Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
100 check-ins occurring around 64bec9e6214c6932.
|
2019-01-25
| ||
| 14:23 | Extend fuzzcheck so that it can process dbsqlfuzz cases. Add a collection of interesting dbsqlfuzz cases to the standard test suite. check-in: e2991a7e user: drh tags: trunk | |
| 14:16 | Fix the xFetch method of the "memdb" VFS (used by deserialize) so that it is robust against corrupt database file. Closed-Leaf check-in: 2c1ef40e user: drh tags: dbsqlfuzz-in-fuzzcheck | |
| 13:42 | Fix a couple of assert() statments in btree.c that could fail with corrupt databases. check-in: 5eb5e828 user: dan tags: trunk | |
| 13:03 | In fuzzcheck, activate vdbe_debug for dbsqlfuzz cases when using the -vvvvv verbosity level or above. check-in: 2e6f7c2a user: drh tags: dbsqlfuzz-in-fuzzcheck | |
| 04:43 | Improved automatic detection of dbsqlfuzz cases in fuzzcheck. check-in: 1ef24e89 user: drh tags: dbsqlfuzz-in-fuzzcheck | |
| 04:00 | Add the ability to process dbsqlfuzz cases in fuzzcheck and add an initial set of interesting dbsqlfuzz cases. check-in: fb9074ff user: drh tags: dbsqlfuzz-in-fuzzcheck | |
|
2019-01-24
| ||
| 17:41 | Fix a buffer overread in fts3 that could occur in a prefix query on a corrupted database. check-in: d0d56893 user: dan tags: trunk | |
| 16:51 | While PRAGMA full_column_names is off (the default) do not make changes to the short_column_names pragma when computing the column names for subqueries. Leaf check-in: 718ead55 user: drh tags: literal-column-names | |
| 16:27 | Fix a problem with running ALTER TABLE on a schema that contains expressions of the type "col IN ()" (empty set on RHS of IN operator). check-in: 2d9cd067 user: dan tags: trunk | |
| 16:07 | Omit deprecated PRAGMAs when compiling with the -DSQLITE_OMIT_DEPRECATED option. check-in: 008112bc user: drh tags: trunk | |
| 15:51 | Make sure the column name flags are restored correctly after an error inside of sqlite3ResultSetOfSelect(). check-in: b1601db7 user: drh tags: trunk | |
| 15:16 | Fix a potential problem with "INSERT INTO ... SELECT * FROM" (or VACUUM) statements on a corrupted database. check-in: db4b4c2c user: dan tags: trunk | |
| 14:16 | Change a integer variable in sqlite3VdbeRecordUnpack() to unsigned in order to avoid any possibility of an integer overflow. check-in: 1b536f6f user: drh tags: trunk | |
| 13:36 | Use memmove() instead of memcpy() in a place where buffers might overlap if the database file is badly corrupted, to prevent warnings from ASAN and valgrind. check-in: 65ad6c55 user: drh tags: trunk | |
| 04:44 | Remove an unreachable branch. check-in: 4c976f17 user: drh tags: trunk | |
|
2019-01-23
| ||
| 20:31 | Fix an assert() in vdbemem.c that could fire if the database was corrupt. check-in: a70958cd user: dan tags: trunk | |
| 19:50 | Fix a problem with renaming a table within a schema that contains a composite query that uses a column alias as an ORDER BY term. check-in: 2ca6b8f8 user: dan tags: trunk | |
| 19:25 | Stricter enforcement of cell sizes when doing balancing operations on the btree, in order to catch file corruption sooner. check-in: 12713f32 user: drh tags: trunk | |
| 19:17 | Fix another fts5 crash that can occur if the database is corrupted. check-in: 44ce8baa user: dan tags: trunk | |
| 16:59 | Fix problems with sub-selects in WINDOW definitions. Also rename-column operations when the column being renamed appears in a WINDOW definition that is part of a VIEW or TRIGGER. check-in: 0387cb3a user: dan tags: trunk | |
| 12:19 | Fix a buffer overwrite triggered by a prefix query on a corrupt fts5 table. check-in: 1d8172a9 user: dan tags: trunk | |
|
2019-01-22
| ||
| 21:17 | Fix a buffer overrun that could occur in fts5 if a prefix query is made on a corrupt database. check-in: 1abc4415 user: dan tags: trunk | |
| 20:18 | Do not run shmlock.test as part of the journaltest permutation. check-in: 5b7d0c78 user: dan tags: trunk | |
| 18:20 | Add the SQLITE_ENABLE_EARLY_CURSOR_CLOSE compile-time option which causes read cursors to be closed after their usefulness ends during a two-pass UPDATE. Leaf check-in: 8790368b user: mistachkin tags: branch-3.26 | |
