New patches: [fix for initialize-user-parameters alex.mizrahi@gmail.com**20080413131102] { hunk ./src/elephant/controller.lisp 143 - (setq variable it)))) + (set variable it)))) } [serializer thread safety (temp) alex.mizrahi@gmail.com**20080528152446] { hunk ./src/elephant/serializer2.lisp 116 - (declare (type fixnum *circularity-initial-hash-size*)) - (if (= 0 (length *circularity-hash-queue*)) - (make-hash-table :test 'eq :size *circularity-initial-hash-size*) - (ele-with-fast-lock (*serializer-fast-lock*) - (vector-pop *circularity-hash-queue*)))) + (or (ele-with-fast-lock (*serializer-fast-lock*) + (and (plusp (length *circularity-hash-queue*)) + (vector-pop *circularity-hash-queue*))) + (make-hash-table :test 'eq + :size (the fixnum *circularity-initial-hash-size*)))) hunk ./src/elephant/serializer2.lisp 143 - (if (= 0 (length *circularity-vector-queue*)) + (or (ele-with-fast-lock (*serializer-fast-lock*) + (and (plusp (length *circularity-vector-queue*)) + (vector-pop *circularity-vector-queue*))) hunk ./src/elephant/serializer2.lisp 147 - :fill-pointer 0 :adjustable t) - (ele-with-fast-lock (*serializer-fast-lock*) - (vector-pop *circularity-vector-queue*)))) + :fill-pointer 0 :adjustable t))) } [race condition in memutil alex.mizrahi@gmail.com**20080530141323] { hunk ./src/memutil/memutil.lisp 152 - (if (= (length *buffer-streams*) 0) - (make-buffer-stream) - (ele-with-fast-lock (*buffer-streams-lock*) - (vector-pop *buffer-streams*)))) + (or (ele-with-fast-lock (*buffer-streams-lock*) + (and (plusp (length *buffer-streams*)) + (vector-pop *buffer-streams*))) + (make-buffer-stream))) } [Fix MAKE-LOCAL-NAME so it produces keywords instead of randomly package-scoped reader-interned symbols. polzer@gnu.org**20080430121024] hunk ./src/db-postmodern/pm-sql.lisp 134 - (read-from-string (format nil "~a~a" (executor-prefix ex) name))) + (intern (format nil "~a~a" (executor-prefix ex) name) "KEYWORD")) [MAP-INVERTED-INDEX-1 test fixed alex.mizrahi@gmail.com**20080604140448] { hunk ./tests/testindexing.lisp 669 - (with-open-store (`(:bdb "/tmp/store/")) hunk ./tests/testindexing.lisp 674 - x) 'foo-1 'str :start "f" :end "fz" :collect t))))) + x) 'foo-1 'str :start "f" :end "fz" :collect t)))) } Context: [Fix map-btree failure to check if first value in a range map is out of range Ryszard Szopa **20080513140503] [CLBUILDreadiness Robert L. Read**20080514052710] [LesliePolzerNilOIDpatch Robert L. Read**20080506032238] [un-disabled tests for SBCL alex.mizrahi@gmail.com**20080410070805] [Disabling threading tests for SBCL Robert L. Read**20080410015544] [Chun Tian's conditional for lispworks slot-definition-allocation sross@common-lisp.net**20080416161010] [spelling errors Robert L. Read**20080408140049] [Changed erroneous statement in tutorial that index comparison uses EQUALP. polzer@gnu.org**20080226123252] [db-postmodern: sync-cache type handling fix alex.mizrahi@gmail.com**20080326222018] [test concurrency extended alex.mizrahi@gmail.com**20080326203147] [transaction restart support plus extended features alex.mizrahi@gmail.com**20080326203005] [DB-POSTMODERN: remove DBPM-ERROR; don't attempt to remove an already prepared statement (pointless since the txn is aborted at the time); defer all errors to txn handler (but warn and print the offending statement) polzer@gnu.org**20080317171254] [DB-POSTMODERN: support transaction retries; handle deadlock; add savepoint utility functions; add warnings to help debugging problematic conditions. polzer@gnu.org**20080306124528] [added BORDEAUX-THREADS dependency and changed PM controller to use it instead of SB-THREAD stuff. polzer@gnu.org**20080306124512] [added concurrency test cases. polzer@gnu.org**20080306124407] [DB-POSTMODERN: reap old connections when a new one is requested. polzer@gnu.org**20080227150322] [Check for unbound slot; potential fix for a compiler error eslick@common-lisp.net**20080226195839] [Fix test dependence for ff-index-test eslick@common-lisp.net**20080226151654] [Improve berkeley DB version agnostic code eslick@common-lisp.net**20080226151453 Added an error message to configure my-config.sexp and made sure we load it directly from my-config.sexp so that we get it right at load time. Prior patch didn't override default until after load time. ] [Support for multiple BDB versions eslick@common-lisp.net**20080226150114] [db-bdb