|
1 | 1 | <!-- doc/src/sgml/release-9.3.sgml --> |
2 | 2 | <!-- See header comment in release.sgml about typical markup --> |
3 | 3 |
|
| 4 | + <sect1 id="release-9-3-24"> |
| 5 | + <title>Release 9.3.24</title> |
| 6 | + |
| 7 | + <formalpara> |
| 8 | + <title>Release date:</title> |
| 9 | + <para>2018-08-09</para> |
| 10 | + </formalpara> |
| 11 | + |
| 12 | + <para> |
| 13 | + This release contains a variety of fixes from 9.3.23. |
| 14 | + For information about new features in the 9.3 major release, see |
| 15 | + <xref linkend="release-9-3"/>. |
| 16 | + </para> |
| 17 | + |
| 18 | + <para> |
| 19 | + The <productname>PostgreSQL</productname> community will stop releasing |
| 20 | + updates for the 9.3.X release series shortly after September 2018. |
| 21 | + Users are encouraged to update to a newer release branch soon. |
| 22 | + </para> |
| 23 | + |
| 24 | + <sect2> |
| 25 | + <title>Migration to Version 9.3.24</title> |
| 26 | + |
| 27 | + <para> |
| 28 | + A dump/restore is not required for those running 9.3.X. |
| 29 | + </para> |
| 30 | + |
| 31 | + <para> |
| 32 | + However, if you are upgrading from a version earlier than 9.3.23, |
| 33 | + see <xref linkend="release-9-3-23"/>. |
| 34 | + </para> |
| 35 | + </sect2> |
| 36 | + |
| 37 | + <sect2> |
| 38 | + <title>Changes</title> |
| 39 | + |
| 40 | + <itemizedlist> |
| 41 | + |
| 42 | + <listitem> |
| 43 | + <para> |
| 44 | + Ensure that updates to the <structfield>relfrozenxid</structfield> |
| 45 | + and <structfield>relminmxid</structfield> values |
| 46 | + for <quote>nailed</quote> system catalogs are processed in a timely |
| 47 | + fashion (Andres Freund) |
| 48 | + </para> |
| 49 | + |
| 50 | + <para> |
| 51 | + Overoptimistic caching rules could prevent these updates from being |
| 52 | + seen by other sessions, leading to spurious errors and/or data |
| 53 | + corruption. The problem was significantly worse for shared catalogs, |
| 54 | + such as <structname>pg_authid</structname>, because the stale cache |
| 55 | + data could persist into new sessions as well as existing ones. |
| 56 | + </para> |
| 57 | + </listitem> |
| 58 | + |
| 59 | + <listitem> |
| 60 | + <para> |
| 61 | + Fix case where a freshly-promoted standby crashes before having |
| 62 | + completed its first post-recovery checkpoint (Michael Paquier, Kyotaro |
| 63 | + Horiguchi, Pavan Deolasee, Álvaro Herrera) |
| 64 | + </para> |
| 65 | + |
| 66 | + <para> |
| 67 | + This led to a situation where the server did not think it had reached |
| 68 | + a consistent database state during subsequent WAL replay, preventing |
| 69 | + restart. |
| 70 | + </para> |
| 71 | + </listitem> |
| 72 | + |
| 73 | + <listitem> |
| 74 | + <para> |
| 75 | + Avoid emitting a bogus WAL record when recycling an all-zero btree |
| 76 | + page (Amit Kapila) |
| 77 | + </para> |
| 78 | + |
| 79 | + <para> |
| 80 | + This mistake has been seen to cause assertion failures, and |
| 81 | + potentially it could result in unnecessary query cancellations on hot |
| 82 | + standby servers. |
| 83 | + </para> |
| 84 | + </listitem> |
| 85 | + |
| 86 | + <listitem> |
| 87 | + <para> |
| 88 | + Improve performance of WAL replay for transactions that drop many |
| 89 | + relations (Fujii Masao) |
| 90 | + </para> |
| 91 | + |
| 92 | + <para> |
| 93 | + This change reduces the number of times that shared buffers are |
| 94 | + scanned, so that it is of most benefit when that setting is large. |
| 95 | + </para> |
| 96 | + </listitem> |
| 97 | + |
| 98 | + <listitem> |
| 99 | + <para> |
| 100 | + Improve performance of lock releasing in standby server WAL replay |
| 101 | + (Thomas Munro) |
| 102 | + </para> |
| 103 | + </listitem> |
| 104 | + |
| 105 | + <listitem> |
