aboutsummaryrefslogtreecommitdiff
path: root/src/backend
Commit message (Collapse)AuthorAge
* Update comment description of geo routines and move comment to moreBruce Momjian2006-06-26
| | | | relevant location.
* Tweak dynahash.c to avoid wasting memory space in non-shared hash tables.Tom Lane2006-06-25
| | | | | | | | palloc() will normally round allocation requests up to the next power of 2, so make dynahash choose allocation sizes that are as close to a power of 2 as possible. Back-patch to 8.1 --- the problem exists further back, but a much larger patch would be needed and it doesn't seem worth taking any risks.
* Remove dllinit.c; it was only needed for long-obsolete Cygwin versions,Tom Lane2006-06-22
| | | | and there was some question about its copyright status.
* pg_stop_backup was calling XLogArchiveNotify() twice for the newly createdTom Lane2006-06-22
| | | | | backup history file. Bug introduced by the 8.1 change to make pg_stop_backup delete older history files. Per report from Masao Fujii.
* Standalone backends need pgstat_bestart() too, per Qingqing Zhou.Tom Lane2006-06-22
|
* Remove ancient kluge that kept nodeAgg.c from crashing on UPDATEs involvingTom Lane2006-06-21
| | | | | | aggregates. We just disallowed that, and AFAICS there should be no other cases where direct (non-aggregated) references to input columns are allowed in a query with aggregation and no GROUP BY.
* Disallow aggregate functions in UPDATE commands (unless within a sub-SELECT).Tom Lane2006-06-21
| | | | | | | | This is disallowed by the SQL spec because it doesn't have any very sensible interpretation. Historically Postgres has allowed it but behaved strangely. As of PG 8.1 a server crash is possible if the MIN/MAX index optimization gets applied; rather than try to "fix" that, it seems best to just enforce the spec restriction. Per report from Josh Drake and Alvaro Herrera.
* Eliminate a gratuitously different wording of the 'cannot use aggregate ↵Tom Lane2006-06-21
| | | | function in check constraint' error message.
* Remove redundant gettimeofday() calls to the extent practical withoutTom Lane2006-06-20
| | | | | | | | | | | changing semantics too much. statement_timestamp is now set immediately upon receipt of a client command message, and the various places that used to do their own gettimeofday() calls to mark command startup are referenced to that instead. I have also made stats_command_string use that same value for pg_stat_activity.query_start for both the command itself and its eventual replacement by <IDLE> or <idle in transaction>. There was some debate about that, but no argument that seemed convincing enough to justify an extra gettimeofday() call.
* Split definitions for md5.c out of crypt.h and into their own headerTom Lane2006-06-20
| | | | | | | | | libpq/md5.h, so that there's a clear separation between backend-only definitions and shared frontend/backend definitions. (Turns out this is reversing a bad decision from some years ago...) Fix up references to crypt.h as needed. I looked into moving the code into src/port, but the headers in src/include/libpq are sufficiently intertwined that it seems more work than it's worth to do that.
* Take the statistics collector out of the loop for monitoring backends'Tom Lane2006-06-19
| | | | | | | current commands; instead, store current-status information in shared memory. This substantially reduces the overhead of stats_command_string and also ensures that pg_stat_activity is fully up to date at all times. Per my recent proposal.
* Don't try to call posix_fadvise() unless <fcntl.h> supplies a declarationTom Lane2006-06-18
| | | | | | | | for it. Hopefully will fix core dump evidenced by some buildfarm members since fadvise patch went in. The actual definition of the function is not ABI-compatible with compiler's default assumption in the absence of any declaration, so it's clearly unsafe to try to call it without seeing a declaration.
* Merge postmaster and postgres command into just postgres. postmasterPeter Eisentraut2006-06-18
| | | | | symlink is kept for now for compatibility. To call single-user mode, use postgres --single.
