| | 69 | |
|---|
| | 70 | <sect1 id="release-8-3-4"> |
|---|
| | 71 | <title>Release 8.3.4</title> |
|---|
| | 72 | |
|---|
| | 73 | <note> |
|---|
| | 74 | <title>Release date</title> |
|---|
| | 75 | <simpara>2008-09-22</simpara> |
|---|
| | 76 | </note> |
|---|
| | 77 | |
|---|
| | 78 | <para> |
|---|
| | 79 | This release contains a variety of fixes from 8.3.3. |
|---|
| | 80 | For information about new features in the 8.3 major release, see |
|---|
| | 81 | <xref linkend="release-8-3"/>. |
|---|
| | 82 | </para> |
|---|
| | 83 | |
|---|
| | 84 | <sect2> |
|---|
| | 85 | <title>Migration to Version 8.3.4</title> |
|---|
| | 86 | |
|---|
| | 87 | <para> |
|---|
| | 88 | A dump/restore is not required for those running 8.3.X. |
|---|
| | 89 | However, if you are upgrading from a version earlier than 8.3.1, |
|---|
| | 90 | see the release notes for 8.3.1. |
|---|
| | 91 | </para> |
|---|
| | 92 | |
|---|
| | 93 | </sect2> |
|---|
| | 94 | |
|---|
| | 95 | <sect2> |
|---|
| | 96 | <title>Changes</title> |
|---|
| | 97 | |
|---|
| | 98 | <itemizedlist> |
|---|
| | 99 | |
|---|
| | 100 | <listitem> |
|---|
| | 101 | <para> |
|---|
| | 102 | Fix bug in btree WAL recovery code (Heikki) |
|---|
| | 103 | </para> |
|---|
| | 104 | |
|---|
| | 105 | <para> |
|---|
| | 106 | Recovery failed if the WAL ended partway through a page split operation. |
|---|
| | 107 | </para> |
|---|
| | 108 | </listitem> |
|---|
| | 109 | |
|---|
| | 110 | <listitem> |
|---|
| | 111 | <para> |
|---|
| | 112 | Fix potential use of wrong cutoff XID for HOT page pruning (Alvaro) |
|---|
| | 113 | </para> |
|---|
| | 114 | |
|---|
| | 115 | <para> |
|---|
| | 116 | This error created a risk of corruption in system |
|---|
| | 117 | catalogs that are consulted by <command>VACUUM</command>: dead tuple versions |
|---|
| | 118 | might be removed too soon. The impact of this on actual database |
|---|
| | 119 | operations would be minimal, since the system doesn't follow MVCC |
|---|
| | 120 | rules while examining catalogs, but it might result in transiently |
|---|
| | 121 | wrong output from <application>pg_dump</application> or other client programs. |
|---|
| | 122 | </para> |
|---|
| | 123 | </listitem> |
|---|
| | 124 | |
|---|
| | 125 | <listitem> |
|---|
| | 126 | <para> |
|---|
| | 127 | Fix potential miscalculation of <structfield>datfrozenxid</structfield> (Alvaro) |
|---|
| | 128 | </para> |
|---|
| | 129 | |
|---|
| | 130 | <para> |
|---|
| | 131 | This error may explain some recent reports of failure to remove old |
|---|
| | 132 | <structname>pg_clog</structname> data. |
|---|
| | 133 | </para> |
|---|
| | 134 | </listitem> |
|---|
| | 135 | |
|---|
| | 136 | <listitem> |
|---|
| | 137 | <para> |
|---|
| | 138 | Fix incorrect HOT updates after <structname>pg_class</structname> is reindexed |
|---|
| | 139 | (Tom) |
|---|
| | 140 | </para> |
|---|
| | 141 | |
|---|
| | 142 | <para> |
|---|
| | 143 | Corruption of <structname>pg_class</structname> could occur if <literal>REINDEX |
|---|
| | 144 | TABLE pg_class</literal> was followed in the same session by an <literal>ALTER |
|---|
| | 145 | TABLE RENAME</literal> or <literal>ALTER TABLE SET SCHEMA</literal> command. |
|---|
| | 146 | </para> |
|---|
| | 147 | </listitem> |
|---|
| | 148 | |
|---|
| | 149 | <listitem> |
|---|
| | 150 | <para> |
|---|
| | 151 | Fix missed <quote>combo cid</quote> case (Karl Schnaitter) |
|---|
| | 152 | </para> |
|---|
| | 153 | |
|---|
| | 154 | <para> |
|---|
| | 155 | This error made rows incorrectly invisible to a transaction in which they |
|---|
| | 156 | had been deleted by multiple subtransactions that all aborted. |
|---|
| | 157 | </para> |
|---|
| | 158 | </listitem> |
|---|
| | 159 | |
|---|
| | 160 | <listitem> |
|---|
| | 161 | <para> |
|---|
| | 162 | Prevent autovacuum from crashing if the table it's currently |
|---|
| | 163 | checking is deleted at just the wrong time (Alvaro) |
|---|
| | 164 | </para> |
|---|
| | 165 | </listitem> |
|---|
| | 166 | |
|---|
| | 167 | <listitem> |
|---|
| | 168 | <para> |
|---|
| | 169 | Widen local lock counters from 32 to 64 bits (Tom) |
|---|
| | 170 | </para> |
|---|
| | 171 | |
|---|
| | 172 | <para> |
|---|
| | 173 | This responds to reports that the counters could overflow in |
|---|
| | 174 | sufficiently long transactions, leading to unexpected <quote>lock is |
|---|
| | 175 | already held</quote> errors. |
