cd85918 | PostgreSQL Daemon | 16 September 2005, 17:15:55 UTC | tag it all beta2 ... | 16 September 2005, 17:15:55 UTC |
0296606 | Bruce Momjian | 16 September 2005, 16:46:09 UTC | Update wording: < 390d388 < 453c451 < removed or have its heap and index files truncated. One > be removed or have its heap and index files truncated. One < * Use a phantom command counter for nested subtransactions to reduce < per-tuple overhead | 16 September 2005, 16:46:09 UTC |
f4516f8 | Teodor Sigaev | 16 September 2005, 14:40:54 UTC | Small fixes | 16 September 2005, 14:40:54 UTC |
3f558e3 | Bruce Momjian | 16 September 2005, 11:24:02 UTC | Add: > o Add single-step debugging of PL/PgSQL functions | 16 September 2005, 11:24:02 UTC |
e4d9b69 | Neil Conway | 16 September 2005, 05:35:41 UTC | Rename pg_complete_relation_size() to pg_total_relation_size(), for the sake of brevity and clarity. Make pg_reload_conf(), pg_rotate_logfile(), and pg_cancel_backend() return a boolean rather than an integer to indicate success or failure. Along the way, make some minor cleanups to dbsize.c -- in particular, use elog() rather than ereport() for "shouldn't happen" error conditions, and remove some of the more flagrant violations of the Postgres indentation conventions. Catalog version bumped. | 16 September 2005, 05:35:41 UTC |
148c00a | Neil Conway | 16 September 2005, 04:13:18 UTC | Update two comments to refer to use the new list API names. | 16 September 2005, 04:13:18 UTC |
b09bfc6 | Tom Lane | 16 September 2005, 03:37:35 UTC | Update release notes for changes between beta1 and beta2. | 16 September 2005, 03:37:35 UTC |
3e579cf | Alvaro Herrera | 16 September 2005, 03:12:32 UTC | Add note about pg_autovacuum settings not being saved on a dump, per Robert Treat. | 16 September 2005, 03:12:32 UTC |
dc06734 | Tom Lane | 16 September 2005, 00:30:05 UTC | Force the size and alignment of LWLock array entries to be either 16 or 32 bytes. This shouldn't make any difference on x86 machines, where the size happened to be 16 bytes anyway, but on 64-bit machines and machines with slock_t int or wider, it will speed array indexing and hopefully reduce SMP cache contention effects. Per recent experimentation. | 16 September 2005, 00:30:05 UTC |
1dd9b09 | Neil Conway | 15 September 2005, 17:44:27 UTC | Copy-editing for GiST README. | 15 September 2005, 17:44:27 UTC |
79fae4a | Teodor Sigaev | 15 September 2005, 16:39:15 UTC | Readme about GiST's algorithms | 15 September 2005, 16:39:15 UTC |
f82b853 | Teodor Sigaev | 15 September 2005, 11:14:18 UTC | 1 Update Snowball sources 2 Makefile fixes | 15 September 2005, 11:14:18 UTC |
3b0453b | Neil Conway | 14 September 2005, 21:14:26 UTC | Various documentation improvements. Add a lot of index entries for the builtin functions (although some more entries are still needed), and remove the duplicate index entries that have been causing collateindex.pl warnings. Consistently use "int" and "bigint", rather than a mix of "int", "integer", "int4", "bigint", and "int8". Make parenthesis style in syntax examples more consistent. Various copy-editing for newly-added documentation and SGML markup fixes. | 14 September 2005, 21:14:26 UTC |
08dc2af | Neil Conway | 14 September 2005, 18:35:38 UTC | Tweak the PL/PgSQL regression tests to catch the recently reported bug in parsing cursor declarations. | 14 September 2005, 18:35:38 UTC |
323f0a6 | Tom Lane | 14 September 2005, 13:46:47 UTC | Fix recent breakage of decl_cursor_arglist syntax, per Michael Paesold. | 14 September 2005, 13:46:47 UTC |
5b0fa0c | Tom Lane | 13 September 2005, 16:16:17 UTC | Fix make_tuple_from_row to support nested rowtypes, per gripe from Roman Neuhauser. Update some obsolete comments for exec_eval_datum, too. | 13 September 2005, 16:16:17 UTC |
d6bc885 | Neil Conway | 13 September 2005, 15:24:57 UTC | Minor SGML markup fixes. | 13 September 2005, 15:24:57 UTC |
c93912b | Alvaro Herrera | 13 September 2005, 01:51:18 UTC | Document the new autovacuum daemon. | 13 September 2005, 01:51:18 UTC |