| 16:44 | Use _strdup() instead of strdup() on Windows builds of the CLI, to avoid a compiler warning reported on the mailing list. check-in: a7126a4f user: drh tags: trunk | |
| 16:43 | Fix the SQLITE_DESERIALIZE_READONLY feature so that it does not cause an assertion fault in the pager. check-in: b9eccef7 user: drh tags: trunk | |
| 16:11 | Update dbfuzz2 to set a maximum database size of 100MiB by default, but with the new --max-db-size N option to change that limit. check-in: 21d6bb78 user: drh tags: trunk | |
| 16:06 | Enhancements to deserialize: (1) Add the SQLITE_FCNTL_SIZE_LIMIT file control to set a maximum size for an in-memory database, defaulting to SQLITE_MEMDB_DEFAULT_MAXSIZE or 1GiB. (2) Honor the SQLITE_DESERIALIZE_READONLY flag. (3) Enhance the TCL interface to support -maxsize N and -readonly BOOLEAN. (4) Add the --maxsize option to the ".open" command and on the command-line for the CLI. check-in: 30f08d58 user: drh tags: trunk | |
| 13:45 | Make sure cursors are opened on all indexes for an UPDATE OR REPLACE regardless of whether or not the indexes are partial or contain columns that might need to be updated. check-in: e148cdad user: drh tags: trunk | |
| 12:21 | Fix another segfault caused by a corrupt fts3 database. check-in: ba3b8412 user: dan tags: trunk | |
| 02:34 | Ensure that a key comparison does not read a collating sequence past the end of the KeyInfo, even if the key field of an index is corrupted by having a string in the last column instead of the ROWID. check-in: 058a8006 user: drh tags: trunk | |
|
2019-01-21
| ||
| 23:18 | Enhance the btree search routine so that it does early detection of impossibly large keys and thereby avoids a large malloc() call. check-in: 3ecaaee6 user: drh tags: trunk | |
| 17:57 | Avoid deferencing a freed pointer following an OOM or SQLITE_CORRUPT error in the fts3 xDestroy method. check-in: 505ed9a4 user: dan tags: trunk | |
| 16:12 | Remove a faulty assert() from fts3. check-in: 6c33a303 user: dan tags: trunk | |
| 16:01 | Make sure the extra bytes allocated for the saved cursor position by check-in [160b1e31c0f27257] are initialized to zero. check-in: 27375649 user: drh tags: trunk | |
| 14:55 | Add the --memtrace option to the CLI. check-in: a1e12fa2 user: drh tags: trunk | |
| 14:49 | Minor fix the fallocate.test module change from [7cd56cad5efead5] check-in: 94fb7a47 user: drh tags: trunk | |
| 13:47 | Add the --max-data and --max-as options to dbfuzz2. Also cause dbfuzz2 to show its maximum RSS size upon exit in standalone mode with the -v option. check-in: 7ce93e82 user: drh tags: trunk | |
|
2019-01-20
| ||
| 00:03 | Add the --max-stack option to dbfuzz2. check-in: c11ae4fe user: drh tags: trunk | |
|
2019-01-19
| ||
| 15:55 | Avoid a division-by-zero error in fts5 caused by a corrupt database. check-in: 928e6221 user: dan tags: trunk | |
| 15:27 | Remove a broken assert() triggered by a "PRAGMA max_page_count = N" invocation, where N is larger than the number of pages in the database image, but smaller than the number of pages in the database file. check-in: 7cd56cad user: dan tags: trunk | |
| 14:07 | Fix a problem with using ALTER TABLE to rename a table or column when the database schema contains a trigger or view that itself contains an expression "<column> AND 0". check-in: 908ff7ff user: dan tags: trunk | |
|
2019-01-18
| ||
| 21:17 | Fix a memory leak introduced by [55c5d72a]. check-in: fbd681dc user: dan tags: trunk | |
| 21:12 | Fix a crash in the fts5vocab module caused by including a "term < NULL" term in a WHERE clause. check-in: 9e717c43 user: dan tags: trunk | |
| 21:03 | Fix an infinite loop caused by a corrupt database in fts3. Also an undefined left-shift in fts5. check-in: 55c5d72a user: dan tags: trunk | |
| 20:15 | Avoid passing a NULL pointer to memcpy in fts5, even if the database is corrupt. check-in: acccc980 user: dan tags: trunk | |
| 19:33 | Typecasts added to the GeoPoly extension to avoid harmless UBSAN warnings. check-in: a1f6a093 user: drh tags: trunk | |
| 19:26 | Fix problems causing undefined left-shift operations in the fts3 snippet() function. check-in: b90dbaed user: dan tags: trunk | |