|---|
| | 176 | </para> |
|---|
| | 177 | </listitem> |
|---|
| | 178 | |
|---|
| | 179 | <listitem> |
|---|
| | 180 | <para> |
|---|
| | 181 | Fix possible duplicate output of tuples during a GiST index scan (Teodor) |
|---|
| | 182 | </para> |
|---|
| | 183 | </listitem> |
|---|
| | 184 | |
|---|
| | 185 | <listitem> |
|---|
| | 186 | <para> |
|---|
| | 187 | Regenerate foreign key checking queries from scratch when either |
|---|
| | 188 | table is modified (Tom) |
|---|
| | 189 | </para> |
|---|
| | 190 | |
|---|
| | 191 | <para> |
|---|
| | 192 | Previously, 8.3 would attempt to replan the query, but would work from |
|---|
| | 193 | previously generated query text. This led to failures if a |
|---|
| | 194 | table or column was renamed. |
|---|
| | 195 | </para> |
|---|
| | 196 | </listitem> |
|---|
| | 197 | |
|---|
| | 198 | <listitem> |
|---|
| | 199 | <para> |
|---|
| | 200 | Fix missed permissions checks when a view contains a simple |
|---|
| | 201 | <literal>UNION ALL</literal> construct (Heikki) |
|---|
| | 202 | </para> |
|---|
| | 203 | |
|---|
| | 204 | <para> |
|---|
| | 205 | Permissions for the referenced tables were checked properly, but not |
|---|
| | 206 | permissions for the view itself. |
|---|
| | 207 | </para> |
|---|
| | 208 | </listitem> |
|---|
| | 209 | |
|---|
| | 210 | <listitem> |
|---|
| | 211 | <para> |
|---|
| | 212 | Add checks in executor startup to ensure that the tuples produced by an |
|---|
| | 213 | <command>INSERT</command> or <command>UPDATE</command> will match the target table's |
|---|
| | 214 | current rowtype (Tom) |
|---|
| | 215 | </para> |
|---|
| | 216 | |
|---|
| | 217 | <para> |
|---|
| | 218 | This situation is believed to be impossible in 8.3, but it can happen in |
|---|
| | 219 | prior releases, so a check seems prudent. |
|---|
| | 220 | </para> |
|---|
| | 221 | </listitem> |
|---|
| | 222 | |
|---|
| | 223 | <listitem> |
|---|
| | 224 | <para> |
|---|
| | 225 | Fix possible repeated drops during <command>DROP OWNED</command> (Tom) |
|---|
| | 226 | </para> |
|---|
| | 227 | |
|---|
| | 228 | <para> |
|---|
| | 229 | This would typically result in strange errors such as <quote>cache |
|---|
| | 230 | lookup failed for relation NNN</quote>. |
|---|
| | 231 | </para> |
|---|
| | 232 | </listitem> |
|---|
| | 233 | |
|---|
| | 234 | <listitem> |
|---|
| | 235 | <para> |
|---|
| | 236 | Fix several memory leaks in XML operations (Kris Jurka, Tom) |
|---|
| | 237 | </para> |
|---|
| | 238 | </listitem> |
|---|
| | 239 | |
|---|
| | 240 | <listitem> |
|---|
| | 241 | <para> |
|---|
| | 242 | Fix <function>xmlserialize()</function> to raise error properly for |
|---|
| | 243 | unacceptable target data type (Tom) |
|---|
| | 244 | </para> |
|---|
| | 245 | </listitem> |
|---|
| | 246 | |
|---|
| | 247 | <listitem> |
|---|
| | 248 | <para> |
|---|
| | 249 | Fix a couple of places that mis-handled multibyte characters in text |
|---|
| | 250 | search configuration file parsing (Tom) |
|---|
| | 251 | </para> |
|---|
| | 252 | |
|---|
| | 253 | <para> |
|---|
| | 254 | Certain characters occurring in configuration files would always cause |
|---|
| | 255 | <quote>invalid byte sequence for encoding</quote> failures. |
|---|
| | 256 | </para> |
|---|
| | 257 | </listitem> |
|---|
| | 258 | |
|---|
| | 259 | <listitem> |
|---|
| | 260 | <para> |
|---|
| | 261 | Provide file name and line number location for all errors reported |
|---|
| | 262 | in text search configuration files (Tom) |
|---|
| | 263 | </para> |
|---|
| | 264 | </listitem> |
|---|
| | 265 | |
|---|
| | 266 | <listitem> |
|---|
| | 267 | <para> |
|---|
| | 268 | Fix <literal>AT TIME ZONE</literal> to first try to interpret its timezone |
|---|
| | 269 | argument as a timezone abbreviation, and only try it as a full timezone |
|---|
| | 270 | name if that fails, rather than the other way around as formerly (Tom) |
|---|
| | 271 | </para> |
|---|
| | 272 | |
|---|
| | 273 | <para> |
|---|
| | 274 | The timestamp input functions have always resolved ambiguous zone names |
|---|
| | 275 | in this order. Making <literal>AT TIME ZONE</literal> do so as well improves |
|---|
| | 276 | consistency, and fixes a compatibility bug introduced in 8.1: |
|---|
| | 277 | in ambiguous cases we now behave the same as 8.0 and before did, |
|---|
| | 278 | since in the older versions <literal>AT TIME ZONE</literal> accepted |
