aboutsummaryrefslogtreecommitdiff
path: root/src/backend
Commit message (Collapse)AuthorAge
* Fix unportable usage of socklen_t: should use ACCEPT_TYPE_ARG3 macroTom Lane2006-01-24
| | | | provided by configure, instead. Per bug #2205.
* Repair longstanding bug in slru/clog logic: it is possible for two backendsTom Lane2006-01-21
| | | | | | | | | | to try to create a log segment file concurrently, but the code erroneously specified O_EXCL to open(), resulting in a needless failure. Before 7.4, it was even a PANIC condition :-(. Correct code is actually simpler than what we had, because we can just say O_CREAT to start with and not need a second open() call. I believe this accounts for several recent reports of hard-to-reproduce "could not create file ...: File exists" errors in both pg_clog and pg_subtrans.
* Fix thinko in autovacuum's test to skip temp tables: want to skip anyTom Lane2006-01-20
| | | | | | temp table not only our own process' tables. It's not real important since vacuum.c will skip temp tables anyway, but might as well make the code do what it claims to do.
* Add some test scaffolding to allow cache-flush stress testing (and I doTom Lane2006-01-19
| | | | mean stress ... system is orders of magnitude slower with this enabled).
* Avoid crashing if relcache flush occurs while trying to load data into anTom Lane2006-01-19
| | | | | | | | | | | | | | | index's support-function cache (in index_getprocinfo). Since none of that data can change for an index that's in active use, it seems sufficient to treat all open indexes the same way we were treating "nailed" system indexes --- that is, just re-read the pg_class row and leave the rest of the relcache entry strictly alone. The pg_class re-read might not be strictly necessary either, but since the reltablespace and relfilenode can change in normal operation it seems safest to do it. (We don't support changing any of the other info about an index at all, at the moment.) Back-patch as far as 8.0. It might be possible to adapt the patch to 7.4, but it would take more work than I care to expend for such a low-probability problem. 7.3 is out of luck for sure.
* It turns out that TablespaceCreateDbspace fails badly if a relcache flushTom Lane2006-01-19
| | | | | | | | | | | | | | occurs when it tries to heap_open pg_tablespace. When control returns to smgrcreate, that routine will be holding a dangling pointer to a closed SMgrRelation, resulting in mayhem. This is of course a consequence of the violation of proper module layering inherent in having smgr.c call a tablespace command routine, but the simplest fix seems to be to change the locking mechanism. There's no real need for TablespaceCreateDbspace to touch pg_tablespace at all --- it's only opening it as a way of locking against a parallel DROP TABLESPACE command. A much better answer is to create a special-purpose LWLock to interlock these two operations. This drops TablespaceCreateDbspace quite a few layers down the food chain and makes it something reasonably safe for smgr to call.
* Fix a tiny memory leak (one List header) in RelationCacheInvalidate().Tom Lane2006-01-19
| | | | | | | This is utterly insignificant in normal operation, but it becomes a problem during cache inval stress testing. The original coding in fact had no leak --- the 8.0 List rewrite created the issue. I wonder whether list_concat should pfree the discarded header?
* Modify pgstats code to reduce performance penalties from oversized stats dataTom Lane2006-01-18
| | | | | | | | | | | | | files: avoid creating stats hashtable entries for tables that aren't being touched except by vacuum/analyze, ensure that entries for dropped tables are removed promptly, and tweak the data layout to avoid storing useless struct padding. Also improve the performance of pgstat_vacuum_tabstat(), and make sure that autovacuum invokes it exactly once per autovac cycle rather than multiple times or not at all. This should cure recent complaints about 8.1 showing much higher stats I/O volume than was seen in 8.0. It'd still be a good idea to revisit the design with an eye to not re-writing the entire stats dataset every half second ... but that would be too much to backpatch, I fear.
* Fix fsync code to test whether F_FULLFSYNC is available, instead ofTom Lane2006-01-17
| | | | assuming it always is on Darwin. Per report from Neil Brandt.