| 18:52 | Avoid integer overflow when computing the array of a bounding box with the rtree_i32 virtual table. check-in: b352f159 user: drh tags: trunk | |
| 17:53 | Handle oversize floating point values carefully when converting to integers for the '%' binary operator. check-in: 048add13 user: drh tags: trunk | |
| 16:06 | Fix a fairly obscure problem allowing an "ALTER TABLE RENAME col TO ..." statement to modify the schema in such a way as to break a reference within a trigger program. check-in: 64bec9e6 user: dan tags: trunk | |
| 14:53 | Use the full 64-bit integer value in the argument to randomblob(). check-in: 05df5f7a user: drh tags: trunk | |
|
2019-01-17
| ||
| 20:19 | Disable the xfer-optimization if the two tables have the same root page due to schema corruption. check-in: f31b3bd2 user: drh tags: trunk | |
| 20:06 | Fix another corruption related crash in fts5. check-in: 4538d9af user: dan tags: trunk | |
| 19:33 | Add the SQLITE_ENABLE_EARLY_CURSOR_CLOSE compile-time option which causes read cursors to be closed after their usefulness ends during a two-pass UPDATE. check-in: 7def6c8e user: drh tags: trunk | |
| 19:11 | Fix a buffer overwrite that could occur when running an fts5 prefix query against a corrupt database. check-in: 3910b563 user: dan tags: trunk | |
| 17:39 | Fix problems with joining two or more fts5_vocab tables that access the same underlying fts5 table. check-in: 49956395 user: dan tags: trunk | |
| 15:47 | Sync up with the latest enhancements on trunk. check-in: 475a179a user: drh tags: apple-osx | |
| 15:40 | Revamp the SrcList allocator routines to be methods of Parse instead of being methods of the "sqlite3" object, so that they can leave better error messages when the SrcList object grows too large. check-in: df08d472 user: drh tags: trunk | |
| 14:34 | Limit the size of SrcList objects to 200 entries (compile-time configurable using -DSQLITE_MAX_SRCLIST=n). The maximum number of tables in a join has always been 64, so this is not a real constraint on capability. Limiting the size of a SrcList prevents DOS attacks (discovered by OSSFuzz) using crazy nexted CTE joins. check-in: 7cac614d user: drh tags: trunk | |
| 04:40 | Fix a corner-case for the logic that cause an insert of a NULL into an INTEGER PRIMARY KEY column to be converted into a valid integer key, when the NULL results from a CASE expression that lacks an ELSE clause. check-in: 9a425051 user: drh tags: trunk | |
| 03:43 | Ensure that the variable-length integer decoder in FTS3 never tries to left-shift a negative number. check-in: 10ffc1fe user: drh tags: trunk | |
| 01:06 | Fix a failure in sqlite3ExprCompare() check-in: 835e2cc5 user: drh tags: trunk | |
|
2019-01-16
| ||
| 20:48 | Fix a problem with fix [b4b57413]. check-in: ca7b7aae user: dan tags: trunk | |
| 19:44 | Fix a problem in the fts3 matchinfo() function with corrupt database handling. check-in: 24ed5fb6 user: dan tags: trunk | |
| 19:26 | Fix a problem in the code generator for sorting results with SRT_EphemTab and a LIMIT clause. check-in: 49fcde2f user: drh tags: trunk | |
| 14:58 | Fix a problem with renaming a column that is used as part of an ORDER BY on a compound SELECT within a database view or trigger. check-in: b4b57413 user: dan tags: trunk | |
| 12:05 | Avoid a dangling pointer comparison when renaming a table that has a trigger that itself contains a window function with an (illegal) column reference in a FOLLOWING expression. check-in: d45bee36 user: dan tags: trunk | |
| 11:38 | Fix a memory leak that could occur in fts3 when handling a corrupt database. check-in: 65cebb06 user: dan tags: trunk | |
|
2019-01-15
| ||
| 20:51 | Fix a problem with ALTER TABLE and vector assignments in UPDATE statements within triggers. check-in: cc6cd753 user: dan tags: trunk | |
| 18:14 | Handle SQL NULL values without crashing in the fts5 snippet() and highlight() functions. check-in: a5e9cc79 user: dan tags: trunk | |
| 16:14 | Fix a buffer overread in fts3 caused by a corrupt record. check-in: e54efd60 user: dan tags: trunk | |
| 15:18 | Have fts5 reject attempts to insert a non-integer, non-null value into a rowid column with SQLITE_MISMATCH. check-in: 4a9483f8 user: dan tags: trunk | |