|---|
| | 279 | <emphasis>only</emphasis> abbreviations. |
|---|
| | 280 | </para> |
|---|
| | 281 | </listitem> |
|---|
| | 282 | |
|---|
| | 283 | <listitem> |
|---|
| | 284 | <para> |
|---|
| | 285 | Fix datetime input functions to correctly detect integer overflow when |
|---|
| | 286 | running on a 64-bit platform (Tom) |
|---|
| | 287 | </para> |
|---|
| | 288 | </listitem> |
|---|
| | 289 | |
|---|
| | 290 | <listitem> |
|---|
| | 291 | <para> |
|---|
| | 292 | Prevent integer overflows during units conversion when displaying a |
|---|
| | 293 | configuration parameter that has units (Tom) |
|---|
| | 294 | </para> |
|---|
| | 295 | </listitem> |
|---|
| | 296 | |
|---|
| | 297 | <listitem> |
|---|
| | 298 | <para> |
|---|
| | 299 | Improve performance of writing very long log messages to syslog (Tom) |
|---|
| | 300 | </para> |
|---|
| | 301 | </listitem> |
|---|
| | 302 | |
|---|
| | 303 | <listitem> |
|---|
| | 304 | <para> |
|---|
| | 305 | Allow spaces in the suffix part of an LDAP URL in |
|---|
| | 306 | <filename>pg_hba.conf</filename> (Tom) |
|---|
| | 307 | </para> |
|---|
| | 308 | </listitem> |
|---|
| | 309 | |
|---|
| | 310 | <listitem> |
|---|
| | 311 | <para> |
|---|
| | 312 | Fix bug in backwards scanning of a cursor on a <literal>SELECT DISTINCT |
|---|
| | 313 | ON</literal> query (Tom) |
|---|
| | 314 | </para> |
|---|
| | 315 | </listitem> |
|---|
| | 316 | |
|---|
| | 317 | <listitem> |
|---|
| | 318 | <para> |
|---|
| | 319 | Fix planner bug that could improperly push down <literal>IS NULL</literal> |
|---|
| | 320 | tests below an outer join (Tom) |
|---|
| | 321 | </para> |
|---|
| | 322 | |
|---|
| | 323 | <para> |
|---|
| | 324 | This was triggered by occurrence of <literal>IS NULL</literal> tests for |
|---|
| | 325 | the same relation in all arms of an upper <literal>OR</literal> clause. |
|---|
| | 326 | </para> |
|---|
| | 327 | </listitem> |
|---|
| | 328 | |
|---|
| | 329 | <listitem> |
|---|
| | 330 | <para> |
|---|
| | 331 | Fix planner bug with nested sub-select expressions (Tom) |
|---|
| | 332 | </para> |
|---|
| | 333 | |
|---|
| | 334 | <para> |
|---|
| | 335 | If the outer sub-select has no direct dependency on the parent query, |
|---|
| | 336 | but the inner one does, the outer value might not get recalculated |
|---|
| | 337 | for new parent query rows. |
|---|
| | 338 | </para> |
|---|
| | 339 | </listitem> |
|---|
| | 340 | |
|---|
| | 341 | <listitem> |
|---|
| | 342 | <para> |
|---|
| | 343 | Fix planner to estimate that <literal>GROUP BY</literal> expressions yielding |
|---|
| | 344 | boolean results always result in two groups, regardless of the |
|---|
| | 345 | expressions' contents (Tom) |
|---|
| | 346 | </para> |
|---|
| | 347 | |
|---|
| | 348 | <para> |
|---|
| | 349 | This is very substantially more accurate than the regular <literal>GROUP |
|---|
| | 350 | BY</literal> estimate for certain boolean tests like <replaceable>col</replaceable> |
|---|
| | 351 | <literal>IS NULL</literal>. |
|---|
| | 352 | </para> |
|---|
| | 353 | </listitem> |
|---|
| | 354 | |
|---|
| | 355 | <listitem> |
|---|
| | 356 | <para> |
|---|
| | 357 | Fix PL/PgSQL to not fail when a <literal>FOR</literal> loop's target variable |
|---|
| | 358 | is a record containing composite-type fields (Tom) |
|---|
| | 359 | </para> |
|---|
| | 360 | </listitem> |
|---|
| | 361 | |
|---|
| | 362 | <listitem> |
|---|
| | 363 | <para> |
|---|
| | 364 | Fix PL/Tcl to behave correctly with Tcl 8.5, and to be more careful |
|---|
| | 365 | about the encoding of data sent to or from Tcl (Tom) |
|---|
| | 366 | </para> |
|---|
| | 367 | </listitem> |
|---|
| | 368 | |
|---|
| | 369 | <listitem> |
|---|
| | 370 | <para> |
|---|
| | 371 | Improve performance of <function>PQescapeBytea()</function> (Rudolf Leitgeb) |
|---|
| | 372 | </para> |
|---|
| | 373 | </listitem> |
|---|
| | 374 | |
|---|
| | 375 | <listitem> |
|---|
| | 376 | <para> |
|---|
| | 377 | On Windows, work around a Microsoft bug by preventing |
|---|
| | 378 | <application>libpq</application> from trying to send more than 64kB per system call |
|---|
| | 379 | (Magnus) |
|---|
| | 380 | </para> |
|---|
| | 381 | </listitem> |
|---|
| | 382 | |
|---|
| | 383 | <listitem> |
|---|
| | 384 | <para> |
|---|
| | 385 | Fix <application>ecpg</application> to handle variables properly in <command>SET</command> |
|---|
