| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | Add another comment, about "deleted" pages that came about through eviction. | Keith Bostic | 2014-04-10 | 1 | -0/+7 |
| * | Add some more comments on skipping deleted pages during reads. | Keith Bostic | 2014-04-10 | 1 | -4/+16 |
| * | Gather most of the fast-delete code into a new file, bt_delete.c. | Keith Bostic | 2014-04-10 | 5 | -285/+331 |
| * | Merge pull request #948 from wiredtiger/new-split-fast-delete | Michael Cahill | 2014-04-10 | 14 | -154/+282 |
| |\ |
|
| | * | Minor cleanup to simplify the tree-walk code, simplify function naming, | Keith Bostic | 2014-04-09 | 3 | -16/+13 |
| | * | Fix the problem where a truncated/deleted page might split (meaning | Keith Bostic | 2014-04-09 | 11 | -109/+192 |
| | * | typo, introduce an infinite loop into the test suite. | Keith Bostic | 2014-04-08 | 1 | -1/+1 |
| | * | WT_REF_SPLIT is possible, I believe, and it's a problem. | Keith Bostic | 2014-04-08 | 3 | -7/+12 |
| | * | Make lint happier, list the WT_REF states we're ignoring. | Keith Bostic | 2014-04-08 | 1 | -2/+5 |
| | * | Clean up the fast-delete code (part of doing a review after merging the | Keith Bostic | 2014-04-08 | 3 | -56/+96 |
| * | | Don't try and extend the file unless we have a valid offset. | Keith Bostic | 2014-04-09 | 1 | -1/+1 |
| * | | Missing WT_RET(), don't lose the error. | Keith Bostic | 2014-04-09 | 1 | -0/+1 |
| * | | The posix_fallocate call on Linux is corrupting already written data, | Keith Bostic | 2014-04-09 | 2 | -21/+66 |
| * | | Check that we don't try to discard pages that are splitting. | Michael Cahill | 2014-04-09 | 1 | -1/+2 |
| * | | Reduce the leaf page size from 1MB to 32KB. | Keith Bostic | 2014-04-08 | 2 | -7/+7 |
| * | | whitespace | Keith Bostic | 2014-04-08 | 1 | -2/+1 |
| |/ |
|
| * | Minor cleanups: | Keith Bostic | 2014-04-07 | 3 | -19/+15 |
| * | Add another case, it's possible to delete fake checkpoints. | Keith Bostic | 2014-04-06 | 1 | -0/+1 |
| * | Add some diagnostic code to confirm the checkpoint array looks like we | Keith Bostic | 2014-04-06 | 1 | -6/+39 |
| * | Set the checkpoint size before calling the checkpoint update function, | Keith Bostic | 2014-04-06 | 1 | -13/+9 |
| * | The live systems allocation/discard extent lists are potentially quite large, | Keith Bostic | 2014-04-06 | 2 | -3/+19 |
| * | Don't terminate a verbose message with a newline, it's not needed. | Keith Bostic | 2014-04-06 | 1 | -1/+1 |
| * | rec_split.c:610:88: error: incompatible pointer types passing 'uint64_t | Keith Bostic | 2014-04-06 | 1 | -2/+1 |
| * | Add some additional wording to a comment, explaining why we can fast-path | Keith Bostic | 2014-04-04 | 1 | -1/+7 |
| * | Merge branch 'develop' into new-split | Keith Bostic | 2014-04-04 | 2 | -9/+9 |
| |\ |
|
| | * | Merge pull request #942 from wiredtiger/man-pages | Michael Cahill | 2014-04-04 | 1 | -1/+1 |
| | |\ |
|
| | | * | Generate Linux man pages when generating docs. | Alex Gorrod | 2014-04-03 | 1 | -1/+1 |
| | * | | If not tracking the object sizes, use the new extent append API instead | Keith Bostic | 2014-04-03 | 1 | -8/+8 |
| | |/ |
|
| * | | The __wt_evict_file() and __wt_sync_file() functions are no longer part | Keith Bostic | 2014-04-04 | 3 | -243/+241 |
| * | | WT_PAGE_TRYLOCK is no longer used, remove it. | Keith Bostic | 2014-04-04 | 1 | -2/+0 |
| * | | Fix lint's version of F_CAS_ATOMIC. | Keith Bostic | 2014-04-04 | 1 | -1/+3 |
| * | | If __wt_epoch call failed, we could return without clearing the | Keith Bostic | 2014-04-04 | 1 | -11/+11 |
| * | | Clean up a comment, no code change. | Keith Bostic | 2014-04-04 | 1 | -10/+12 |
| * | | Fix a comment, we no longer acquire the page's spinlock. | Keith Bostic | 2014-04-04 | 1 | -5/+5 |
| * | | Fix a race between reconciliation and freeing obsolete updates on a page. No... | Michael Cahill | 2014-04-04 | 3 | -6/+23 |
| * | | Use the standard transaction visibility check for cleaning up cached overflow... | Michael Cahill | 2014-04-04 | 1 | -10/+2 |
| * | | Revert to asserting that a page is clean after `__wt_rec_write` returns zero,... | Michael Cahill | 2014-04-04 | 1 | -8/+3 |
| * | | Simplify pinning and releasing a transaction ID required to look at page inde... | Michael Cahill | 2014-04-04 | 9 | -81/+64 |
| * | | Merge branch 'new-split' into new-split-fixes | Michael Cahill | 2014-04-04 | 1 | -6/+1 |
| |\ \ |
|
| | * | | We have to hold the overflow lock regardless of whether or not the on-page | Keith Bostic | 2014-04-03 | 1 | -6/+1 |
| * | | | Merge branch 'new-split' into new-split-fixes | Michael Cahill | 2014-04-04 | 1 | -18/+30 |
| |\ \ \
| |/ / |
|
| | * | | Don't write removed overflow keys that are associated with removed, | Keith Bostic | 2014-04-03 | 1 | -18/+30 |
| * | | | Merge branch 'new-split' into new-split-fixes | Keith Bostic | 2014-04-03 | 3 | -15/+45 |
| |\ \ \
| |/ / |
|
| | * | | Rewrite the big comment about how we're handling removed on-page | Keith Bostic | 2014-04-03 | 1 | -11/+31 |
| | * | | Fix up a comment in the walk code: we can't see restart on an ascent | Keith Bostic | 2014-04-03 | 1 | -4/+9 |
| | * | | Clarify a comment. | Keith Bostic | 2014-04-03 | 1 | -0/+5 |
| * | | | Catch a case where snap_min is being cleared while cursors are open. This tu... | Michael Cahill | 2014-04-03 | 5 | -4/+12 |
| * | | | Make sure we always publish a required transaction ID before looking at page ... | Michael Cahill | 2014-04-03 | 7 | -21/+82 |
| * | | | Flip the logic around when to deepen the tree -- there is no point trying to ... | Michael Cahill | 2014-04-03 | 1 | -14/+10 |
| * | | | It's okay if an LRU walk returns WT_NOTFOUND: that just means some page was l... | Michael Cahill | 2014-04-03 | 1 | -1/+4 |