| 106 | + <para> |
| 107 | + Ensure a table's cached index list is correctly rebuilt after an index |
| 108 | + creation fails partway through (Peter Geoghegan) |
| 109 | + </para> |
| 110 | + |
| 111 | + <para> |
| 112 | + Previously, the failed index's OID could remain in the list, causing |
| 113 | + problems later in the same session. |
| 114 | + </para> |
| 115 | + </listitem> |
| 116 | + |
| 117 | + <listitem> |
| 118 | + <para> |
| 119 | + Fix misoptimization of equivalence classes involving composite-type |
| 120 | + columns (Tom Lane) |
| 121 | + </para> |
| 122 | + |
| 123 | + <para> |
| 124 | + This resulted in failure to recognize that an index on a composite |
| 125 | + column could provide the sort order needed for a mergejoin on that |
| 126 | + column. |
| 127 | + </para> |
| 128 | + </listitem> |
| 129 | + |
| 130 | + <listitem> |
| 131 | + <para> |
| 132 | + Fix SQL-standard <literal>FETCH FIRST</literal> syntax to allow |
| 133 | + parameters (<literal>$<replaceable>n</replaceable></literal>), as the |
| 134 | + standard expects (Andrew Gierth) |
| 135 | + </para> |
| 136 | + </listitem> |
| 137 | + |
| 138 | + <listitem> |
| 139 | + <para> |
| 140 | + Fix failure to schema-qualify some object names |
| 141 | + in <function>getObjectDescription</function> output |
| 142 | + (Kyotaro Horiguchi, Tom Lane) |
| 143 | + </para> |
| 144 | + |
| 145 | + <para> |
| 146 | + Names of collations, conversions, and text search objects |
| 147 | + were not schema-qualified when they should be. |
| 148 | + </para> |
| 149 | + </listitem> |
| 150 | + |
| 151 | + <listitem> |
| 152 | + <para> |
| 153 | + Widen <command>COPY FROM</command>'s current-line-number counter |
| 154 | + from 32 to 64 bits (David Rowley) |
| 155 | + </para> |
| 156 | + |
| 157 | + <para> |
| 158 | + This avoids two problems with input exceeding 4G lines: <literal>COPY |
| 159 | + FROM WITH HEADER</literal> would drop a line every 4G lines, not only |
| 160 | + the first line, and error reports could show a wrong line number. |
| 161 | + </para> |
| 162 | + </listitem> |
| 163 | + |
| 164 | + <listitem> |
| 165 | + <para> |
| 166 | + Add a string freeing function |
| 167 | + to <application>ecpg</application>'s <filename>pgtypes</filename> |
| 168 | + library, so that cross-module memory management problems can be |
| 169 | + avoided on Windows (Takayuki Tsunakawa) |
| 170 | + </para> |
| 171 | + |
| 172 | + <para> |
| 173 | + On Windows, crashes can ensue if the <function>free</function> call |
| 174 | + for a given chunk of memory is not made from the same DLL |
| 175 | + that <function>malloc</function>'ed the memory. |
| 176 | + The <filename>pgtypes</filename> library sometimes returns strings |
| 177 | + that it expects the caller to free, making it impossible to follow |
| 178 | + this rule. Add a <function>PGTYPESchar_free()</function> function |
| 179 | + that just wraps <function>free</function>, allowing applications |
| 180 | + to follow this rule. |
| 181 | + </para> |
| 182 | + </listitem> |
| 183 | + |
| 184 | + <listitem> |
| 185 | + <para> |
| 186 | + Fix <application>ecpg</application>'s support for <type>long |
| 187 | + long</type> variables on Windows, as well as other platforms that |
| 188 | + declare <function>strtoll</function>/<function>strtoull</function> |
| 189 | + nonstandardly or not at all (Dang Minh Huong, Tom Lane) |
| 190 | + </para> |
| 191 | + </listitem> |
| 192 | + |
| 193 | + <listitem> |
| 194 | + <para> |
| 195 | + Fix misidentification of SQL statement type in PL/pgSQL, when a rule |
| 196 | + change causes a change in the semantics of a statement intra-session |
| 197 | + (Tom Lane) |
| 198 | + </para> |
| 199 | + |
| 200 | + <para> |