| | 386 | commands (Michael) |
|---|
| | 387 | </para> |
|---|
| | 388 | </listitem> |
|---|
| | 389 | |
|---|
| | 390 | <listitem> |
|---|
| | 391 | <para> |
|---|
| | 392 | Improve <application>pg_dump</application> and <application>pg_restore</application>'s |
|---|
| | 393 | error reporting after failure to send a SQL command (Tom) |
|---|
| | 394 | </para> |
|---|
| | 395 | </listitem> |
|---|
| | 396 | |
|---|
| | 397 | <listitem> |
|---|
| | 398 | <para> |
|---|
| | 399 | Fix <application>pg_ctl</application> to properly preserve postmaster |
|---|
| | 400 | command-line arguments across a <literal>restart</literal> (Bruce) |
|---|
| | 401 | </para> |
|---|
| | 402 | </listitem> |
|---|
| | 403 | |
|---|
| | 404 | <listitem> |
|---|
| | 405 | <para> |
|---|
| | 406 | Fix erroneous WAL file cutoff point calculation in |
|---|
| | 407 | <application>pg_standby</application> (Simon) |
|---|
| | 408 | </para> |
|---|
| | 409 | </listitem> |
|---|
| | 410 | |
|---|
| | 411 | <listitem> |
|---|
| | 412 | <para> |
|---|
| | 413 | Update time zone data files to <application>tzdata</application> release 2008f (for |
|---|
| | 414 | DST law changes in Argentina, Bahamas, Brazil, Mauritius, Morocco, |
|---|
| | 415 | Pakistan, Palestine, and Paraguay) |
|---|
| | 416 | </para> |
|---|
| | 417 | </listitem> |
|---|
| | 418 | |
|---|
| | 419 | </itemizedlist> |
|---|
| | 420 | |
|---|
| | 421 | </sect2> |
|---|
| | 422 | </sect1> |
|---|
| | 3772 | </sect2> |
|---|
| | 3773 | </sect1> |
|---|
| | 3774 | |
|---|
| | 3775 | <sect1 id="release-8-2-10"> |
|---|
| | 3776 | <title>Release 8.2.10</title> |
|---|
| | 3777 | |
|---|
| | 3778 | <note> |
|---|
| | 3779 | <title>Release date</title> |
|---|
| | 3780 | <simpara>2008-09-22</simpara> |
|---|
| | 3781 | </note> |
|---|
| | 3782 | |
|---|
| | 3783 | <para> |
|---|
| | 3784 | This release contains a variety of fixes from 8.2.9. |
|---|
| | 3785 | For information about new features in the 8.2 major release, see |
|---|
| | 3786 | <xref linkend="release-8-2"/>. |
|---|
| | 3787 | </para> |
|---|
| | 3788 | |
|---|
| | 3789 | <sect2> |
|---|
| | 3790 | <title>Migration to Version 8.2.10</title> |
|---|
| | 3791 | |
|---|
| | 3792 | <para> |
|---|
| | 3793 | A dump/restore is not required for those running 8.2.X. |
|---|
| | 3794 | However, if you are upgrading from a version earlier than 8.2.7, |
|---|
| | 3795 | see the release notes for 8.2.7. |
|---|
| | 3796 | </para> |
|---|
| | 3797 | |
|---|
| | 3798 | </sect2> |
|---|
| | 3799 | |
|---|
| | 3800 | <sect2> |
|---|
| | 3801 | <title>Changes</title> |
|---|
| | 3802 | |
|---|
| | 3803 | <itemizedlist> |
|---|
| | 3804 | |
|---|
| | 3805 | <listitem> |
|---|
| | 3806 | <para> |
|---|
| | 3807 | Fix bug in btree WAL recovery code (Heikki) |
|---|
| | 3808 | </para> |
|---|
| | 3809 | |
|---|
| | 3810 | <para> |
|---|
| | 3811 | Recovery failed if the WAL ended partway through a page split operation. |
|---|
| | 3812 | </para> |
|---|
| | 3813 | </listitem> |
|---|
| | 3814 | |
|---|
| | 3815 | <listitem> |
|---|
| | 3816 | <para> |
|---|
| | 3817 | Fix potential miscalculation of <structfield>datfrozenxid</structfield> (Alvaro) |
|---|
| | 3818 | </para> |
|---|
| | 3819 | |
|---|
| | 3820 | <para> |
|---|
| | 3821 | This error may explain some recent reports of failure to remove old |
|---|
| | 3822 | <structname>pg_clog</structname> data. |
|---|
| | 3823 | </para> |
|---|
| | 3824 | </listitem> |
|---|
| | 3825 | |
|---|
| | 3826 | <listitem> |
|---|
| | 3827 | <para> |
|---|
| | 3828 | Widen local lock counters from 32 to 64 bits (Tom) |
|---|
| | 3829 | </para> |
|---|
| | 3830 | |
|---|
| | 3831 | <para> |
|---|
| | 3832 | This responds to reports that the counters could overflow in |
|---|
| | 3833 | sufficiently long transactions, leading to unexpected <quote>lock is |
|---|
| | 3834 | already held</quote> errors. |
|---|
| | 3835 | </para> |
|---|
| | 3836 | </listitem> |
|---|
| | 3837 | |
|---|
| | 3838 | <listitem> |
|---|
| | 3839 | <para> |
|---|
| | 3840 | Fix possible duplicate output of tuples during a GiST index scan (Teodor) |
|---|
| | 3841 | </para> |
|---|
| | 3842 | </listitem> |
|---|
| | 3843 | |
|---|
| | 3844 | <listitem> |
|---|
| | 3845 | <para> |
|---|
| | 3846 | Fix missed permissions checks when a view contains a simple |
|---|
| | 3847 | <literal>UNION ALL</literal> construct (Heikki) |
|---|
| | 3848 | </para> |
