running the tests
+=================
+
+NOTE: You must have given the --enable-tap-tests argument to configure.
+Also, to use "make installcheck", you must have built and installed
+contrib/hstore in addition to the core code.
+
+Run
+    make check
+or
+    make installcheck
+You can use "make installcheck" if you previously did "make install".
+In that case, the code in the installation tree is tested.  With
+"make check", a temporary installation tree is built from the current
+sources and then tested.
+
+Either way, this test initializes, starts, and stops several test Postgres
+clusters.
+
+See src/test/perl/README for more info about running these tests.
diff --git a/src/test/icu/t/010_database.pl b/src/test/icu/t/010_database.pl
new file mode 100644
index 0000000000..4cc8907b42
--- /dev/null
+++ b/src/test/icu/t/010_database.pl
@@ -0,0 +1,58 @@
+# Copyright (c) 2022, PostgreSQL Global Development Group
+
+use strict;
+use warnings;
+use PostgreSQL::Test::Cluster;
+use PostgreSQL::Test::Utils;
+use Test::More;
+
+if ($ENV{with_icu} ne 'yes')
+{
+       plan skip_all => 'ICU not supported by this build';
+}
+
+my $node1 = PostgreSQL::Test::Cluster->new('node1');
+$node1->init;
+$node1->start;
+
+$node1->safe_psql('postgres',
+       q{CREATE DATABASE dbicu LOCALE_PROVIDER icu LOCALE 'C' ICU_LOCALE 
'en-u-kf-upper' TEMPLATE template0});
+
+$node1->safe_psql('dbicu',
+q{
+CREATE COLLATION upperfirst (provider = icu, locale = 'en-u-kf-upper');
+CREATE TABLE icu (def text, en text COLLATE "en-x-icu", upfirst text COLLATE 
upperfirst);
+INSERT INTO icu VALUES ('a', 'a', 'a'), ('b', 'b', 'b'), ('A', 'A', 'A'), 
('B', 'B', 'B');
+});
+
+is($node1->safe_psql('dbicu', q{SELECT def FROM icu ORDER BY def}),
+       qq(A
+a
+B
+b),
+       'sort by database default locale');
+
+is($node1->safe_psql('dbicu', q{SELECT def FROM icu ORDER BY def COLLATE 
"en-x-icu"}),
+       qq(a
+A
+b
+B),
+       'sort by explicit collation standard');
+
+is($node1->safe_psql('dbicu', q{SELECT def FROM icu ORDER BY en COLLATE 
upperfirst}),
+       qq(A
+a
+B
+b),
+       'sort by explicit collation upper first');
+
+
+# Test error cases in CREATE DATABASE involving locale-related options
+
+my ($ret, $stdout, $stderr) = $node1->psql('postgres',
+       q{CREATE DATABASE dbicu LOCALE_PROVIDER icu TEMPLATE template0});
+isnt($ret, 0, "ICU locale must be specified for ICU provider: exit code not 
0");
+like($stderr, qr/ERROR:  ICU locale must be specified/, "ICU locale must be 
specified for ICU provider: error message");
+
+
+done_testing();
diff --git a/src/test/regress/expected/collate.icu.utf8.out 
b/src/test/regress/expected/collate.icu.utf8.out
index 9699ca16cf..d4c8c6de38 100644
--- a/src/test/regress/expected/collate.icu.utf8.out
+++ b/src/test/regress/expected/collate.icu.utf8.out
@@ -1029,14 +1029,12 @@ CREATE COLLATION test0 FROM "C"; -- fail, duplicate name
 ERROR:  collation "test0" already exists
 do $$
 BEGIN
-  EXECUTE 'CREATE COLLATION test1 (provider = icu, lc_collate = ' ||
-          quote_literal(current_setting('lc_collate')) ||
-          ', lc_ctype = ' ||
-          quote_literal(current_setting('lc_ctype')) || ');';
+  EXECUTE 'CREATE COLLATION test1 (provider = icu, locale = ' ||
+          quote_literal(current_setting('lc_collate')) || ');';
 END
 $$;
-CREATE COLLATION test3 (provider = icu, lc_collate = 'en_US.utf8'); -- fail, 
need lc_ctype
-ERROR:  parameter "lc_ctype" must be specified
+CREATE COLLATION test3 (provider = icu, lc_collate = 'en_US.utf8'); -- fail, 
needs "locale"
+ERROR:  parameter "locale" must be specified
 CREATE COLLATION testx (provider = icu, locale = 'nonsense'); /* never fails 
with ICU */  DROP COLLATION testx;
 CREATE COLLATION test4 FROM nonsense;
 ERROR:  collation "nonsense" for encoding "UTF8" does not exist
diff --git a/src/test/regress/sql/collate.icu.utf8.sql 
b/src/test/regress/sql/collate.icu.utf8.sql
index 242a7ce6b7..b0ddc7db44 100644
--- a/src/test/regress/sql/collate.icu.utf8.sql
+++ b/src/test/regress/sql/collate.icu.utf8.sql
@@ -366,13 +366,11 @@ CREATE SCHEMA test_schema;
 CREATE COLLATION test0 FROM "C"; -- fail, duplicate name
 do $$
 BEGIN
-  EXECUTE 'CREATE COLLATION test1 (provider = icu, lc_collate = ' ||
-          quote_literal(current_setting('lc_collate')) ||
-          ', lc_ctype = ' ||
-          quote_literal(current_setting('lc_ctype')) || ');';
+  EXECUTE 'CREATE COLLATION test1 (provider = icu, locale = ' ||
+          quote_literal(current_setting('lc_collate')) || ');';
 END
 $$;
-CREATE COLLATION test3 (provider = icu, lc_collate = 'en_US.utf8'); -- fail, 
need lc_ctype
+CREATE COLLATION test3 (provider = icu, lc_collate = 'en_US.utf8'); -- fail, 
needs "locale"
 CREATE COLLATION testx (provider = icu, locale = 'nonsense'); /* never fails 
with ICU */  DROP COLLATION testx;
 
 CREATE COLLATION test4 FROM nonsense;

base-commit: 705e20f8550c0e8e47c0b6b20b5f5ffd6ffd9e33
-- 
2.35.1

Reply via email to