* Repair problems with the result of lookup_rowtype_tupdesc() possibly beingTom Lane2006-01-17
| | | | | | | discarded by cache flush while still in use. This is a minimal patch that just copies the tupdesc anywhere it could be needed across a flush. Applied to back branches only; Neil Conway is working on a better long-term solution for HEAD.
* We neglected to apply domain constraints on UNKNOWN parameters toNeil Conway2006-01-12
| | | | prepared statements, per report from David Wheeler.
* Repair "Halloween problem" in EvalPlanQual: a tuple that's been inserted byTom Lane2006-01-12
| | | | | | | | our own command (or more generally, xmin = our xact and cmin >= current command ID) should not be seen as good. Else we may try to update rows we already updated. This error was inserted last August while fixing the even bigger problem that the old coding wouldn't see *any* tuples inserted by our own transaction as good. Per report from Euler Taveira de Oliveira.
* Improve error messages for missing-FROM-entry cases, as per recent discussion.Tom Lane2006-01-10
|
* Add RelationOpenSmgr() calls to ensure rd_smgr is valid when we try toTom Lane2006-01-07
| | | | | | use it. While it normally has been opened earlier during btree index build, testing shows that it's possible for the link to be closed again if an sinval reset occurs while the index is being built.
* During CatCacheRemoveCList, we must now remove any members that areTom Lane2006-01-07
| | | | | | | | | dead and have become unreferenced. Before 8.1, such members were left for AtEOXact_CatCache() to clean up, but now AtEOXact_CatCache isn't supposed to have anything to do. In an assert-enabled build this bug leads to an assertion failure at transaction end, but in a non-assert build the dead member is effectively just a small memory leak. Per report from Jeremy Drake.
* Fix failure to apply domain constraints to a NULL constant that's added toTom Lane2006-01-06
| | | | an INSERT target list during rule rewriting. Per report from John Supplee.
* Fix Windows-only postmaster code to reject a connection request and continue,Tom Lane2006-01-06
| | | | | | | rather than elog(FATAL), when there is no more room in ShmemBackendArray. This is a security issue since too many connection requests arriving close together could cause the postmaster to shut down, resulting in denial of service. Reported by Yoshiyuki Asaba, fixed by Magnus Hagander.
* Convert Assert checking for empty page into a regular test and elog.Tom Lane2006-01-06
| | | | | The consequences of overwriting a non-empty page are bad enough that we should not omit this test in production builds.
* Fix ReadBuffer() to correctly handle the case where it's trying to extendTom Lane2006-01-06
| | | | | | | | | the relation but it finds a pre-existing valid buffer. The buffer does not correspond to any page known to the kernel, so we *must* do smgrextend to ensure that the space becomes allocated. The 7.x branches all do this correctly, but the corner case got lost somewhere during 8.0 bufmgr rewrites. (My fault no doubt :-( ... I think I assumed that such a buffer must be not-BM_VALID, which is not so.)
* Translation updatePeter Eisentraut2006-01-05
|
* Arrange to set the LC_XXX environment variables to match our locale setup.Tom Lane2006-01-05
| | | | Back-patch of previous fix in HEAD for plperl-vs-locale issue.
* Fix another case in which autovacuum would fail while analyzingTom Lane2006-01-04
| | | | expressional indexes. Per report from Brian Hirt.
* Reset flex state in a way that works for both flex 2.5.4 and 2.5.31.Tom Lane2006-01-02
|
* Rewrite ProcessConfigFile() to avoid misbehavior at EOF, as per reportTom Lane2006-01-01
| | | | | | | | from Andrus Moor. The former state-machine-style coding wasn't actually doing much except obscuring the control flow, and it didn't extend readily to fix this case, so I just took it out. Also, add a YY_FLUSH_BUFFER call to ensure the lexer is reset correctly if the previous scan failed partway through the file.
* Remove DOS line endings ("\r\n") from several .po files. DOS line endingsNeil Conway2006-01-01
| | | | | are inconsistent with the rest of the .po files, and apparently cause problems for Sun's cc. Per report on IRC from "bitvector2".
* Repair EXPLAIN failure when trying to display a plan condition that involvesTom Lane2005-12-30
| | | | | | selection of a field from the result of a function returning RECORD. I believe this case is new in 8.1; it's due to the addition of OUT parameters. Per example from Michael Fuhr.