2a11a49 | Tom Lane | 12 September 2005, 23:09:05 UTC | Fix cut-and-paste-o in TCP keepalives documentation. | 12 September 2005, 23:09:05 UTC |
fe9f96b | Neil Conway | 12 September 2005, 22:39:30 UTC | Minor documentation tweak: make the capitalization of a section title consistent with the other sections in this chapter. | 12 September 2005, 22:39:30 UTC |
f59b05c | Tom Lane | 12 September 2005, 22:20:16 UTC | Ensure that any memory leaked during an error inside the bgwriter is recovered. I did not see any actual leak while testing this in CVS tip, but 8.0 definitely has a problem with leaking the space temporarily palloc'd by BufferSync(). In any case this seems a good idea to forestall similar problems in future. Per report from Arjen van der Meijden. | 12 September 2005, 22:20:16 UTC |
9813fc1 | Neil Conway | 12 September 2005, 22:11:38 UTC | Make the documentation of GUC variables a separate chapter, rather than a section of the "Server Run-time Environment" chapter. Also, move the SGML for the new chapter to a separate file and fix the resulting fallout. | 12 September 2005, 22:11:38 UTC |
37a3ed6 | Tom Lane | 12 September 2005, 19:17:45 UTC | Update manual's introduction to indexes to reflect 8.1 changes; in particular the addition of bitmap scans and the relaxation of rules about when multicolumn indexes can be used. Also some minor editorial work in other parts of the chapter. | 12 September 2005, 19:17:45 UTC |
f7259f5 | Neil Conway | 12 September 2005, 18:40:19 UTC | Fix a mistake in the documentation for SPI_getbinval(), per Michael Fuhr. | 12 September 2005, 18:40:19 UTC |
b93a72a | Tom Lane | 12 September 2005, 16:43:29 UTC | Forgot to add pg_pltemplate to the list of shared system catalogs that appears in the REINDEX ref page. | 12 September 2005, 16:43:29 UTC |
c959d37 | Michael Meskes | 12 September 2005, 11:57:53 UTC | Fixed transaction command handling to not ignore savepoints and to correctly check for errors. | 12 September 2005, 11:57:53 UTC |
4fe4563 | Neil Conway | 12 September 2005, 05:10:12 UTC | Add a few items to the release notes, and make a few more minor cleanups. | 12 September 2005, 05:10:12 UTC |
cc3bda3 | Tom Lane | 12 September 2005, 02:26:33 UTC | Tweak TCP-keepalive code so that an invalid setting doesn't cause us to drop connections unceremoniously. Also some other marginal cleanups: don't query getsockopt() repeatedly if it fails, and avoid having the apparent definition of struct Port depend on which system headers you might have included or not. Oliver Jowett and Tom Lane. | 12 September 2005, 02:26:33 UTC |
f7a5f90 | Neil Conway | 11 September 2005, 23:40:33 UTC | Many small improvements and copy edits to the release notes. Use "—" rather than "-" for the "dash" character. Correct SGML markup. Make references to the names of contributors more consistent. Rewrite a bit of prose, and make some other similar cleanups. | 11 September 2005, 23:40:33 UTC |
c7d1a8d | Tom Lane | 11 September 2005, 04:10:25 UTC | Fix some corner-case bugs in _sendSQLLine's parsing of SQL commands > found in a pg_dump archive. It had problems with dollar-quote tags broken across bufferload boundaries (this may explain bug report from Rod Taylor), also with dollar-quote literals of the form $a$a$..., and was also confused about the rules for backslash in double quoted identifiers (ie, they're not special). Also put in placeholder support for E'...' literals --- this will need more work later. | 11 September 2005, 04:10:25 UTC |
e1a7d1b | Tom Lane | 11 September 2005, 00:36:14 UTC | Avoid changing stdin/stdout to binary mode on Windows unless that is really the source or destination of the archive. I think this will resolve recent complaints that password prompting is broken in pg_restore on Windows. Note that password prompting and reading from stdin is an unworkable combination on Windows ... but that was true anyway. | 11 September 2005, 00:36:14 UTC |