* Fix a couple of obvious problems in DROP IF EXISTS patch.Tom Lane2006-06-16
|
* DROP ... IF EXISTS for the following cases:Andrew Dunstan2006-06-16
| | | | language, tablespace, trigger, rule, opclass, function, aggregate. operator, and cast.
* Fix problems with cached tuple descriptors disappearing while still in useTom Lane2006-06-16
| | | | | | | | | | by creating a reference-count mechanism, similar to what we did a long time ago for catcache entries. The back branches have an ugly solution involving lots of extra copies, but this way is more efficient. Reference counting is only applied to tupdescs that are actually in caches --- there seems no need to use it for tupdescs that are generated in the executor, since they'll go away during plan shutdown by virtue of being in the per-query memory context. Neil Conway and Tom Lane
* Test for POSIX_FADV_DONTNEED to use posix_fadvise().Bruce Momjian2006-06-16
|
* Use posix_fadvise() to avoid kernel caching of WAL contents on WAL fileBruce Momjian2006-06-15
| | | | | | close. ITAGAKI Takahiro
* Remove the limit on the number of entries allowed in catcaches, andTom Lane2006-06-15
| | | | | | | | | | | | | | | | | | remove the infrastructure needed to enforce the limit, ie, the global LRU list of cache entries. On small-to-middling databases this wins because maintaining the LRU list is a waste of time. On large databases this wins because it's better to keep more cache entries (we assume such users can afford to use some more per-backend memory than was contemplated in the Berkeley-era catcache design). This provides a noticeable improvement in the speed of psql \d on a 10000-table database, though it doesn't make it instantaneous. While at it, use per-catcache settings for the number of hash buckets per catcache, rather than the former one-size-fits-all value. It's a bit silly to be using the same number of hash buckets for, eg, pg_am and pg_attribute. The specific values I used might need some tuning, but they seem to be in the right ballpark based on CATCACHE_STATS results from the standard regression tests.
* Do a CHECK_FOR_INTERRUPTS immediately before terminating a fastpathTom Lane2006-06-14
| | | | | | | | | function call. Previously, there may have been no CHECK_FOR_INTERRUPTS at all in the fastpath code path, making it impossible to cancel an operation such as \lo_import externally. This addition doesn't ensure you can cancel, since your SIGINT may arrive while the backend is idle waiting for the client, but it gives the largest window we can easily provide. Noted while experimenting with new control-C code for psql.
* Avoid use of C commment inside C comment from recent Win32 int overflow patch.Bruce Momjian2006-06-12
|
* Prevent Win32 from displaying a popup box on backend crash. Instead letBruce Momjian2006-06-12
| | | | | | the postmaster deal with it. Magnus Hagander
* Win32 can't catch the exception thrown by INT_MIN / -1 or INT_MIN * -1,Bruce Momjian2006-06-12
| | | | | | | so on that platform we test for those before the computation and throw an "out of range" error. Backpatch to 8.1.X.
* Minor code cleanup: make the WIN32 case less gratuitously different fromTom Lane2006-06-12
| | | | the other platform-specific cases in ps_status.
* Fix Assert failure when a fastpath function call is attempted inside anTom Lane2006-06-11
| | | | | | | | | | | | | | | already-aborted transaction block. GetSnapshotData throws an Assert if not in a valid transaction; hence we mustn't attempt to set a snapshot for the function until after checking for aborted transaction. This is harmless AFAICT if Asserts aren't enabled (GetSnapshotData will compute a bogus snapshot, but it doesn't matter since HandleFunctionRequest will throw an error shortly anywy). Hence, not a major bug. Along the way, add some ability to log fastpath calls when statement logging is turned on. This could probably stand to be improved further, but not logging anything is clearly undesirable. Backpatched as far as 8.0; bug doesn't exist before that.
* Revert sampling patch for EXPLAIN ANALYZE; it turns out to be too unreliableTom Lane2006-06-09
| | | | | because node timing is much less predictable than the patch expects. I kept the API change for InstrStopNode, however.