|---|
| | 3849 | |
|---|
| | 3850 | <para> |
|---|
| | 3851 | Permissions for the referenced tables were checked properly, but not |
|---|
| | 3852 | permissions for the view itself. |
|---|
| | 3853 | </para> |
|---|
| | 3854 | </listitem> |
|---|
| | 3855 | |
|---|
| | 3856 | <listitem> |
|---|
| | 3857 | <para> |
|---|
| | 3858 | Add checks in executor startup to ensure that the tuples produced by an |
|---|
| | 3859 | <command>INSERT</command> or <command>UPDATE</command> will match the target table's |
|---|
| | 3860 | current rowtype (Tom) |
|---|
| | 3861 | </para> |
|---|
| | 3862 | |
|---|
| | 3863 | <para> |
|---|
| | 3864 | <command>ALTER COLUMN TYPE</command>, followed by re-use of a previously |
|---|
| | 3865 | cached plan, could produce this type of situation. The check protects |
|---|
| | 3866 | against data corruption and/or crashes that could ensue. |
|---|
| | 3867 | </para> |
|---|
| | 3868 | </listitem> |
|---|
| | 3869 | |
|---|
| | 3870 | <listitem> |
|---|
| | 3871 | <para> |
|---|
| | 3872 | Fix possible repeated drops during <command>DROP OWNED</command> (Tom) |
|---|
| | 3873 | </para> |
|---|
| | 3874 | |
|---|
| | 3875 | <para> |
|---|
| | 3876 | This would typically result in strange errors such as <quote>cache |
|---|
| | 3877 | lookup failed for relation NNN</quote>. |
|---|
| | 3878 | </para> |
|---|
| | 3879 | </listitem> |
|---|
| | 3880 | |
|---|
| | 3881 | <listitem> |
|---|
| | 3882 | <para> |
|---|
| | 3883 | Fix <literal>AT TIME ZONE</literal> to first try to interpret its timezone |
|---|
| | 3884 | argument as a timezone abbreviation, and only try it as a full timezone |
|---|
| | 3885 | name if that fails, rather than the other way around as formerly (Tom) |
|---|
| | 3886 | </para> |
|---|
| | 3887 | |
|---|
| | 3888 | <para> |
|---|
| | 3889 | The timestamp input functions have always resolved ambiguous zone names |
|---|
| | 3890 | in this order. Making <literal>AT TIME ZONE</literal> do so as well improves |
|---|
| | 3891 | consistency, and fixes a compatibility bug introduced in 8.1: |
|---|
| | 3892 | in ambiguous cases we now behave the same as 8.0 and before did, |
|---|
| | 3893 | since in the older versions <literal>AT TIME ZONE</literal> accepted |
|---|
| | 3894 | <emphasis>only</emphasis> abbreviations. |
|---|
| | 3895 | </para> |
|---|
| | 3896 | </listitem> |
|---|
| | 3897 | |
|---|
| | 3898 | <listitem> |
|---|
| | 3899 | <para> |
|---|
| | 3900 | Fix datetime input functions to correctly detect integer overflow when |
|---|
| | 3901 | running on a 64-bit platform (Tom) |
|---|
| | 3902 | </para> |
|---|
| | 3903 | </listitem> |
|---|
| | 3904 | |
|---|
| | 3905 | <listitem> |
|---|
| | 3906 | <para> |
|---|
| | 3907 | Prevent integer overflows during units conversion when displaying a |
|---|
| | 3908 | configuration parameter that has units (Tom) |
|---|
| | 3909 | </para> |
|---|
| | 3910 | </listitem> |
|---|
| | 3911 | |
|---|
| | 3912 | <listitem> |
|---|
| | 3913 | <para> |
|---|
| | 3914 | Improve performance of writing very long log messages to syslog (Tom) |
|---|
| | 3915 | </para> |
|---|
| | 3916 | </listitem> |
|---|
| | 3917 | |
|---|
| | 3918 | <listitem> |
|---|
| | 3919 | <para> |
|---|
| | 3920 | Allow spaces in the suffix part of an LDAP URL in |
|---|
| | 3921 | <filename>pg_hba.conf</filename> (Tom) |
|---|
| | 3922 | </para> |
|---|
| | 3923 | </listitem> |
|---|
| | 3924 | |
|---|
| | 3925 | <listitem> |
|---|
| | 3926 | <para> |
|---|
| | 3927 | Fix bug in backwards scanning of a cursor on a <literal>SELECT DISTINCT |
|---|
| | 3928 | ON</literal> query (Tom) |
|---|
| | 3929 | </para> |
|---|
| | 3930 | </listitem> |
|---|
| | 3931 | |
|---|
| | 3932 | <listitem> |
|---|
| | 3933 | <para> |
|---|
| | 3934 | Fix planner bug with nested sub-select expressions (Tom) |
|---|
| | 3935 | </para> |
|---|
| | 3936 | |
|---|
| | 3937 | <para> |
|---|
| | 3938 | If the outer sub-select has no direct dependency on the parent query, |
|---|
| | 3939 | but the inner one does, the outer value might not get recalculated |
|---|
| | 3940 | for new parent query rows. |
|---|
| | 3941 | </para> |
|---|
| | 3942 | </listitem> |
|---|
| | 3943 | |
|---|
| | 3944 | <listitem> |
|---|
| | 3945 | <para> |
|---|
| | 3946 | Fix planner to estimate that <literal>GROUP BY</literal> expressions yielding |
|---|
| | 3947 | boolean results always result in two groups, regardless of the |