d134889 | Bruce Momjian | 10 September 2005, 05:21:52 UTC | Update URL for FAQ_MINGW. | 10 September 2005, 05:21:52 UTC |
35979e6 | Tom Lane | 09 September 2005, 06:51:12 UTC | Given its current definition that depends on time(NULL), timetz_zone is certainly no longer immutable, but must indeed be marked volatile. I wonder if it should use the value of now() (that is, transaction start time) so that it could be marked stable. But it's probably not important enough to be worth changing the code for ... indeed, I'm not even going to force an initdb for this catalog change, seeing that we just did one a few hours ago. | 09 September 2005, 06:51:12 UTC |
f2ebd01 | Tom Lane | 09 September 2005, 06:46:14 UTC | timestamptz_izone should return the input, not NULL, when the input is a non-finite timestamp, for consistency with related functions. In other words: +infinity rotated to a different timezone is still +infinity. | 09 September 2005, 06:46:14 UTC |
a239af0 | Tom Lane | 09 September 2005, 02:31:50 UTC | Fix the various forms of AT TIME ZONE to accept either timezones found in the zic database or zone names found in the date token table. This preserves the old ability to do AT TIME ZONE 'PST' along with the new ability to do AT TIME ZONE 'PST8PDT'. Per gripe from Bricklen Anderson. Also, fix some inconsistencies in usage of TZ_STRLEN_MAX --- the old code had the potential for one-byte buffer overruns, though given alignment considerations it's unlikely there was any real risk. | 09 September 2005, 02:31:50 UTC |
48123de | Tom Lane | 08 September 2005, 20:07:42 UTC | Create the pg_pltemplate system catalog to hold template information for procedural languages. This replaces the hard-wired table I had originally proposed as a stopgap solution. For the moment, the initial contents only include languages shipped with the core distribution. | 08 September 2005, 20:07:42 UTC |
c18cabe | Tom Lane | 08 September 2005, 16:49:04 UTC | Update regression tests for new USA timezone data. Mea culpa for not realizing that the regression tests could be affected. | 08 September 2005, 16:49:04 UTC |
69d48ce | Tom Lane | 07 September 2005, 21:39:25 UTC | Update timezone data files to release 2005m of the zic database. Among other changes, this reflects the recently passed change in USA daylight savings rules. | 07 September 2005, 21:39:25 UTC |
e0dedd0 | Tom Lane | 05 September 2005, 23:50:49 UTC | Implement a preliminary 'template' facility for procedural languages, as per my recent proposal. For now the template data is hard-wired in proclang.c --- this should be replaced later by a new shared system catalog, but we don't want to force initdb during 8.1 beta. This change lets us cleanly load existing dump files even if they contain outright wrong information about a PL's support functions, such as a wrong path to the shared library or a missing validator function. Also, we can revert the recent kluges to make pg_dump dump PL support functions that are stored in pg_catalog. While at it, I removed the code in pg_regress that replaced $libdir with a hardcoded path for temporary installations. This is no longer needed given our support for relocatable installations. | 05 September 2005, 23:50:49 UTC |
e35e6b1 | Tom Lane | 05 September 2005, 18:59:38 UTC | Back out prior patch and instead just suppress SubqueryScan elimination when there are extra resjunk columns in the child node. I found some additional cases involving Append nodes that weren't handled by the prior patch, and it's not clear how to fix them in the same way without breaking inheritance cases. So the prudent path seems to be to narrow the scope of the optimization. | 05 September 2005, 18:59:38 UTC |
2da2d34 | Tom Lane | 05 September 2005, 18:05:13 UTC | Clean up version comparison/display code, per gripe from Michael Fuhr. | 05 September 2005, 18:05:13 UTC |
