Hallvard Furuseth [Wed, 18 Jun 2014 21:01:14 +0000 (23:01 +0200)]
Fix MDB_GET_MULTIPLE/MDB_NEXT_MULTIPLE doc.
Howard Chu [Wed, 18 Jun 2014 20:37:04 +0000 (13:37 -0700)]
Tweak ovpage search
Use num*20, chosen from empirical testing
Howard Chu [Wed, 18 Jun 2014 19:29:13 +0000 (12:29 -0700)]
Fix some overwrite quirks
Was skipping the overwrite in some cases, if the cmp function
said the new and old already matched. We should always perform
the overwrite anyway, since the cmp function may only be doing
a prefix compare and the data may actually be different.
Howard Chu [Wed, 18 Jun 2014 16:07:39 +0000 (09:07 -0700)]
Tweak prev patch for ovpage search
Howard Chu [Wed, 18 Jun 2014 10:02:08 +0000 (03:02 -0700)]
Shorten search for ovpage space
Howard Chu [Wed, 18 Jun 2014 09:53:17 +0000 (02:53 -0700)]
Howard Chu [Fri, 13 Jun 2014 17:46:27 +0000 (10:46 -0700)]
Bump version to 0.9.12
Howard Chu [Wed, 11 Jun 2014 09:27:30 +0000 (02:27 -0700)]
ITS#7875 fix #7681 regression
Due to commit
18a07eb7c2dc33372455a6040984cd6b699b41a5
Alexey Naumov [Sun, 25 May 2014 11:10:21 +0000 (17:10 +0600)]
Fix possible crash in DLL_THREAD_DETACH (explodes in .Net Framework)
Hallvard Furuseth [Fri, 30 May 2014 08:50:04 +0000 (10:50 +0200)]
Comment ~MDB_APPEND, from earlier commit message.
Hallvard Furuseth [Fri, 30 May 2014 08:47:56 +0000 (10:47 +0200)]
Factor out mdb_cmp_clong()
Hallvard Furuseth [Fri, 30 May 2014 08:47:56 +0000 (10:47 +0200)]
More for big MDB_MAXKEYSIZE with MDB_DUPSORT.
With mdb_cursor_put(new key, data item big enough to need a sub-DB):
- Track cursors as when adding other nodes. The "!do_sub" is now
unneeded. Before ITS#7861 it distinguished inserting key from data.
- Set MDB_TXN_ERROR if an empty sub-DB was created. I.e. adding
key:subDB succeeded, but inserting the data item failed.
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
Factor out MDB_TXN_ERROR
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
Simplify MDB_MULTIPLE
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
ITS#7834 Fix MDB_MULTIPLE writing multiple keys.
Update the state before looping, so the key is not duplicated.
Broke in
4c8f57615c5ca7b014c038e59c1045182e74f5ad (mostly).
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
ITS#7861 Re-fix MDB_db.md_entries.
mdb_cursor_del(): Fix md_entries with MDB_NODUPDATA.
mdb_cursor_put(): Check in xcursor whether an item got inserted. Count
data instead of keys after
13d6d59c512065db3741c20c65b4d9928b986968.
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
ITS#7771 mdb_cursor_put doc: Cursor moves on error
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
Document DBI issues, extend MDB_BAD_VALSIZE.
Hallvard Furuseth [Fri, 30 May 2014 07:06:08 +0000 (09:06 +0200)]
More ITS#7793 Check key->mv_size with MDB_CURRENT.
Since non-NULL key must be provided now.
Howard Chu [Tue, 20 May 2014 15:47:28 +0000 (08:47 -0700)]
Fix mdb_rebalance collapsing root
Shift the rest of the cursor stack as needed
Howard Chu [Thu, 3 Apr 2014 19:51:41 +0000 (12:51 -0700)]
ITS#7829 more for mdb_rebalance
Make sure to propagate rebalance results all the
way back to original caller's cursor.