* Add regression tests for CSV and \., and add automatic quoting of aBruce Momjian2005-12-28
| | | | | | | single column dump that has a \. value, so the load works properly. I also added documentation describing this issue. Backpatch to 8.1.X.
* Our code had:Bruce Momjian2005-12-27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | if (c == '\\' && cstate->line_buf.len == 0) The problem with that is the because of the input and _output_ buffering, cstate->line_buf.len could be zero even if we are not on the first character of a line. In fact, for a typical line, it is zero for all characters on the line. The proper solution is to introduce a boolean, first_char_in_line, that we set as we enter the loop and clear once we process a character. I have restructured the line-reading code in copy.c by: o merging the CSV/non-CSV functions into a single function o used macros to centralize and clarify the buffering code o updated comments o renamed client_encoding_only to encoding_embeds_ascii o added a high-bit test to the encoding_embeds_ascii test for performance o in CSV mode, allow a backslash followed by a non-period to continue being processed as a data value There should be no performance impact from this patch because it is functionally equivalent. If you apply the patch you will see copy.c is much clearer in this area now and might suggest additional optimizations. I have also attached a 8.1-only patch to fix the CSV \. handling bug with no code restructuring.
* Fix long standing Asian multibyte charsets bug.Tatsuo Ishii2005-12-24
| | | | | | | | | | | See: Subject: [HACKERS] bugs with certain Asian multibyte charsets From: Tatsuo Ishii <ishii@sraoss.co.jp> To: pgsql-hackers@postgresql.org Date: Sat, 24 Dec 2005 18:25:33 +0900 (JST) for more details.
* Fix for rearranging encoding id ISO-8859-5 to ISO-8859-8.Tatsuo Ishii2005-12-24
| | | | | | | | | | | Also make the code more robust by searching for target encoding in the internal charset map. Problem reported by Sagi Bashari on 2005/12/21. See "[BUGS] BUG #2120: Crash when doing UTF8<->ISO_8859_8 encoding conversion" on pgsql-bugs list for more details. Back patch for 8.1_STABLE.
* Adjust string comparison so that only bitwise-equal strings are consideredTom Lane2005-12-22
| | | | | | | | | | | | equal: if strcoll claims two strings are equal, check it with strcmp, and sort according to strcmp if not identical. This fixes inconsistent behavior under glibc's hu_HU locale, and probably under some other locales as well. Also, take advantage of the now-well-defined behavior to speed up texteq, textne, bpchareq, bpcharne: they may as well just do a bitwise comparison and not bother with strcoll at all. NOTE: affected databases may need to REINDEX indexes on text columns to be sure they are self-consistent.
* Defend against crash while processing Describe Statement or Describe PortalTom Lane2005-12-14
| | | | | | messages, when client attempts to execute these outside a transaction (start one) or in a failed transaction (reject message, except for COMMIT/ROLLBACK statements which we can handle). Per report from Francisco Figueiredo Jr.
* Fix problem with whole-row Vars referencing sub-select outputs, perTom Lane2005-12-14
| | | | | example from Jim Dew. Add some simple regression tests, since this is an area we seem to break regularly :-(
* Teach deparsing of CASE expressions to cope with the simplified formsTom Lane2005-12-10
| | | | | | that simplify_boolean_equality() may leave behind. This is only relevant if the user writes something a bit silly, like CASE x=y WHEN TRUE THEN. Per example from Michael Fuhr; may or may not explain bug #2106.
* Translation updatesPeter Eisentraut2005-12-09
|
* Fix bgwriter's failure to release buffer pins and open files after anTom Lane2005-12-08
| | | | | error. This probably explains bug #2099 and could also account for mysterious VACUUM hangups.
* In a nestloop inner indexscan, it's OK to use pushed-down baserestrictinfoTom Lane2005-12-06
| | | | | | clauses even if it's an outer join. This is a corner case since such clauses could only arise from weird OUTER JOIN ON conditions, but worth fixing. Per example from Ron at cheapcomplexdevices.com.