|---|
| | 3948 | expressions' contents (Tom) |
|---|
| | 3949 | </para> |
|---|
| | 3950 | |
|---|
| | 3951 | <para> |
|---|
| | 3952 | This is very substantially more accurate than the regular <literal>GROUP |
|---|
| | 3953 | BY</literal> estimate for certain boolean tests like <replaceable>col</replaceable> |
|---|
| | 3954 | <literal>IS NULL</literal>. |
|---|
| | 3955 | </para> |
|---|
| | 3956 | </listitem> |
|---|
| | 3957 | |
|---|
| | 3958 | <listitem> |
|---|
| | 3959 | <para> |
|---|
| | 3960 | Fix PL/PgSQL to not fail when a <literal>FOR</literal> loop's target variable |
|---|
| | 3961 | is a record containing composite-type fields (Tom) |
|---|
| | 3962 | </para> |
|---|
| | 3963 | </listitem> |
|---|
| | 3964 | |
|---|
| | 3965 | <listitem> |
|---|
| | 3966 | <para> |
|---|
| | 3967 | Fix PL/Tcl to behave correctly with Tcl 8.5, and to be more careful |
|---|
| | 3968 | about the encoding of data sent to or from Tcl (Tom) |
|---|
| | 3969 | </para> |
|---|
| | 3970 | </listitem> |
|---|
| | 3971 | |
|---|
| | 3972 | <listitem> |
|---|
| | 3973 | <para> |
|---|
| | 3974 | On Windows, work around a Microsoft bug by preventing |
|---|
| | 3975 | <application>libpq</application> from trying to send more than 64kB per system call |
|---|
| | 3976 | (Magnus) |
|---|
| | 3977 | </para> |
|---|
| | 3978 | </listitem> |
|---|
| | 3979 | |
|---|
| | 3980 | <listitem> |
|---|
| | 3981 | <para> |
|---|
| | 3982 | Improve <application>pg_dump</application> and <application>pg_restore</application>'s |
|---|
| | 3983 | error reporting after failure to send a SQL command (Tom) |
|---|
| | 3984 | </para> |
|---|
| | 3985 | </listitem> |
|---|
| | 3986 | |
|---|
| | 3987 | <listitem> |
|---|
| | 3988 | <para> |
|---|
| | 3989 | Fix <application>pg_ctl</application> to properly preserve postmaster |
|---|
| | 3990 | command-line arguments across a <literal>restart</literal> (Bruce) |
|---|
| | 3991 | </para> |
|---|
| | 3992 | </listitem> |
|---|
| | 3993 | |
|---|
| | 3994 | <listitem> |
|---|
| | 3995 | <para> |
|---|
| | 3996 | Update time zone data files to <application>tzdata</application> release 2008f (for |
|---|
| | 3997 | DST law changes in Argentina, Bahamas, Brazil, Mauritius, Morocco, |
|---|
| | 3998 | Pakistan, Palestine, and Paraguay) |
|---|
| | 3999 | </para> |
|---|
| | 4000 | </listitem> |
|---|
| | 4001 | |
|---|
| | 4002 | </itemizedlist> |
|---|
| | 4003 | |
|---|
| | 8133 | <sect1 id="release-8-1-14"> |
|---|
| | 8134 | <title>Release 8.1.14</title> |
|---|
| | 8135 | |
|---|
| | 8136 | <note> |
|---|
| | 8137 | <title>Release date</title> |
|---|
| | 8138 | <simpara>2008-09-22</simpara> |
|---|
| | 8139 | </note> |
|---|
| | 8140 | |
|---|
| | 8141 | <para> |
|---|
| | 8142 | This release contains a variety of fixes from 8.1.13. |
|---|
| | 8143 | For information about new features in the 8.1 major release, see |
|---|
| | 8144 | <xref linkend="release-8-1"/>. |
|---|
| | 8145 | </para> |
|---|
| | 8146 | |
|---|
| | 8147 | <sect2> |
|---|
| | 8148 | <title>Migration to Version 8.1.14</title> |
|---|
| | 8149 | |
|---|
| | 8150 | <para> |
|---|
| | 8151 | A dump/restore is not required for those running 8.1.X. |
|---|
| | 8152 | However, if you are upgrading from a version earlier than 8.1.2, |
|---|
| | 8153 | see the release notes for 8.1.2. |
|---|
| | 8154 | </para> |
|---|
| | 8155 | |
|---|
| | 8156 | </sect2> |
|---|
| | 8157 | |
|---|
| | 8158 | <sect2> |
|---|
| | 8159 | <title>Changes</title> |
|---|
| | 8160 | |
|---|
| | 8161 | <itemizedlist> |
|---|
| | 8162 | |
|---|
| | 8163 | <listitem> |
|---|
| | 8164 | <para> |
|---|
| | 8165 | Widen local lock counters from 32 to 64 bits (Tom) |
|---|
| | 8166 | </para> |
|---|
| | 8167 | |
|---|
| | 8168 | <para> |
|---|
| | 8169 | This responds to reports that the counters could overflow in |
|---|
| | 8170 | sufficiently long transactions, leading to unexpected <quote>lock is |
|---|
| | 8171 | already held</quote> errors. |
|---|
| | 8172 | </para> |
|---|
| | 8173 | </listitem> |
|---|
| | 8174 | |
|---|
| | 8175 | <listitem> |
|---|
| | 8176 | <para> |
|---|
| | 8177 | Fix possible duplicate output of tuples during a GiST index scan (Teodor) |
|---|
| | 8178 | </para> |
|---|
| | 8179 | </listitem> |
|---|
| | 8180 | |
|---|
| | 8181 | <listitem> |
|---|
| | 8182 | <para> |
|---|