Hallvard Furuseth [Wed, 2 Apr 2014 05:02:14 +0000 (07:02 +0200)]
ITS#7829 Simplify last commit
Howard Chu [Mon, 31 Mar 2014 20:42:19 +0000 (13:42 -0700)]
ITS#7829 more for prev commit
Howard Chu [Fri, 28 Mar 2014 20:51:59 +0000 (13:51 -0700)]
ITS#7829 fix rebalance
Fixup cursor state after node_move/page_merge
Howard Chu [Wed, 26 Mar 2014 15:59:43 +0000 (08:59 -0700)]
ITS#7793/7828
Fix md_entries count, test for insert on md_entries increment
Howard Chu [Wed, 26 Mar 2014 09:41:26 +0000 (02:41 -0700)]
More for ITS#7793/7828
Note MDB_CURRENT/MDB_DUPSORT requirement
Howard Chu [Wed, 26 Mar 2014 09:33:51 +0000 (02:33 -0700)]
ITS#7828, #7793
Clarify MDB_CURRENT key usage, simplify freelist_save
Howard Chu [Wed, 26 Mar 2014 09:18:20 +0000 (02:18 -0700)]
Revert "ITS#7793 mdb_cursor_put(): Fix MDB_CURRENT."
This reverts commit
4e954f8ae83d5b39736bab2da0eaf2bdf5ee3961.
Caused ITS#7828.
Hallvard Furuseth [Sun, 23 Mar 2014 12:06:03 +0000 (13:06 +0100)]
Revert "Update MDB_PANIC description" (thinko).
This reverts commit
6bd34d0534f1c8bfd73deaf70febd3b58a2f9d5a.
Hallvard Furuseth [Sat, 22 Mar 2014 21:27:10 +0000 (22:27 +0100)]
Update MDB_PANIC description
Hallvard Furuseth [Sat, 22 Mar 2014 21:27:10 +0000 (22:27 +0100)]
Tweak mdb_mid2l_insert() call from ITS#7775.
It does not return MDB error codes.
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:43 +0000 (12:42 +0100)]
mtest: Accept an MDB_NOTFOUND case
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:43 +0000 (12:42 +0100)]
Handle more errors. Invalidate txn if needed.
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:42 +0000 (12:42 +0100)]
Don't use non-existent node info in branch/leaf2
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:42 +0000 (12:42 +0100)]
Move code out of mdb_cursor_del0() and mdb_del().
Prepares for the next commits.
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:42 +0000 (12:42 +0100)]
More checks for closed DBIs and invalidated txns.
Factor txn/DBI-checks out to TXN_DBI_EXIST().
mdb_audit(): Skip closed DBIs.
mdb_cursor_renew(), mdb_stat(): Check DBI and txn.
mdb_cursor_count(): Check txn.
mdb_dbi_flags(): Check DBI.
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:42 +0000 (12:42 +0100)]
mdb_page_list(): Support & display all page types
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:42 +0000 (12:42 +0100)]
mdb_txn_renew0(): Fix un-mutexed me_flags update.
Replace env flag MDB_LIVE_READER with field me_live_reader.
Hallvard Furuseth [Sun, 16 Mar 2014 11:42:42 +0000 (12:42 +0100)]
Doxygen fixes
Howard Chu [Sun, 16 Mar 2014 10:55:03 +0000 (03:55 -0700)]
Rename samples
They're meant to be read and studied, not run.
Howard Chu [Thu, 6 Mar 2014 04:40:12 +0000 (20:40 -0800)]
ITS#7815 more for page_split
This case didn't come up in testing. Needs finer grained test.
Howard Chu [Thu, 6 Mar 2014 04:04:42 +0000 (20:04 -0800)]
ITS#7815 fix mdb_page_split
Was leaving cursor pointed at wrong page if the newindx was
in the right sibling page, and the parent was also split.
Hallvard Furuseth [Wed, 19 Feb 2014 20:25:40 +0000 (21:25 +0100)]
mtest*.c, sample-*.c: Add/note error checks.
Hallvard Furuseth [Wed, 19 Feb 2014 20:25:40 +0000 (21:25 +0100)]
mtest[46]: Use different DBs when different flags.
Hallvard Furuseth [Tue, 28 Jan 2014 13:27:27 +0000 (14:27 +0100)]
ITS#7793 mdb_cursor_put(): Fix MDB_CURRENT.
Ignore key, broken by
5bda3565a9bfaa6cd54053faeafcc06da15bc00c
and some older code. Document and clarify MDB_CURRENT usage.
Also affects non-MDB_CURRENT put() with empty data and a key
which matches by the mdb_set_compare function but not by memcmp.