updated to BerkeleyDB 4.6 kazennikov@gmail.com**20071230140731 Changed only BDB constants as upgrade 4.5 -> 4.6 they were changed. A kind of hack perhaps. But it works. The testing was not excessive, but it works well for my project. ] [add test for STRING types (as opposed to SIMPLE-STRING types) polzer@gnu.org**20080222081256] [Refactor UTF{16,32}LE serializers. polzer@gnu.org**20080222084824] [Enable multiple process connections to a BDB data-store via DB_REGISTER option eslick@common-lisp.net**20080225222713] [Enable multi-store indexed classes eslick@common-lisp.net**20080223184504] [Change semantics of transaction :retry-count from tries to retries eslick@common-lisp.net**20080221031015] [Minor edits, fixed a comment, fixed a db-lisp out of date error eslick@common-lisp.net**20080221024151] [Alex's patch for 8.3 read@robertlread.net**20080217223512 I entered here the patch from Alex of 2088/02/16 which apparently makes this compatible with Postgres 8.3. It is green for me on all tests on Posgres 8.1, so I am committing it. ] [mtype change in dcm read@robertlread.net**20080215135054] [controller-doc-improvement read@robertlread.net**20080210155716] [tutorial read@robertlread.net**20080203161532] [db-postmodern: transaction retry handling, serializable isolation alex.mizrahi@gmail.com**20080318155444] [db-postmodern: small sync cache fix alex.mizrahi@gmail.com**20080318155129 cache was discarded in empty txn ] [moved cache-instance into initial-persistent-setup alex.mizrahi@gmail.com**20080120142436 because it was bypassed by recreate-instance otherwise ] [accessor name in tests change alex.mizrahi@gmail.com**20080116222405] [db-postmodern: pm-btree initialization fixed alex.mizrahi@gmail.com**20080116222316] [recreate-instance stuff improved alex.mizrahi@gmail.com**20080116220138] [Fix instance deserialization to bypass initialization protocol sross@common-lisp.net*-20071214141938] [non-keyword-accessors sross@common-lisp.net**20080113173616 allows lispworks to run tests. ] [function-call-key-form sross@common-lisp.net**20080113173547] [documentation type fix read@robertlread.net**20080111151124] [Fix the use of internal symbol of sb-kernel in memutils Leonardo Varuzza **20071230000120 memutil.lisp use the functions sb-kernel::copy-*-from-system-area, which aren't exported in the latest version of sbcl. Fix it adding the :: when appropriate ] [db-bdb bugfix: when bdb key comparison compared only the first half of utf16 strings kazennikov@gmail.com**20071230141055] [db-postmodern: removed specialized map-index alex.mizrahi@gmail.com**20080107134012 because pure cursor version works fine and is more robust ] [cursor-duplicate removed from db-postmodern Henrik Hjelte*-20071124163701] [db-postmodern removed possiblity of using NIL as a key in btrees Henrik Hjelte**20071124163828] [cursor-duplicate removed from db-postmodern Henrik Hjelte**20071124163701] [Ensure set-db-synch is defined before pset is loaded sross@common-lisp.net**20071214145041] [Fix instance deserialization to bypass initialization protocol sross@common-lisp.net**20071214141938] [db-postmodern fix map-index optimization bug Henrik Hjelte **20080104151644] [db-postmodern: cursors re-implemented alex.mizrahi@gmail.com**20071215191805] [db-postmodern: optimized form-slot-key for persistent-slot-reader alex.mizrahi@gmail.com**20071207200835 it uses SBCL internal function now, for other implementation it's less optimized. ] [db-postmodern: small example update alex.mizrahi@gmail.com**20071207200630] [db-postmodern: optimized map-index for -by-value case alex.mizrahi@gmail.com**20071207195402] [Fix to from-end traversal of new map-index eslick@common-lisp.net**20071130223524] [New map-index implementation eslick@common-lisp.net**20071130222620] [Cheaper get-instance-by-value eslick@common-lisp.net**20071130222520] [removed a little compiler warning (typo) Henrik Hjelte**20071122151929] [remove kind-hints parameter from add-index Henrik Hjelte**20071122151046 Probably a coming feature from Ian, but right now it breaks the generic function add-index and thus postmodern, so I removed it for now. ] [TAG ELEPHANT-0-9-1 ieslick@common-lisp.net**20071116153634] [Fixes to enable the docs to build (on OS X / SBCL using 'make' in elephant/doc) eslick@common-lisp.net**20071104204802] [a little comment update Henrik Hjelte**20071106080259] [postmodern