| 14:44 | Fix a harmless memory leak in the Lemon parser generator utility program. check-in: 1caff0fb user: drh tags: trunk | |
| 14:31 | Fix a problem triggered by DELETE statements with WHERE clauses that use the OR-optimization on some virtual tables. check-in: ecf5caa7 user: dan tags: trunk | |
|
2019-01-14
| ||
| 20:44 | Add scalar function "prefix_length()" to ext/misc/prefixes.c. check-in: 0bf15505 user: dan tags: trunk | |
| 19:13 | Fix a problem causing some Tcl test cases to fail with errors like "expected: [<multiline-whitespace>], got: []". check-in: 11b8a4cb user: dan tags: trunk | |
| 16:16 | Add the "prefixes" table-valued function in the ext/misc folder. check-in: f0fd2163 user: drh tags: trunk | |
| 15:35 | Fix a problem causing a crash if an fts5vocab table was created to query an fts3/4 FTS index. check-in: 9cd64ce4 user: dan tags: trunk | |
| 15:09 | Fix a problem with querying a corrupt fts3 database. check-in: 65e50706 user: dan tags: trunk | |
| 13:32 | When parsing a corrupt record using sqlite3VdbeRecordUnpack(), make sure that the resulting UnpackedRecord object is completely initialized, to avoid (harmless) MSAN warnings. check-in: ddc3697e user: drh tags: trunk | |
| 11:56 | Have fts3 ignore empty sets of parenthesis if built with SQLITE_ENABLE_FTS3_PARENTHESIS. check-in: c93c6b45 user: dan tags: trunk | |
| 05:48 | Avoid reading off the front of a page buffer when balancing a corrupt btree page. check-in: cb505090 user: drh tags: trunk | |
|
2019-01-13
| ||
| 20:23 | In dbfuzz2, avoid using a malloc in the LLVMFuzzerInitialize() initializer routine, so that no memory leaks are reported. Also, show the version of SQLite being used when the -v option is on. check-in: 824f9324 user: drh tags: trunk | |
| 20:17 | Relax the minimum size database file constraint on the dbtotxt utility program. check-in: 97e723d7 user: drh tags: trunk | |
| 20:17 | Improved detection of database corruption while balancing pages from an auto_vacuum database with overflow pages. Test cases in TH3. check-in: 35f04235 user: drh tags: trunk | |
| 00:58 | Move a local variable declaration into the outermost scope in which it is used. This fixes an ASAN warning. check-in: ac3b6021 user: drh tags: trunk | |
|
2019-01-12
| ||
| 21:30 | Prevent unsigned 32-bit integer overflow from leading to a buffer overread inside of an assert(). The problem fixed here is no reachable in production code. check-in: 0f850a25 user: drh tags: trunk | |
| 20:55 | Fix another problem with handling corrupt records in fts5_decode(). check-in: 726e398b user: dan tags: trunk | |
| 16:19 | Fix an off-by-one error on a Goto in the code generator, that only causes problems for a REPLACE on an INTEGER PRIMARY KEY in non-debug builds. Test case in TH3. check-in: e35eb877 user: drh tags: trunk | |
| 14:58 | Fix a problem with corrupt database handling in the fts3 matchinfo() function. check-in: 703646b1 user: dan tags: trunk | |
| 00:45 | Improved detection of shadow table corruption in the fts5_decode() SQL function. check-in: b74e5f3f user: drh tags: trunk | |
| 00:12 | Improved shadow table corruption detection in the matchinfo() function of FTS3. check-in: 567be3bb user: drh tags: trunk | |
| 00:07 | Indicate that the database may be corrupt in the fts3corrupt4.test test script. check-in: 473626d5 user: drh tags: trunk | |
|
2019-01-11
| ||
| 23:08 | Use 64-bit APIs in the fileio.c extension. check-in: b49d56a0 user: drh tags: trunk | |
| 21:34 | Fix a segfault in fts3 prompted by a corrupted database. check-in: 2d7b1d1d user: dan tags: trunk | |
| 19:27 | Fix another problem with corrupt databases in fts5. check-in: 9bf811ea user: dan tags: trunk | |
| 17:41 | Fix a problem handling a negative value in the "number-of-pages" database header field. Also a problem with running "REINDEX tbl" against a virtual table for which the SQL passed to sqlite3_declare_vtab() contains PRIMARY KEY or UNIQUE constraints. check-in: 556dd892 user: dan tags: trunk | |
| 17:20 | Improved detection and reporting of errors in the readfile() extension SQL function. check-in: d2f0b5a4 user: drh tags: trunk | |
| 17:19 | Omit errors about missing SAVEPOINTs when aborting the .archive command in the CLI. check-in: 2a47387b user: drh tags: trunk | |