* Fix bootstrap.c so that database startup process and bgwriter properly releaseTom Lane2006-06-08
| | | | | | | | | | | | LWLocks during a panic exit. This avoids the possible self-deadlock pointed out by Qingqing Zhou. Also, I noted that an error during LoadFreeSpaceMap() or BuildFlatFiles() would result in exit(0) which would leave the postmaster thinking all is well. Added a critical section to ensure such errors don't allow startup to proceed. Backpatched to 8.1. The 8.0 code is a bit different and I'm not sure if the problem exists there; given we've not seen this reported from the field, I'm going to be conservative about backpatching any further.
* Remove Win32 file, moved to /port.Bruce Momjian2006-06-08
| | | | Magnus Hagander
* Remove obsolete comment about VACUUM FULL: it takes buffer content locksTom Lane2006-06-08
| | | | | now, and must do so to ensure bgwriter doesn't write a page that is in process of being compacted.
* Allow timezone names in SQL strings,Bruce Momjian2006-06-07
| | | | | | '2006-05-24 21:11 Americas/New_York'::timestamptz Joachim Wieland
* Prepare code to be built by MSVC:Bruce Momjian2006-06-07
| | | | | | | | | | o remove many WIN32_CLIENT_ONLY defines o add WIN32_ONLY_COMPILER define o add 3rd argument to open() for portability o add include/port/win32_msvc directory for system includes Magnus Hagander
* Per previous analysis, the most correct notion of SampleOverhead is thatTom Lane2006-06-07
| | | | | | | it is just the total time to do INSTR_TIME_SET_CURRENT(), and not any of the other code involved in InstrStartNode/InstrStopNode. Even though I fear we may end up reverting this patch altogether, we may as well have the most correct version in our CVS archive.
* Have REINDEX use NOTICE, so it can be suppressed.Bruce Momjian2006-06-07
|
* Remove "fuzzy comparison" logic in qsort comparison function forTom Lane2006-06-07
| | | | | | | | | | | | choose_bitmap_and(). It was way too fuzzy --- per comment, it was meant to be 1% relative difference, but was actually coded as 0.01 absolute difference, thus causing selectivities of say 0.001 and 0.000000000001 to be treated as equal. I believe this thinko explains Maxim Boguk's recent complaint. While we could change it to a relative test coded like compare_fuzzy_path_costs(), there's a bigger problem here, which is that any fuzziness at all renders the comparison function non-transitive, which could confuse qsort() to the point of delivering completely wrong results. So forget the whole thing and just do an exact comparison.
* Use INFO-level for reindex messages, rather than NOTICE.Bruce Momjian2006-06-07
| | | | Euler Taveira de Oliveira
* Make the planner estimate costs for nestloop inner indexscans on the basisTom Lane2006-06-06
| | | | | | | | | | | | | | | | | | | | | that the Mackert-Lohmann formula applies across all the repetitions of the nestloop, not just each scan independently. We use the M-L formula to estimate the number of pages fetched from the index as well as from the table; that isn't what it was designed for, but it seems reasonably applicable anyway. This makes large numbers of repetitions look much cheaper than before, which accords with many reports we've received of overestimation of the cost of a nestloop. Also, change the index access cost model to charge random_page_cost per index leaf page touched, while explicitly not counting anything for access to metapage or upper tree pages. This may all need tweaking after we get some field experience, but in simple tests it seems to be giving saner results than before. The main thing is to get the infrastructure in place to let cost_index() and amcostestimate functions take repeated scans into account at all. Per my recent proposal. Note: this patch changes pg_proc.h, but I did not force initdb because the changes are basically cosmetic --- the system does not look into pg_proc to decide how to call an index amcostestimate function, and there's no way to call such a function from SQL at all.