| | 8183 | Add checks in executor startup to ensure that the tuples produced by an |
|---|
| | 8184 | <command>INSERT</command> or <command>UPDATE</command> will match the target table's |
|---|
| | 8185 | current rowtype (Tom) |
|---|
| | 8186 | </para> |
|---|
| | 8187 | |
|---|
| | 8188 | <para> |
|---|
| | 8189 | <command>ALTER COLUMN TYPE</command>, followed by re-use of a previously |
|---|
| | 8190 | cached plan, could produce this type of situation. The check protects |
|---|
| | 8191 | against data corruption and/or crashes that could ensue. |
|---|
| | 8192 | </para> |
|---|
| | 8193 | </listitem> |
|---|
| | 8194 | |
|---|
| | 8195 | <listitem> |
|---|
| | 8196 | <para> |
|---|
| | 8197 | Fix <literal>AT TIME ZONE</literal> to first try to interpret its timezone |
|---|
| | 8198 | argument as a timezone abbreviation, and only try it as a full timezone |
|---|
| | 8199 | name if that fails, rather than the other way around as formerly (Tom) |
|---|
| | 8200 | </para> |
|---|
| | 8201 | |
|---|
| | 8202 | <para> |
|---|
| | 8203 | The timestamp input functions have always resolved ambiguous zone names |
|---|
| | 8204 | in this order. Making <literal>AT TIME ZONE</literal> do so as well improves |
|---|
| | 8205 | consistency, and fixes a compatibility bug introduced in 8.1: |
|---|
| | 8206 | in ambiguous cases we now behave the same as 8.0 and before did, |
|---|
| | 8207 | since in the older versions <literal>AT TIME ZONE</literal> accepted |
|---|
| | 8208 | <emphasis>only</emphasis> abbreviations. |
|---|
| | 8209 | </para> |
|---|
| | 8210 | </listitem> |
|---|
| | 8211 | |
|---|
| | 8212 | <listitem> |
|---|
| | 8213 | <para> |
|---|
| | 8214 | Fix datetime input functions to correctly detect integer overflow when |
|---|
| | 8215 | running on a 64-bit platform (Tom) |
|---|
| | 8216 | </para> |
|---|
| | 8217 | </listitem> |
|---|
| | 8218 | |
|---|
| | 8219 | <listitem> |
|---|
| | 8220 | <para> |
|---|
| | 8221 | Improve performance of writing very long log messages to syslog (Tom) |
|---|
| | 8222 | </para> |
|---|
| | 8223 | </listitem> |
|---|
| | 8224 | |
|---|
| | 8225 | <listitem> |
|---|
| | 8226 | <para> |
|---|
| | 8227 | Fix bug in backwards scanning of a cursor on a <literal>SELECT DISTINCT |
|---|
| | 8228 | ON</literal> query (Tom) |
|---|
| | 8229 | </para> |
|---|
| | 8230 | </listitem> |
|---|
| | 8231 | |
|---|
| | 8232 | <listitem> |
|---|
| | 8233 | <para> |
|---|
| | 8234 | Fix planner bug with nested sub-select expressions (Tom) |
|---|
| | 8235 | </para> |
|---|
| | 8236 | |
|---|
| | 8237 | <para> |
|---|
| | 8238 | If the outer sub-select has no direct dependency on the parent query, |
|---|
| | 8239 | but the inner one does, the outer value might not get recalculated |
|---|
| | 8240 | for new parent query rows. |
|---|
| | 8241 | </para> |
|---|
| | 8242 | </listitem> |
|---|
| | 8243 | |
|---|
| | 8244 | <listitem> |
|---|
| | 8245 | <para> |
|---|
| | 8246 | Fix planner to estimate that <literal>GROUP BY</literal> expressions yielding |
|---|
| | 8247 | boolean results always result in two groups, regardless of the |
|---|
| | 8248 | expressions' contents (Tom) |
|---|
| | 8249 | </para> |
|---|
| | 8250 | |
|---|
| | 8251 | <para> |
|---|
| | 8252 | This is very substantially more accurate than the regular <literal>GROUP |
|---|
| | 8253 | BY</literal> estimate for certain boolean tests like <replaceable>col</replaceable> |
|---|
| | 8254 | <literal>IS NULL</literal>. |
|---|
| | 8255 | </para> |
|---|
| | 8256 | </listitem> |
|---|
| | 8257 | |
|---|
| | 8258 | <listitem> |
|---|
| | 8259 | <para> |
|---|
| | 8260 | Fix PL/PgSQL to not fail when a <literal>FOR</literal> loop's target variable |
|---|
| | 8261 | is a record containing composite-type fields (Tom) |
|---|
| | 8262 | </para> |
|---|
| | 8263 | </listitem> |
|---|
| | 8264 | |
|---|
| | 8265 | <listitem> |
|---|
| | 8266 | <para> |
|---|
| | 8267 | Fix PL/Tcl to behave correctly with Tcl 8.5, and to be more careful |
|---|
| | 8268 | about the encoding of data sent to or from Tcl (Tom) |
|---|
| | 8269 | </para> |
|---|
| | 8270 | </listitem> |
|---|
| | 8271 | |
|---|
| | 8272 | <listitem> |
|---|
| | 8273 | <para> |
|---|
| | 8274 | Fix PL/Python to work with Python 2.5 |
|---|
| | 8275 | </para> |
|---|
| | 8276 | |
|---|
| | 8277 | <para> |
|---|
| | 8278 | This is a back-port of fixes made during the 8.2 development cycle. |