removed ugly-fix from pm-btree-index Henrik Hjelte**20071106080216 and made char-columns hardcoded (removed other option). ] [random test for serializer Henrik Hjelte**20071101144320] [POSTMODERN-tests include hints to configure postgres Henrik Hjelte**20071101102627] [db-postmodern fixed buggy cursor-delete fix secondary-cursor Henrik Hjelte**20071101100700] [postmodern remove obsolete comment about weak tables Henrik Hjelte**20071031023318] [postmodern texinfo file Henrik Hjelte**20071030185853] [db-postmodern update the ugly map-index quick fix Henrik Hjelte**20071030181310] [db-postmodern secondary cursor should be closed after removing values Henrik Hjelte**20071030181154] [Postmodern backend: connection spec now accepts :port keyword argument, to specify the port. Similar to Postmodern's connection spec syntax. tjg@pentaside.org**20071024152639] [Fix some test harness issues for lispworks eslick@common-lisp.net**20071028225431] [Fix a typo eslick@common-lisp.net**20071028223403] [Fix signaling test to bind error appropriate for the given lisp eslick@common-lisp.net**20071028192553] [Patch to use core lisp comparison predicates, including fixes to sql cursors and removing a test that works by accident under BDB due to the inability to compare standard objects eslick@common-lisp.net**20071028191358] [Fix bugs that showed up in migration test suite; some test harness detritus and a bug in the SQL txn handling implementation eslick@common-lisp.net**20071024025205] [(#18) Preliminary migration-oriented GC, very slow. Also added warning print vars and did some preliminary work on (#48) eslick@common-lisp.net**20071024010932] [(#40) Allow delete while mapping; add tests; fix more test dependencies; fix bug in map-index eslick@common-lisp.net**20071023031831] [(#19) Fixed increment cursor on cursor-put eslick@common-lisp.net**20071023004356] [Fix bugs in recent changes and tests for change class and character indexing; tests are green eslick@common-lisp.net**20071023003026] [(#7) Delete slot data during instance edits in change-class and redefine-class; optional warning conditions eslick@common-lisp.net**20071022235935] [Fix a defaults bug in manual transaction handling eslick@common-lisp.net**20071022235855] [Fixed a bug in cursor-prev and added a test for standard btree cursors (was missing! wow!) eslick@common-lisp.net**20071022212318] [Add test for characters as index keys eslick@common-lisp.net**20071022194149] [Fix character comparison in BDB data store and lisp-compare functions eslick@common-lisp.net**20071022162238] [Fixed mop test dependencies for fiveam conversion eslick@common-lisp.net**20071022140438] [Fix lisp comparisons for map-index to mirror 0.9p1 eslick@common-lisp.net**20071022135848 Forgot to push patch to lisp-compare<=/=/< functions from 0.9 CVS to 091 darcs. Fixed in merge branch. ] [Fix FiveAM test dependencies, some Allegro issues, some mis-merges, etc. See diff for details. eslick@common-lisp.net**20071019213723] [resolve merge conflicts between eslick working branch and postmodern branch eslick@common-lisp.net**20071019160331] [Most recent edits; small bug fixes and query testing ieslick@common-lisp.net**20071019153850] [Add test for unicode values in slots from 0.9p1 ieslick@common-lisp.net**20070627145755] [Enable UTF32le unicode support for OpenMCL 1.1 ieslick@common-lisp.net**20070627143750] [Pset wasn't persistent ieslick@common-lisp.net**20070612135351] [Last CVS branch bug fix: utf16 serializer ieslick@common-lisp.net**20070606170001] [Keep to date with CVS for release tag 0.9 ieslick@common-lisp.net**20070531195635] [First pass complete cl-prevalence backend; still a few bugs to work out ieslick@common-lisp.net**20070509001453] [Cleanup and export instance caching so data stores can override the protocol ieslick@common-lisp.net**20070509001327] [Update elephant code version to 0.9.1 ieslick@common-lisp.net**20070509001216] [Cleanup persistent object printing to handle unbound oids ieslick@common-lisp.net**20070509001123] [Simple augmentation of debugging model in deserializer ieslick@common-lisp.net**20070507031150] [Fix map-legacy-names bug for null case ieslick@common-lisp.net**20070507031026] [Prevalence fixes for duplicate and get cursor operations ieslick@common-lisp.net**20070505174213] [Test duplicate operations without depending