* While making the seq_page_cost changes, I was struck by the fact thatTom Lane2006-06-05
| | | | | | | | | | | | | cost_nonsequential_access() is really totally inappropriate for its only remaining use, namely estimating I/O costs in cost_sort(). The routine was designed on the assumption that disk caching might eliminate the need for some re-reads on a random basis, but there's nothing very random in that sense about sort's access pattern --- it'll always be picking up the oldest outputs. If we had a good fix on the effective cache size we might consider charging zero for I/O unless the sort temp file size exceeds it, but that's probably putting much too much faith in the parameter. Instead just drop the logic in favor of a fixed compromise between seq_page_cost and random_page_cost per page of sort I/O.
* Increase the default value of cpu_index_tuple_cost from 0.001 to 0.005.Tom Lane2006-06-05
| | | | | | This shouldn't affect simple indexscans much, while for bitmap scans that are touching a lot of index rows, this seems to bring the estimates more in line with reality. Per recent discussion.
* Add a GUC parameter seq_page_cost, and use that everywhere we formerlyTom Lane2006-06-05
| | | | | | | | assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
* Don't choke during startup if the environment offers an invalid valueTom Lane2006-06-03
| | | | | | | | | | for LC_MESSAGES; instead, just press forward, leaving the effective setting at 'C'. There is not any very good reason to complain when we are going to replace the value soon with whatever postgresql.conf says. This change should solve the occasionally-reported problem of initdb failing with 'failed to initialize lc_messages'; the current theory is that that is a reflection of either wrong LANG/LC_MESSAGES or completely broken locale support.
* Fix copy-and-pasteo in Russian translation: message complaining aboutTom Lane2006-06-03
| | | | | HAVE_INT64_TIMESTAMP was mentioning PG_CONTROL_VERSION instead. Victor Snezhko
* Make PG_MODULE_MAGIC required in shared libraries that are loaded intoTom Lane2006-05-31
| | | | | the server. Per discussion, there seems no point in a waiting period before making this required.
* Magic blocks don't do us any good unless we use 'em ... so install oneTom Lane2006-05-30
| | | | in every shared library.
* Code review for magic-block patch. Remove separate header file pgmagic.h,Tom Lane2006-05-30
| | | | | | | | | | as this seems only likely to create headaches for module developers. Put the macro in the pre-existing fmgr.h file instead. Avoid being too cute about how many fields we can cram into a word, and avoid trying to fetch from a library we've already unlinked. Along the way, it occurred to me that the magic block really ought to be 'const' so it can be stored in the program text area. Do the same for the existing data blocks for PG_FUNCTION_INFO_V1 functions.
* Code review for EXPLAIN patch. Fix some typos, make it behave sanelyTom Lane2006-05-30
| | | | | across multiple loops, get rid of the shaky assumption that exactly one tuple is returned per node iteration.
* PostmasterIsAlive test really ought to be in the inner loop for safety.Tom Lane2006-05-30
|
* Add pgmagic header block to store compile-time constants:Bruce Momjian2006-05-30
| | | | | | | | | | | | | | | | | | | | | It now only checks four things: Major version number (7.4 or 8.1 for example) NAMEDATALEN FUNC_MAX_ARGS INDEX_MAX_KEYS The three constants were chosen because: 1. We document them in the config page in the docs 2. We mark them as changable in pg_config_manual.h 3. Changing any of these will break some of the more popular modules: FUNC_MAX_ARGS changes fmgr interface, every module uses this NAMEDATALEN changes syscache interface, every PL as well as tsearch uses this INDEX_MAX_KEYS breaks tsearch and anything using GiST. Martijn van Oosterhout
* Make EXPLAIN sampling smarter, to avoid excessive sampling delay.Bruce Momjian2006-05-30
| | | | Martijn van Oosterhout
* Clarify the HINT for 'checkpoint request failed', per recent complaintTom Lane2006-05-30
| | | | demonstrating that its intent wasn't obvious.
* Move conversion dependency patch to the proper branch, out of 8.1.X,Bruce Momjian2006-05-30
| | | | into HEAD.
generated by cgit v1.2.3 (git 2.25.1) at 2025-05-28 09:04:44 +0000