Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | Minor edit of the architecture page. | Michael Cahill | 2012-02-20 | 1 | -19/+19 | |
| | ||||||
* | Add a "Follow me on Google Groups" banner. | Michael Cahill | 2012-02-20 | 3 | -1/+32 | |
| | ||||||
* | Pass the configuration stack through __wt_btree_reopen: opening for salvage ↵ | Michael Cahill | 2012-02-20 | 3 | -5/+8 | |
| | | | | looks for the "force" configuration. | |||||
* | In __btree_set_root, initialize newcfg to avoid freeing a bad pointer on error. | Michael Cahill | 2012-02-20 | 1 | -0/+1 | |
| | ||||||
* | Fill in the marketing part of the "WiredTiger Architecture" page. | Keith Bostic | 2012-02-19 | 4 | -18/+112 | |
| | | | | First part of #170. | |||||
* | Import the "fork me on github" image into the tree (we don't want to | Keith Bostic | 2012-02-18 | 2 | -1/+1 | |
| | | | | | depend on a link to GitHub, and we really don't want to depend on a link through an akamai cache). | |||||
* | Add a clean target to the docs Makefile, add -a to dist/s_docs that | Keith Bostic | 2012-02-18 | 2 | -1/+11 | |
| | | | | cleans the build. | |||||
* | Replace the __wt_btree_reopen() function with __wt_conn_reopen_btree() | Keith Bostic | 2012-02-18 | 4 | -49/+16 | |
| | | | | | | | that does a full close/open pair instead of trying to fast-track it. This will be slower, but reopens aren't common, and this avoids having to worry about state held in the WT_BTREE structure across a verify or salvage operation. | |||||
* | Don't remove flexlint.out, that's not an expected file. | Keith Bostic | 2012-02-17 | 1 | -3/+2 | |
| | ||||||
* | Don't test for shared libraries unless Python is configured. | Keith Bostic | 2012-02-17 | 1 | -5/+5 | |
| | ||||||
* | Add a paragraph to the install page about building from the a GitHub | Keith Bostic | 2012-02-17 | 3 | -24/+53 | |
| | | | | | | | | | clone. Move port information into the main page (not great, but shouldn't be in the "install" page, either). Re-order the introduction a bit. | |||||
* | ../src/schema/schema_drop.c:28: error: ‘ret’ may be used uninitialized ↵ | Michael Cahill | 2012-02-17 | 1 | -0/+1 | |
| | | | | in this function | |||||
* | When parsing config strings, continue to the end of the string in case of ↵ | Michael Cahill | 2012-02-17 | 1 | -3/+4 | |
| | | | | | | repeated keys. closes #124 | |||||
* | Move "root:F" and "version:F" entries for files into the value for "file:F", ↵ | Michael Cahill | 2012-02-16 | 12 | -179/+164 | |
| | | | | so there is only a single record per file. | |||||
* | Add a link to the Google Group from the front page of the documentation. | Michael Cahill | 2012-02-17 | 1 | -2/+7 | |
| | ||||||
* | Note the wiredtiger_open config string "multiprocess" is not yet | Keith Bostic | 2012-02-16 | 2 | -2/+3 | |
| | | | | supported. | |||||
* | Move connection-list handling from the btree code to conn/conn_btree.c. | Keith Bostic | 2012-02-14 | 7 | -172/+247 | |
| | ||||||
* | Don't use "aspell --mode=ccpp" when spell-checking the documentation, | Keith Bostic | 2012-02-10 | 2 | -1/+2 | |
| | | | | | | there's no C-language code there (and it turns out that RedHat/CentOS don't support aspell in current releases). Makes #165 better, but it doesn't completely fix it. | |||||
* | Mention the Google group mailing list in the README (displayed on the ↵ | Michael Cahill | 2012-02-09 | 2 | -5/+11 | |
| | | | | project page). | |||||
* | Add a "fork me on github" ribbon to the documentation. | Michael Cahill | 2012-02-09 | 1 | -0/+1 | |
| | ||||||
* | Save and restore session->btree in schema ops to simplify calling code. | Michael Cahill | 2012-02-09 | 3 | -14/+32 | |
| | | | | closes #164 | |||||
* | Clarify the comments around WT_REF_EVICTING so we don't accidentally break ↵ | Michael Cahill | 2012-02-09 | 1 | -2/+8 | |
| | | | | | | it in future. refs #161 | |||||
* | Remove WT_PAGE_FORCE_EVICT: the WT_REF_EVICTING state makes it obsolete. | Michael Cahill | 2012-02-09 | 4 | -48/+24 | |
| | ||||||
* | Update documentation to cover cursor duplication. | Michael Cahill | 2012-02-09 | 1 | -2/+2 | |
| | | | | | | | refs #161 --HG-- extra : rebase_source : 19979afe98569f2348b730e93913b2b6362587a1 | |||||
* | Fixed indentation to not use tabs. | Don Anderson | 2012-02-08 | 1 | -2/+2 | |
| | ||||||
* | WT_NOTFOUND is only returned by a cursor function that can't find a | Keith Bostic | 2012-02-08 | 1 | -2/+3 | |
| | | | | | | | key/value pair to return, not by the cursor:open method. Catch the exception instead. Part of #163. | |||||
* | WT_NOTFOUND is only returned by a cursor function that can't find a | Keith Bostic | 2012-02-08 | 1 | -2/+2 | |
| | | | | | | | key/value pair to return, not by the cursor:open method. Update to check for ENOENT instead. Part of #163. | |||||
