|
1 | | -<!-- $PostgreSQL: pgsql/doc/src/sgml/release-9.0.sgml,v 2.52 2010/08/24 23:45:32 tgl Exp $ --> |
| 1 | +<!-- $PostgreSQL: pgsql/doc/src/sgml/release-9.0.sgml,v 2.53 2010/08/25 00:47:57 tgl Exp $ --> |
2 | 2 |
|
3 | 3 | <sect1 id="release-9-0"> |
4 | 4 | <title>Release 9.0</title> |
|
8 | 8 | <simpara>2010-??-??</simpara> |
9 | 9 | </note> |
10 | 10 |
|
11 | | - <para>CURRENT AS OF 2010-07-29</para> |
| 11 | + <para>CURRENT AS OF 2010-08-24</para> |
12 | 12 |
|
13 | 13 | <sect2> |
14 | 14 | <title>Overview</title> |
|
276 | 276 | </para> |
277 | 277 | </listitem> |
278 | 278 |
|
| 279 | + <listitem> |
| 280 | + <para> |
| 281 | + Array input now considers only plain ASCII whitespace characters |
| 282 | + to be potentially ignorable; it will never ignore non-ASCII characters, |
| 283 | + even if they are whitespace according to some locales (Tom Lane) |
| 284 | + </para> |
| 285 | + |
| 286 | + <para> |
| 287 | + This avoids some corner cases where array values could be interpreted |
| 288 | + differently depending on the server's locale settings. |
| 289 | + </para> |
| 290 | + </listitem> |
| 291 | + |
279 | 292 | <listitem> |
280 | 293 | <para> |
281 | 294 | Improve standards compliance of <link |
|
1115 | 1128 | </para> |
1116 | 1129 | </listitem> |
1117 | 1130 |
|
| 1131 | + <listitem> |
| 1132 | + <para> |
| 1133 | + Fix failure of <literal>ALTER TABLE <replaceable>table</> ADD COLUMN |
| 1134 | + <replaceable>col</> serial</literal> when done by non-owner of table |
| 1135 | + (Tom Lane) |
| 1136 | + </para> |
| 1137 | + </listitem> |
| 1138 | + |
1118 | 1139 | </itemizedlist> |
1119 | 1140 |
|
1120 | 1141 | </sect4> |
|
1688 | 1709 | <listitem> |
1689 | 1710 | <para> |
1690 | 1711 | Allow aggregate functions to use <link |
1691 | | - linkend="syntax-aggregates"><literal>ORDER BY</></> (Andrew Gierth) |
| 1712 | + linkend="syntax-aggregates"><literal>ORDER BY</></link> (Andrew Gierth) |
1692 | 1713 | </para> |
1693 | 1714 |
|
1694 | 1715 | <para> |
@@ -2379,6 +2400,17 @@ if TG_OP = 'INSERT' AND NEW.col1 = ... then |
2379 | 2400 | </para> |
2380 | 2401 | </listitem> |
2381 | 2402 |
|
| 2403 | + <listitem> |
| 2404 | + <para> |
| 2405 | + <application>pg_restore</> now complains if any command-line arguments |
| 2406 | + remain after the switches and optional file name (Tom Lane) |
| 2407 | + </para> |
| 2408 | + |
| 2409 | + <para> |
| 2410 | + Previously, it silently ignored any such arguments. |
| 2411 | + </para> |
| 2412 | + </listitem> |
| 2413 | + |
2382 | 2414 | </itemizedlist> |
2383 | 2415 |
|
2384 | 2416 | </sect4> |
@@ -3067,6 +3099,21 @@ if TG_OP = 'INSERT' AND NEW.col1 = ... then |
3067 | 3099 | </para> |
3068 | 3100 | </listitem> |
3069 | 3101 |
|
| 3102 | + <listitem> |
| 3103 | + <para> |
| 3104 | + Custom typanalyze functions should no longer rely on |
| 3105 | + <structname>VacAttrStats</>.<structfield>attr</> to determine the type |
| 3106 | + of data they will be passed (Tom Lane) |
| 3107 | + </para> |
| 3108 | + |
| 3109 | + <para> |
| 3110 | + This was changed to allow collection of statistics on index columns |
| 3111 | + for which the storage type is different from the underlying column |
| 3112 | + data type. There are new fields that tell the actual datatype being |
| 3113 | + analyzed. |
| 3114 | + </para> |
| 3115 | + </listitem> |
| 3116 | + |
3070 | 3117 | </itemizedlist> |
3071 | 3118 |
|
3072 | 3119 | </sect4> |
|