| 201 | + This error led to assertion failures, or in rare cases, failure to |
| 202 | + enforce the <literal>INTO STRICT</literal> option as expected. |
| 203 | + </para> |
| 204 | + </listitem> |
| 205 | + |
| 206 | + <listitem> |
| 207 | + <para> |
| 208 | + Fix password prompting in client programs so that echo is properly |
| 209 | + disabled on Windows when <literal>stdin</literal> is not the |
| 210 | + terminal (Matthew Stickney) |
| 211 | + </para> |
| 212 | + </listitem> |
| 213 | + |
| 214 | + <listitem> |
| 215 | + <para> |
| 216 | + Further fix mis-quoting of values for list-valued GUC variables in |
| 217 | + dumps (Tom Lane) |
| 218 | + </para> |
| 219 | + |
| 220 | + <para> |
| 221 | + The previous fix for quoting of <varname>search_path</varname> and |
| 222 | + other list-valued variables in <application>pg_dump</application> |
| 223 | + output turned out to misbehave for empty-string list elements, and it |
| 224 | + risked truncation of long file paths. |
| 225 | + </para> |
| 226 | + </listitem> |
| 227 | + |
| 228 | + <listitem> |
| 229 | + <para> |
| 230 | + Make <application>pg_upgrade</application> check that the old server |
| 231 | + was shut down cleanly (Bruce Momjian) |
| 232 | + </para> |
| 233 | + |
| 234 | + <para> |
| 235 | + The previous check could be fooled by an immediate-mode shutdown. |
| 236 | + </para> |
| 237 | + </listitem> |
| 238 | + |
| 239 | + <listitem> |
| 240 | + <para> |
| 241 | + Fix crash in <filename>contrib/ltree</filename>'s |
| 242 | + <function>lca()</function> function when the input array is empty |
| 243 | + (Pierre Ducroquet) |
| 244 | + </para> |
| 245 | + </listitem> |
| 246 | + |
| 247 | + <listitem> |
| 248 | + <para> |
| 249 | + Fix various error-handling code paths in which an incorrect error code |
| 250 | + might be reported (Michael Paquier, Tom Lane, Magnus Hagander) |
| 251 | + </para> |
| 252 | + </listitem> |
| 253 | + |
| 254 | + <listitem> |
| 255 | + <para> |
| 256 | + Rearrange makefiles to ensure that programs link to freshly-built |
| 257 | + libraries (such as <filename>libpq.so</filename>) rather than ones |
| 258 | + that might exist in the system library directories (Tom Lane) |
| 259 | + </para> |
| 260 | + |
| 261 | + <para> |
| 262 | + This avoids problems when building on platforms that supply old copies |
| 263 | + of <productname>PostgreSQL</productname> libraries. |
| 264 | + </para> |
| 265 | + </listitem> |
| 266 | + |
| 267 | + <listitem> |
| 268 | + <para> |
| 269 | + Update time zone data files to <application>tzdata</application> |
| 270 | + release 2018e for DST law changes in North Korea, plus historical |
| 271 | + corrections for Czechoslovakia. |
| 272 | + </para> |
| 273 | + |
| 274 | + <para> |
| 275 | + This update includes a redefinition of <quote>daylight savings</quote> |
| 276 | + in Ireland, as well as for some past years in Namibia and |
| 277 | + Czechoslovakia. In those jurisdictions, legally standard time is |
| 278 | + observed in summer, and daylight savings time in winter, so that the |
| 279 | + daylight savings offset is one hour behind standard time not one hour |
| 280 | + ahead. This does not affect either the actual UTC offset or the |
| 281 | + timezone abbreviations in use; the only known effect is that |
| 282 | + the <structfield>is_dst</structfield> column in |
| 283 | + the <structname>pg_timezone_names</structname> view will now be true |
| 284 | + in winter and false in summer in these cases. |
| 285 | + </para> |
| 286 | + </listitem> |
| 287 | + |
| 288 | + </itemizedlist> |
| 289 | + |
| 290 | + </sect2> |
| 291 | + </sect1> |
| 292 | + |
4 | 293 | <sect1 id="release-9-3-23"> |
5 | 294 | <title>Release 9.3.23</title> |
6 | 295 |
|
|