0372894 | Tom Lane | 05 September 2005, 17:25:01 UTC | For non-projecting plan node types such as Limit, set_plan_references has to recopy the input plan node's targetlist if it removes a SubqueryScan node just below the non-projecting node. For simplicity I made it recopy always. Per bug report from Allan Wang and Michael Fuhr. | 05 September 2005, 17:25:01 UTC |
2739e1d | Andrew Dunstan | 05 September 2005, 14:44:05 UTC | Add a note about CSV lines with trailing spaces. Suggestion by Darcy Buskermolen, reworded by me. | 05 September 2005, 14:44:05 UTC |
656fe39 | Bruce Momjian | 05 September 2005, 13:59:08 UTC | Conditionally output the server version number in psql if it doesn't exactly match the client version number, and warn about major version mismatches. | 05 September 2005, 13:59:08 UTC |
e8f3541 | Alvaro Herrera | 04 September 2005, 13:40:24 UTC | Fix typo in Richard's last name. | 04 September 2005, 13:40:24 UTC |
a17dbec | Tom Lane | 03 September 2005, 15:55:00 UTC | libpgport routines need nonstandard palloc to work on Windows. Propagate hack that's in dirmod.c to copydir.c. | 03 September 2005, 15:55:00 UTC |
b72fe49 | Bruce Momjian | 03 September 2005, 00:41:30 UTC | Add idea: < cmin/cmax pair and is stored in local memory. > cmin/cmax pair and is stored in local memory. Another idea is to > store both cmin and cmax only in local memory. | 03 September 2005, 00:41:30 UTC |
11b5c55 | Bruce Momjian | 02 September 2005, 21:50:54 UTC | Add log display label for unnamed server-side portals. | 02 September 2005, 21:50:54 UTC |
0802078 | Tom Lane | 02 September 2005, 21:25:30 UTC | Fix the spelling. | 02 September 2005, 21:25:30 UTC |
264cd5f | Bruce Momjian | 02 September 2005, 21:23:05 UTC | Update. < have its heap and index files truncated. One issue is < that no other backend should be able to add to the table < at the same time, which is something that is currently < allowed. > removed or have its heap and index files truncated. One > issue is that no other backend should be able to add to > the table at the same time, which is something that is > currently allowed. | 02 September 2005, 21:23:05 UTC |
ef6d5b4 | Bruce Momjian | 02 September 2005, 21:10:59 UTC | Update: > o Allow COPY on a newly-created table to skip WAL logging 450a452,456 > On crash recovery, the table involved in the COPY would > have its heap and index files truncated. One issue is > that no other backend should be able to add to the table > at the same time, which is something that is currently > allowed. | 02 September 2005, 21:10:59 UTC |
f8ec875 | Bruce Momjian | 02 September 2005, 21:01:20 UTC | Update: < o Allow COPY into an empty table to skip WAL logging > o Allow COPY into an new table to skip WAL logging | 02 September 2005, 21:01:20 UTC |
4506ec9 | Bruce Momjian | 02 September 2005, 21:00:59 UTC | Update listen_addresses description. Darcy Buskermolen | 02 September 2005, 21:00:59 UTC |
5bddded | Bruce Momjian | 02 September 2005, 20:11:34 UTC | Add: > * Use UTF8 encoding for NLS messages so all server encodings can > read them properly < o %Add support for Unicode < < To fix this, the data needs to be converted to/from UTF16/UTF8 < so the Win32 wcscoll() can be used, and perhaps other functions < like towupper(). However, UTF8 already works with normal < locales but provides no ordering or character set classes. | 02 September 2005, 20:11:34 UTC |
3bea1a3 | Bruce Momjian | 02 September 2005, 19:51:37 UTC | Update wording: < could only see committed rows from another transaction. However, > could only see rows from another completed transaction. However, 981c981 < proper visibility of the row, for example, for cursors. > proper visibility of the row's cmin, for example, for cursors. | 02 September 2005, 19:51:37 UTC |
da08559 | Bruce Momjian | 02 September 2005, 19:39:15 UTC | Add: > o Allow COPY into an empty table to skip WAL logging | 02 September 2005, 19:39:15 UTC |