* | Change __session_drop to ignore force and do no specific mapping of the | Keith Bostic | 2012-02-08 | 1 | -15/+1 | |
| | | | | | | | | | | | return from __wt_schema_drop (we check force and do that mapping in __wt_schema_drop itself, as well as the functions that do the actual work for __wt_schema_drop). Minor change, remove a couple of places where we clear the local variable ret -- it's unnecessary, done by the SESSION_API_CALL macro. Part of #163. | |||||
* | Change __drop_table to not return WT_NOTFOUND if the table isn't found | Keith Bostic | 2012-02-08 | 1 | -5/+12 | |
| | | | | | | | | | | in the schema file and "force" is set. Change __wt_schema_drop to not only map WT_NOTFOUND to ENOENT, but map WT_NOTFOUND to 0 if "force" is set. The underlying drop functions should have handled the problem, but this should make sure. Part of #163. | |||||
* | Minor clarification of the meaning of WT_NOTFOUND: it's a cursor operation | Keith Bostic | 2012-02-08 | 1 | -4/+4 | |
| | | | | only, not a "search" operation. Part of #163. | |||||
* | Minor clarification of the meaning of WT_NOTFOUND: it's a cursor operation | Keith Bostic | 2012-02-08 | 1 | -1/+1 | |
| | | | | only, not a "search" operation. Part of #163. | |||||
* | Minor clarification of the meaning of WT_NOTFOUND: it's a cursor operation | Keith Bostic | 2012-02-08 | 1 | -5/+5 | |
| | | | | only, not a "search" operation. | |||||
* | If a session or connection method is about to return WT_NOTFOUND (some | Keith Bostic | 2012-02-08 | 3 | -39/+40 | |
| | | | | | | | underlying object was not found), map it to ENOENT, only cursor methods return WT_NOTFOUND. Fixes #163. | |||||
* | Fix with automake version < 1.11, use foreign mode so that fewer top-level ↵ | Michael Cahill | 2012-02-08 | 3 | -14/+3 | |
| | | | | files are required. | |||||
* | Don't fail tests if libtool munges the utility executable name (to 'lt-wt'). | Michael Cahill | 2012-02-08 | 1 | -2/+2 | |
| | ||||||
* | Switch to quiet builds by default. | Michael Cahill | 2012-02-08 | 1 | -0/+1 | |
| | ||||||
* | Cleanup test imports, use unittest2 if available so tests pass on Python 2.6. | Michael Cahill | 2012-02-08 | 40 | -203/+95 | |
| | ||||||
* | bzip2_compress.c:133: error: pointer targets in assignment differ in signedness | Michael Cahill | 2012-02-07 | 1 | -4/+4 | |
| | ||||||
* | SWIG 2.0.4 no longer mentions a LICENSE file, so there's reason we need to | Keith Bostic | 2012-02-06 | 1 | -22/+0 | |
| | | | | have one, as far as I can tell. | |||||
* | Implement cursor duplication via WT_SESSION::open_cursor. | Michael Cahill | 2012-02-06 | 3 | -3/+66 | |
| | | | | | | | closes #161 --HG-- extra : rebase_source : d81a47356b3fe2441707cf82d2c3f81c27e62810 | |||||
* | Remove a GCC-specific function attribution used for build testing. | Michael Cahill | 2012-02-06 | 2 | -2/+1 | |
| | | | | | --HG-- extra : rebase_source : 0ab3eaa41269e81c4e1f0d952306c87f97fc1f3b | |||||
* | Now that addresses are variable-size, it's theoretically possible to | Keith Bostic | 2012-02-03 | 1 | -3/+3 | |
| | | | | | | | | | | | configure a btree internal page size that's unable to store 2 maximum sized address cookies (a 255B/WT_BM_MAX_ADDR_COOKIE size address with a 512B internal page size). It's not a problem with the default block manager, but if/when we write a block manager extension API, we need to either pass it the internal page size, or have it set the maximum address cookie it can return so we can compare the two. | |||||
* | Now that addresses are variable-size, it's theoretically possible to | Keith Bostic | 2012-02-03 | 2 | -2/+22 | |
| | | | | | | | | | | | configure a btree internal page size that's unable to store 2 maximum sized address cookies (a 255B/WT_BM_MAX_ADDR_COOKIE size address with a 512B internal page size). It's not a problem with the default block manager, but if/when we write a block manager extension API, we need to either pass it the internal page size, or have it set the maximum address cookie it can return so we can compare the two. | |||||
* | Add checks to the session.truncate method to ensure the start/stop | Keith Bostic | 2012-02-03 | 1 | -0/+13 | |
| | | | | | | cursors reference the same object and have been initialized. closes #157 | |||||
* | WiredTiger release 1.0 | Michael Cahill | 2012-02-03 | 1 | -0/+1 | |
| | ||||||
* | Update to 1.0release-1.01.0.0 | Michael Cahill | 2012-02-03 | 5 | -12/+15 | |
| | ||||||
* | Fix a script status message. | Michael Cahill | 2012-02-03 | 1 | -1/+1 | |
| | ||||||
* | Add a navigation link to the license. | Michael Cahill | 2012-02-03 | 4 | -3/+4 | |
| | ||||||
* | Make s_docs quiet if Python isn't found. | Michael Cahill | 2012-02-03 | 1 | -4/+5 | |
| | ||||||
* | merge tip | Michael Cahill | 2012-02-03 | 465 | -0/+91147 | |
|\ |