on primary key ordering part 1 ieslick@common-lisp.net**20070505170910] [A tiny convenience. rread@common-lisp.net**20070911160015] [Cleaning up some type declarations rread@common-lisp.net**20070911155928] [Somehow this was fixed before, and then regressed again in the rread@common-lisp.net**20070911155714 current code. I have added a test which actually exercises, according the the XP discipline. It is a very inelegant test, but it is a hard to exercise problem. ] [Made enable-sync-cache more efficient and safe Henrik Hjelte**20070926015756] [db-postmodern ignore errors around close cursor Henrik Hjelte**20070921053113] [added sh script for flushing logs sample alex.mizrahi@gmail.com**20070920095806] [fixes in pm-cache alex.mizrahi@gmail.com**20070920071646] [global-sync-cache alex.mizrahi@gmail.com**20070919113321] [TAG PostmodernImprovements ieslick@common-lisp.net**20070822200524] [un-disabled instance caching alex.mizrahi@gmail.com**20070913141409] [txn btree value cache alex.mizrahi@gmail.com**20070912191553] [pm-btree make-plpgsql-insert/update duplicates handling fixed alex.mizrahi@gmail.com**20070905124306] [fix type declaration in get-instances-by-range alex.mizrahi@gmail.com**20070904123721] [intern to proper package in make-derived-name alex.mizrahi@gmail.com**20070904123559] [db-postmodern safe-ignore-postgres-error on create-language Henrik Hjelte**20070823153914] [do-test-spec jumps into debugger by default Henrik Hjelte**20070823145758] [db-postmodern do not use postmoderns connection pooling Henrik Hjelte**20070823094751] [db-postmodern remove meaningless function Henrik Hjelte**20070823094715] [db-postmodern rename with-conn with-postmodern-conn Henrik Hjelte**20070823091456] [db-postmodern create-language on initialization (Robert L. Read) Henrik Hjelte**20070823090504] [db-postmodern ignore-errors changed to handler-case Henrik Hjelte**20070823085526] [db-postmodern update in initialization code Henrik Hjelte**20070823085018] [db-postmodern bugfix with transaction handling Henrik Hjelte**20070823084932] [db-postmodern attempt to solve uncommon error 42P01 with prepared statements Henrik Hjelte**20070823084724] [changed test for serializer Henrik Hjelte**20070822020617] [db-postmodern: execute-transaction does not use magic macro Henrik Hjelte**20070820201508] [db-postmodern print-object on cursor Henrik Hjelte**20070814090413] [db-postmodern bugfix again map-index patch is always needed Henrik Hjelte**20070814090244] [db-postmodern print-object on pm-btree Henrik Hjelte**20070814081715] [db-postmodern ignore-errors when preparing-query Henrik Hjelte**20070808093908 An ugly fix that should be solved at some point ] [db-postmodern minor changes Henrik Hjelte**20070731073031] [db-postmodern make char-columns default Henrik Hjelte**20070726054334] [db-postmodern some with-trans-and-vars changed to with-vars Henrik Hjelte**20070725220146] [db-postmodern cursor-set refactoring and bugfix Henrik Hjelte**20070725220115] [typo fix Ties Stuij**20070724111443] [bugfix cursor-set on empty tree Henrik Hjelte**20070722140929] [db-postmodern some refactoring Henrik Hjelte**20070722113311] [db-postmodern misc cleaning up Henrik Hjelte**20070722075640] [db-postmodern some cleaning up Henrik Hjelte**20070722074116] [make a fixture for indexing tests Henrik Hjelte**20070722070302] [db-postmodern fix for problem with map-index Henrik Hjelte**20070722061401 elephant apparently requires btrees with strings as keys to be sorted correctly according to lisp-compare<=. This works for cl-sql which sorts things in memory, but not for the postmodern design which relies on the database to sort things. We need to either change elephant or implement db-postmodern differently. This fix changes elephant internals temporarily when using the db-postmodern backend. Not pretty. ] [some more basic indexing tests Henrik Hjelte**20070722060733] [testcase pcursor2-on-string Henrik Hjelte**20070722060604] [db-postmodern new implementation of cursor-set Henrik Hjelte**20070722030035 Solved a bug with map-index for strings. Works with testcases but needs cleaning up.. ] [test larger-indexing-with-string Henrik Hjelte**20070721180537 fails on postmodern, works with bdb and clsql ] [two more indexing-basic tests for completeness Henrik Hjelte**20070721160313] [two more trivial map-index tests Henrik