35e9b1c | Tom Lane | 02 September 2005, 19:02:20 UTC | Clean up a couple of ad-hoc computations of the maximum number of tuples on a page, as suggested by ITAGAKI Takahiro. Also, change a few places that were using some other estimates of max-items-per-page to consistently use MaxOffsetNumber. This is conservatively large --- we could have used the new MaxHeapTuplesPerPage macro, or a similar one for index tuples --- but those places are simply declaring a fixed-size buffer and assuming it will work, rather than actively testing for overrun. It seems safer to size these buffers in a way that can't overflow even if the page is corrupt. | 02 September 2005, 19:02:20 UTC |
962a4bb | Tom Lane | 02 September 2005, 18:55:32 UTC | In copy_file, use a palloc'd buffer instead of just a local char array; a local array isn't guaranteed to have any particular alignment, and so it could slow down the data transfer. | 02 September 2005, 18:55:32 UTC |
fad7e8e | Bruce Momjian | 02 September 2005, 15:22:44 UTC | Add text: * Merge xmin/xmax/cmin/cmax back into three header fields Before subtransactions, there used to be only three fields needed to store these four values. This was possible because only the current transaction looks at the cmin/cmax values. If the current transaction created and expired the row the fields stored where xmin (same as xmax), cmin, cmax, and if the transaction was expiring a row from a another transaction, the fields stored were xmin (cmin was not needed), xmax, and cmax. Such a system worked because a transaction could only see committed rows from another transaction. However, subtransactions can see rows from outer transactions, and once the subtransaction completes, the outer transaction continues, requiring the storage of all four fields. With subtransactions, an outer transaction can create a row, a subtransaction expire it, and when the subtransaction completes, the outer transaction still has to have proper visibility of the row, for example, for cursors. One possible solution is to create a phantom cid which represents a cmin/cmax pair and is stored in local memory. | 02 September 2005, 15:22:44 UTC |
6c0398a | Tom Lane | 02 September 2005, 03:19:53 UTC | Add some notes about how pg_dump relates to the practices recommended under 'Populating a Database'. | 02 September 2005, 03:19:53 UTC |
9a412be | Tom Lane | 02 September 2005, 00:57:57 UTC | Update performance discussion for 8.1. Add a little more explanatory material in the EXPLAIN section, update examples to match current reality, show examples of bitmap indexscans as well as plain ones. | 02 September 2005, 00:57:57 UTC |
55af2a4 | Bruce Momjian | 02 September 2005, 00:49:57 UTC | Duplicate: < * Maintain a map of recently-expired rows < < This allows vacuum to target specific pages for possible free space < without requiring a sequential scan. < Update entry: > One complexity is that index entries still have to be vacuumed, and > doing this without an index scan (by using the heap values to find the > index entry) might be slow and unreliable, especially for user-defined > index functions. | 02 September 2005, 00:49:57 UTC |
ce128c0 | Tom Lane | 01 September 2005, 22:02:44 UTC | Correct comment about why contrib/xml2 isn't built automatically. | 01 September 2005, 22:02:44 UTC |
e32a2ec | Tom Lane | 01 September 2005, 20:01:53 UTC | Update storage.sgml to reflect current reality. | 01 September 2005, 20:01:53 UTC |
a635ab4 | Tom Lane | 01 September 2005, 18:15:42 UTC | Short-circuit AllocSetReset if nothing has been palloc'd in the memory context since the previous AllocSetReset. Original patch by Atsushi Ogawa, editorialized on a little bit by Tom Lane. | 01 September 2005, 18:15:42 UTC |
91a21b5 | Bruce Momjian | 01 September 2005, 16:18:51 UTC | 8.0.3 is now the most recent version. | 01 September 2005, 16:18:51 UTC |
fc9b645 | Bruce Momjian | 01 September 2005, 16:18:08 UTC | Update disk space computation because oids are now optional. | 01 September 2005, 16:18:08 UTC |
14720f7 | Tom Lane | 01 September 2005, 15:34:31 UTC | Fix unportable uses of <ctype.h> functions. Per Sergey Koposov. | 01 September 2005, 15:34:31 UTC |