* Allow to_char(interval) and to_char(time) to use AM/PM specifications.Bruce Momjian2005-12-03
| | | | | | | | | | | Map them to a single day, so '30 hours' is 'AM'. Have to_char(interval) and to_char(time) use "HH", "HH12" as 12-hour intervals, rather than bypass and print the full interval hours. This is neeeded because to_char(time) is mapped to interval in this function. Intervals should use "HH24", and document suggestion. Allow "D" format specifiers for interval/time.
* Rearrange code in ExecInitBitmapHeapScan so that we don't initialize theTom Lane2005-12-02
| | | | | | | | child plan nodes until we have acquired lock on the relation to scan. The relative order of initialization of plan nodes isn't real important in other cases, but it's critical here because one is supposed to lock a relation before its indexes, not vice versa. The original coding was at least vulnerable to deadlock against DROP INDEX, and perhaps worse things.
* Retry in FileRead and FileWrite if Windows returns ERROR_NO_SYSTEM_RESOURCES.Tom Lane2005-12-01
| | | | | | Also add a retry for Unixen returning EINTR, which hasn't been reported as an issue but at least theoretically could be. Patch by Qingqing Zhou, some minor adjustments by me.
* Check for overflow in strtol() while parsing datetime inputs.Tom Lane2005-12-01
| | | | Michael Fuhr.
* Rearrange code in pg_atoi() to avoid assuming that isspace() cannotTom Lane2005-11-30
| | | | change errno. No reported bugs here, but why take a chance?
* Tweak choose_bitmap_and() heuristics in the light of example provided in bugTom Lane2005-11-30
| | | | | | | #2075: consider an index redundant if any of its index conditions were already used, rather than if all of them were. Also, make the selectivity comparison a bit fuzzy, so that very small differences in estimated selectivities don't skew the results.
* Tweak hash join code to use an additional heuristic for deciding whetherTom Lane2005-11-28
| | | | | | | | it's worth probing the outer relation for emptiness before building the hash table. To wit, if we're rescanning a join previously performed, remember whether we found it nonempty the previous time, and don't bother with the probe if it was nonempty. This buys back the performance lost in examples like Mario Weilguni's.
* Install a more future-proof fix for the snapshot-unset bug just found.Alvaro Herrera2005-11-28
| | | | Per suggestion from Tom Lane.
* Recent changes to allow hash join to exit early given empty input fromTom Lane2005-11-28
| | | | | | | one child or the other had a problem: they did not leave the node in a state that ExecReScanHashJoin would understand. In particular it would tend to fail to reset the child plans when needed. Per report from Mario Weilguni.
* Set a snapshot before running analyze on a single table, to avoid aAlvaro Herrera2005-11-28
| | | | crash when analyzing tables with expressional indexes.
* Get rid of ExecAssignResultTypeFromOuterPlan() and make all plan node typesTom Lane2005-11-23
| | | | | | | | | | | generate their output tuple descriptors from their target lists (ie, using ExecAssignResultTypeFromTL()). We long ago fixed things so that all node types have minimally valid tlists, so there's no longer any good reason to have two different ways of doing it. This change is needed to fix bug reported by Hayden James: the fix of 2005-11-03 to emit the correct column names after optimizing away a SubqueryScan node didn't work if the new top-level plan node used ExecAssignResultTypeFromOuterPlan to generate its tupdesc, since the next plan node down won't have the correct column labels.
* Fix problems with rewriter failing to set Query.hasSubLinks when insertingTom Lane2005-11-23
| | | | | | | a SubLink expression into a rule query. Pre-8.1 we essentially did this unconditionally; 8.1 tries to do it only when needed, but was missing a couple of cases. Per report from Kyle Bateman. Add some regression test cases covering this area.
* Re-run pgindent, fixing a problem where comment lines after a blankBruce Momjian2005-11-22
| | | | | | | | | comment line where output as too long, and update typedefs for /lib directory. Also fix case where identifiers were used as variable names in the backend, but as typedefs in ecpg (favor the backend for indenting). Backpatch to 8.1.X.
generated by cgit v1.2.3 (git 2.25.1) at 2025-08-04 04:59:43 +0000