Hjelte**20070721153616] [removed-kv tests are merged to one test Henrik Hjelte**20070721150018 removed old outcommented code ] [testcollections changed to fiveam Henrik Hjelte**20070721145633] [fiveam make a default testsuite elephant-tests Henrik Hjelte**20070721145533] [Test framework changed to FiveAM Henrik Hjelte**20070721044404] [file elephant-tests split into several files Henrik Hjelte**20070721022740] [pm-indexed-btree, remove cache slot ties**20070717090830 ] [postmodern disable cache Henrik Hjelte**20070714161443 Because it may cause conflicts when different processes access the same database ] [PuttingIndexOnMd5column rread@common-lisp.net**20070710202059 I haven't gotten Henrik's approval of this yet, but it seems obvious (and my test bear out) that this is the column that should be indexed. ] [TryingToSurviveSerializationErrors rread@common-lisp.net**20070710201826 This is an attempt to survive serialization errrors. Rob may be the only person who ever has these (having a live database since around 0.3), and this method is almost a non-method --- but at least it doesn't make your whole attempt to load a database grind to a halt. ] [keysize test rread@common-lisp.net**20070710201629 This is a test of how big a key can get. It was introduced when debugging the postmodern stuff. Unfortunately, there is a limit, which we should work to overcome. ] [db-postmodern some cleaning Henrik Hjelte**20070709191312] [Fixed Roberts bug with large blobs Henrik Hjelte**20070709190719 Becuause the bob columb in blob had a unique index. Now the index is on a md5 value of the bob. ] [stress-test make subsets work with sbcl Henrik Hjelte**20070709093108] [db-postmodern send binary data instead of base64 Henrik Hjelte**20070709092113] [db-postmodern blob table bid column is now a 64 bit integer Henrik Hjelte**20070708152711] [db-postmodern, treat strings as objects. Henrik Hjelte**20070706033714 To avoid this: Database error 54000: index row size 3172 exceeds btree maximum, 2713 Values larger than 1/3 of a buffer page cannot be indexed. Consider a function index of an MD5 hash of the value, or use full text indexing. ] [db-postmodern some cleaning Henrik Hjelte**20070706033345] [updated install-bdb.sh in henrik/contrib Henrik Hjelte **20070608144005] [btree-exec-prepared rewritten Henrik Hjelte**20070706005537] [postmodern integration rread@common-lisp.net**20070704214111] [postmodern tests file Henrik Hjelte **20070511171430] [db-postmodern bug fix for cursor-previous et al Henrik Hjelte **20070511170648] [testcollections, cursor-previous variants, tests may cheat Henrik Hjelte **20070511154905 because of loop for while m. It makes the test silently leave as soon as m is false despite what the value for p and v is. If the m is moved to the always clause everything is tested. ] [postmodern cursor-set need to be able to move to next key Henrik Hjelte **20070511070159] [upgrade-btree-type in db-postmodern Henrik Hjelte **20070510143154 which fixed a bug shown by test pset ] [make Makefile documentation build on sbcl Henrik Hjelte **20070510132113] [make db-postmodern compile and load Henrik Hjelte **20070510093421] [initial import db-postmodern Henrik Hjelte **20070510080504] [Prevalence backend BTree index implementation; all but 11 tests pass (with fake serializer) ieslick@common-lisp.net**20070505043545] [Fix btree-index forbidden ops factorization ieslick@common-lisp.net**20070507023022] [Enable subclass override of controller object cache ieslick@common-lisp.net**20070507022608] [Add specific conditions for errors in the deserializer than implementations can dispatch on; also generalized database error signal is exported ieslick@common-lisp.net**20070507022025] [Remove user file my-config.sexp that should not be repo ieslick@common-lisp.net**20070505221208] [Bug fix for primary key missing in map-index ieslick@common-lisp.net**20070505043920] [Factor our forbidden secondary cursor ops ieslick@common-lisp.net**20070502191218] [Resolve README Conflict ieslick@common-lisp.net**20070502181707] [Prevalence store basic ops + BTrees + some index + some prep for new index architecture ieslick@common-lisp.net**20070502174250] [Initial Import of CVS 0.9.1 RC1 ieslick@common-lisp.net**20070430224026] [Empty repository ieslick@common-lisp.net**20070308220933] Patch bundle hash: 57ef1ab320758e28950b6b363bbeb99d4a42c28e