bc06506 | Bruce Momjian | 30 August 2005, 23:25:23 UTC | Add: > * Eventually enable escape_string_warning and standard_conforming_strings | 30 August 2005, 23:25:23 UTC |
59b4cef | Bruce Momjian | 30 August 2005, 18:57:48 UTC | Add regression tests for disabling constraints. Gavin Sherry | 30 August 2005, 18:57:48 UTC |
571be6f | Tom Lane | 30 August 2005, 15:48:28 UTC | Remove a couple of obsolete statements about how many buffers you can fit into a 512K shared memory segment. | 30 August 2005, 15:48:28 UTC |
b740be2 | Tom Lane | 30 August 2005, 02:47:37 UTC | Log send() failures when sending to statistics process (but only in assert-enabled builds). This is a temporary measure to see if we can learn anything about those intermittent stats test failures in the buildfarm. | 30 August 2005, 02:47:37 UTC |
e038e89 | Tom Lane | 30 August 2005, 02:09:29 UTC | Improve table describing shared memory parameters. | 30 August 2005, 02:09:29 UTC |
cc36795 | Tom Lane | 30 August 2005, 01:37:38 UTC | Mention max_prepared_transactions in PREPARE TRANSACTION reference page, per Bruce's suggestion. | 30 August 2005, 01:37:38 UTC |
13065c7 | Tom Lane | 30 August 2005, 01:08:47 UTC | DropTableSpace forgot to remove dependency on tablespace's owner. Per report from Jaime Casanova. | 30 August 2005, 01:08:47 UTC |
139b1f6 | Tom Lane | 30 August 2005, 01:07:54 UTC | Fix misleading comment. | 30 August 2005, 01:07:54 UTC |
78ef2d3 | Tom Lane | 30 August 2005, 00:58:48 UTC | Update documentation about shared memory sizing to reflect current reality. | 30 August 2005, 00:58:48 UTC |
037709e | Tom Lane | 29 August 2005, 21:38:18 UTC | Reduce default value of max_prepared_transactions from 50 to 5. This saves nearly 700kB in the default shared memory segment size, which seems worthwhile, and it is a feature that many users won't use anyway. Per Heikki's argument, there is no point in a compromise value --- those who are using 2PC at all will probably want it at least equal to max_connections. But we can't set it to zero by default without breaking the prepared_xacts regression test. | 29 August 2005, 21:38:18 UTC |
8aec77f | Tom Lane | 29 August 2005, 19:39:39 UTC | Fix platform-specific test for path prefix-ness: move it into path.c where it can be done right. Allow explicit use of absolute DataDir path. Per Dave Page. | 29 August 2005, 19:39:39 UTC |
6b44d79 | Tom Lane | 29 August 2005, 01:32:00 UTC | Fix misspelled error message. | 29 August 2005, 01:32:00 UTC |
1a6fe83 | Bruce Momjian | 29 August 2005, 00:47:35 UTC | Allow Win32 libpq will use it's minimal pthread implementation, and ecpg will use pthreadGC2. Dave Page | 29 August 2005, 00:47:35 UTC |
2d03390 | Tom Lane | 29 August 2005, 00:41:34 UTC | Sigh, looks like you need '.set mips2' before you can access MIPS SYNC instruction. | 29 August 2005, 00:41:34 UTC |
46a0eee | Tom Lane | 28 August 2005, 22:47:20 UTC | Tweak nodeBitmapAnd to stop evaluating sub-plan scans if it finds it's got an empty bitmap after any step; the remaining subplans can no longer affect the result. Per a suggestion from Ilia Kantor. | 28 August 2005, 22:47:20 UTC |
75e90bb | Bruce Momjian | 28 August 2005, 22:21:46 UTC | Fix initdb quoting for Win32 paths in final examples, per Dave Page. | 28 August 2005, 22:21:46 UTC |
3b85fc3 | Bruce Momjian | 28 August 2005, 21:42:28 UTC | Update pthread_self() on Win32 to return DWORD. | 28 August 2005, 21:42:28 UTC |
a78a530 | Bruce Momjian | 28 August 2005, 18:52:37 UTC | Add: > * Add function to return the thread safety status of libpq and ecpg | 28 August 2005, 18:52:37 UTC |
8046c1c | Bruce Momjian | 28 August 2005, 18:49:01 UTC | Use GetCurrentThreadId, per Magnus. | 28 August 2005, 18:49:01 UTC |
7319ab9 | Tom Lane | 28 August 2005, 18:26:01 UTC | Add a SYNC instruction to the S_UNLOCK sequence for MIPS. | 28 August 2005, 18:26:01 UTC |
0c172b2 | Bruce Momjian | 28 August 2005, 16:37:48 UTC | Add comment about pthread_self() cast. | 28 August 2005, 16:37:48 UTC |