|---|
| | 8279 | </para> |
|---|
| | 8280 | </listitem> |
|---|
| | 8281 | |
|---|
| | 8282 | <listitem> |
|---|
| | 8283 | <para> |
|---|
| | 8284 | Improve <application>pg_dump</application> and <application>pg_restore</application>'s |
|---|
| | 8285 | error reporting after failure to send a SQL command (Tom) |
|---|
| | 8286 | </para> |
|---|
| | 8287 | </listitem> |
|---|
| | 8288 | |
|---|
| | 8289 | <listitem> |
|---|
| | 8290 | <para> |
|---|
| | 8291 | Fix <application>pg_ctl</application> to properly preserve postmaster |
|---|
| | 8292 | command-line arguments across a <literal>restart</literal> (Bruce) |
|---|
| | 8293 | </para> |
|---|
| | 8294 | </listitem> |
|---|
| | 8295 | |
|---|
| | 8296 | <listitem> |
|---|
| | 8297 | <para> |
|---|
| | 8298 | Update time zone data files to <application>tzdata</application> release 2008f (for |
|---|
| | 8299 | DST law changes in Argentina, Bahamas, Brazil, Mauritius, Morocco, |
|---|
| | 8300 | Pakistan, Palestine, and Paraguay) |
|---|
| | 8301 | </para> |
|---|
| | 8302 | </listitem> |
|---|
| | 8303 | |
|---|
| | 8304 | </itemizedlist> |
|---|
| | 8305 | |
|---|
| | 8306 | </sect2> |
|---|
| | 8307 | </sect1> |
|---|
| | 8308 | |
|---|
| | 12169 | |
|---|
| | 12170 | <sect1 id="release-8-0-18"> |
|---|
| | 12171 | <title>Release 8.0.18</title> |
|---|
| | 12172 | |
|---|
| | 12173 | <note> |
|---|
| | 12174 | <title>Release date</title> |
|---|
| | 12175 | <simpara>2008-09-22</simpara> |
|---|
| | 12176 | </note> |
|---|
| | 12177 | |
|---|
| | 12178 | <para> |
|---|
| | 12179 | This release contains a variety of fixes from 8.0.17. |
|---|
| | 12180 | For information about new features in the 8.0 major release, see |
|---|
| | 12181 | <xref linkend="release-8-0"/>. |
|---|
| | 12182 | </para> |
|---|
| | 12183 | |
|---|
| | 12184 | <sect2> |
|---|
| | 12185 | <title>Migration to Version 8.0.18</title> |
|---|
| | 12186 | |
|---|
| | 12187 | <para> |
|---|
| | 12188 | A dump/restore is not required for those running 8.0.X. |
|---|
| | 12189 | However, if you are upgrading from a version earlier than 8.0.6, |
|---|
| | 12190 | see the release notes for 8.0.6. |
|---|
| | 12191 | </para> |
|---|
| | 12192 | |
|---|
| | 12193 | </sect2> |
|---|
| | 12194 | |
|---|
| | 12195 | <sect2> |
|---|
| | 12196 | <title>Changes</title> |
|---|
| | 12197 | |
|---|
| | 12198 | <itemizedlist> |
|---|
| | 12199 | |
|---|
| | 12200 | <listitem> |
|---|
| | 12201 | <para> |
|---|
| | 12202 | Widen local lock counters from 32 to 64 bits (Tom) |
|---|
| | 12203 | </para> |
|---|
| | 12204 | |
|---|
| | 12205 | <para> |
|---|
| | 12206 | This responds to reports that the counters could overflow in |
|---|
| | 12207 | sufficiently long transactions, leading to unexpected <quote>lock is |
|---|
| | 12208 | already held</quote> errors. |
|---|
| | 12209 | </para> |
|---|
| | 12210 | </listitem> |
|---|
| | 12211 | |
|---|
| | 12212 | <listitem> |
|---|
| | 12213 | <para> |
|---|
| | 12214 | Add checks in executor startup to ensure that the tuples produced by an |
|---|
| | 12215 | <command>INSERT</command> or <command>UPDATE</command> will match the target table's |
|---|
| | 12216 | current rowtype (Tom) |
|---|
| | 12217 | </para> |
|---|
| | 12218 | |
|---|
| | 12219 | <para> |
|---|
| | 12220 | <command>ALTER COLUMN TYPE</command>, followed by re-use of a previously |
|---|
| | 12221 | cached plan, could produce this type of situation. The check protects |
|---|
| | 12222 | against data corruption and/or crashes that could ensue. |
|---|
| | 12223 | </para> |
|---|
| | 12224 | </listitem> |
|---|
| | 12225 | |
|---|
| | 12226 | <listitem> |
|---|
| | 12227 | <para> |
|---|
| | 12228 | Fix datetime input functions to correctly detect integer overflow when |
|---|
| | 12229 | running on a 64-bit platform (Tom) |
|---|
| | 12230 | </para> |
|---|
| | 12231 | </listitem> |
|---|
| | 12232 | |
|---|
| | 12233 | <listitem> |
|---|
| | 12234 | <para> |
|---|
| | 12235 | Improve performance of writing very long log messages to syslog (Tom) |
|---|
| | 12236 | </para> |
|---|
| | 12237 | </listitem> |
|---|
| | 12238 | |
|---|
| | 12239 | <listitem> |
|---|
| | 12240 | <para> |
|---|
| | 12241 | Fix bug in backwards scanning of a cursor on a <literal>SELECT DISTINCT |
|---|
| | 12242 | ON</literal> query (Tom) |
|---|
| | 12243 | </para> |
|---|
| | 12244 | </listitem> |
|---|
| | 12245 | |
|---|