0845538 | Tom Lane | 28 August 2005, 16:31:37 UTC | dumpUserConfig failed (in a pretty harmless way, but failed nonetheless) to cope with a group name when dumping from a pre-8.1 installation. Per report from Stefan Kaltenbrunner. | 28 August 2005, 16:31:37 UTC |
974e3cf | Tom Lane | 27 August 2005, 22:37:00 UTC | cost_agg really ought to charge something per output tuple; else there are cases where it appears to have zero run cost. | 27 August 2005, 22:37:00 UTC |
4e5fbb3 | Tom Lane | 27 August 2005, 22:13:44 UTC | Change the division of labor between grouping_planner and query_planner so that the latter estimates the number of groups that grouping will produce. This is needed because it is primarily query_planner that makes the decision between fast-start and fast-finish plans, and in the original coding it was unable to make more than a crude rule-of-thumb choice when the query involved grouping. This revision helps us make saner choices for queries like SELECT ... GROUP BY ... LIMIT, as in a recent example from Mark Kirkwood. Also move the responsibility for canonicalizing sort_pathkeys and group_pathkeys into query_planner; this information has to be available anyway to support the first change, and doing it this way lets us get rid of compare_noncanonical_pathkeys entirely. | 27 August 2005, 22:13:44 UTC |
9e56c5a | Tom Lane | 27 August 2005, 18:44:03 UTC | Windows needs WSAStartup() before getaddrinfo() will work. Andrew Dunstan | 27 August 2005, 18:44:03 UTC |
5a7d369 | Tom Lane | 27 August 2005, 18:04:49 UTC | Fix two separate bugs in setrefs.c. set_subqueryscan_references needs to copy the whole plan tree before invoking adjust_plan_varnos(); else if there is any multiply-linked substructure, the latter might increment some Var's varno twice. Previously there were some retail copyObject calls inside adjust_plan_varnos, but it seems a lot safer to just dup the whole tree first. Also, set_inner_join_references was trying to avoid work by not recursing if a BitmapHeapScan's bitmapqualorig contained no outer references; which was OK at the time the code was written, I think, but now that create_bitmap_scan_plan removes duplicate clauses from bitmapqualorig it is possible for that field to be NULL while outer references still remain in the qpqual and/or contained indexscan nodes. For safety, always recurse even if the BitmapHeapScan looks to be outer reference free. Per reports from Michael Fuhr and Oleg Bartunov. | 27 August 2005, 18:04:49 UTC |
5824d02 | Tom Lane | 27 August 2005, 16:22:48 UTC | Get the MIPS assembler syntax right. Also add a separate sync command; the reference I consulted yesterday said SC does a SYNC, but apparently this is not true on newer MIPS processors, so be safe. | 27 August 2005, 16:22:48 UTC |
338e28e | Bruce Momjian | 27 August 2005, 14:13:21 UTC | Fix typo: sql_information_info -> sql_implementation_info Robert Treat | 27 August 2005, 14:13:21 UTC |
a1a19bb | Bruce Momjian | 27 August 2005, 13:52:07 UTC | Add: > * Allow user-defined types to specify a type modifier at table creation > time | 27 August 2005, 13:52:07 UTC |
542faa8 | Bruce Momjian | 27 August 2005, 13:38:43 UTC | Update: > > Another issue is whether underlying table changes should be reflected > in the view, e.g. should SELECT * show additional columns if they > are added after the view is created. | 27 August 2005, 13:38:43 UTC |
846319d | Tom Lane | 26 August 2005, 22:04:42 UTC | Another try at the inlined MIPS spinlock code. Can't test this myself, but for sure it's not any more broken than the prior version. | 26 August 2005, 22:04:42 UTC |
f9244df | Bruce Momjian | 26 August 2005, 20:38:57 UTC | Add: > o Issue a warning if a change-on-restart-only postgresql.conf value > is modified and the server config files are reloaded > o Mark change-on-restart-only values in postgresql.conf 205a209 > o Fix SELECT '0.01 years'::interval, '0.01 months'::interval | 26 August 2005, 20:38:57 UTC |