Howard Chu [Wed, 15 Jan 2014 17:09:48 +0000 (09:09 -0800)]
Bump version to 0.9.11
Howard Chu [Wed, 15 Jan 2014 01:12:48 +0000 (17:12 -0800)]
Note changes for LMDB 0.9.11
Hallvard Furuseth [Mon, 13 Jan 2014 19:46:08 +0000 (20:46 +0100)]
Revert "ITS#7772 Reduce spurious sub-page growth."
This reverts commit
75494c08b49c3ef77dce8a00f2bc44c83658c1a7,
which must await a fix to ITS#7771 (DUPSORT cursor tracking).
Howard Chu [Mon, 13 Jan 2014 01:37:04 +0000 (17:37 -0800)]
Make sure mapsize changes take effect on Windows.
Hallvard Furuseth [Mon, 6 Jan 2014 22:21:54 +0000 (23:21 +0100)]
ITS#7775 Drop <assert.h>, drop/replace assert()s.
db_mid2l_insert(): Move assert to mdb.c.
mdb_cursor_set(): Previous assert cannot fail now.
mdb_cursor_put(): Check mc/key and return EINVAL.
mdb_cursor_dbi(): No error return, so just segfault if cursor==NULL.
Howard Chu [Mon, 6 Jan 2014 22:21:44 +0000 (23:21 +0100)]
ITS#7775 Cleanup mdb_node_del()
Hallvard Furuseth [Mon, 6 Jan 2014 22:17:38 +0000 (23:17 +0100)]
ITS#7775 Add mdb_env_set_assert()
Hallvard Furuseth [Mon, 6 Jan 2014 22:17:38 +0000 (23:17 +0100)]
Add mdb_env_<set,get>_userctx()
Hallvard Furuseth [Mon, 6 Jan 2014 22:17:37 +0000 (23:17 +0100)]
ITS#7377 Invalidate txn on page-allocation errors
This should likely be reverted when all callers handle these errors.
Hallvard Furuseth [Sun, 5 Jan 2014 15:37:26 +0000 (16:37 +0100)]
Wrap __func__ in mdb_func_
Hallvard Furuseth [Sun, 5 Jan 2014 15:26:21 +0000 (16:26 +0100)]
Invalidate the txn on corruption/internal errors
Howard Chu [Fri, 3 Jan 2014 20:40:28 +0000 (12:40 -0800)]
ITS#7775 cleanup unnecessary assert()s
Hallvard Furuseth [Tue, 31 Dec 2013 01:16:19 +0000 (02:16 +0100)]
mdb_page_malloc(): Move VGMEMP_ALLOC before memset
Hallvard Furuseth [Sun, 29 Dec 2013 14:58:21 +0000 (15:58 +0100)]
ITS#7771 mdb_cursor_del0(): Fix xcursor tracking.
Hallvard Furuseth [Sun, 29 Dec 2013 14:57:33 +0000 (15:57 +0100)]
sprintf format: pid_t need not be an int.
Hallvard Furuseth [Sun, 29 Dec 2013 14:09:51 +0000 (15:09 +0100)]
Add mdb_dbg_pgno(): (sub)pagenum for debug output.
Hallvard Furuseth [Sun, 29 Dec 2013 14:09:05 +0000 (15:09 +0100)]
ITS#7772 Reduce spurious sub-page growth.
mdb_cursor_put(): Use free sub-page space added by a
previous put(existing DUPSORT item), if any.
Tweak NUMKEYS code to avoid a new signedness warning.
Hallvard Furuseth [Sun, 29 Dec 2013 13:54:17 +0000 (14:54 +0100)]
ITS#7665 Drop lmdb.h doc on non-overlapping flags.
It was inaccurate and misplaced. Verify it in mdb.c instead.
Hallvard Furuseth [Sun, 29 Dec 2013 13:52:30 +0000 (14:52 +0100)]
ITS#7515 has been fixed.
Hallvard Furuseth [Sun, 29 Dec 2013 13:52:30 +0000 (14:52 +0100)]
Drop UNLOCK_MUTEX_R() on mdb_reader_pid() error
Hallvard Furuseth [Sat, 21 Dec 2013 11:31:57 +0000 (12:31 +0100)]
Fix commit "Raise safe max MDB_MAXKEYSIZE."
I.e.
d69d2ce2307e9d0e2167843935c386b05a869c31, it left out some changes.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:13 +0000 (11:57 +0100)]
Replace MDB_node.mn_offset[], to ease debugging.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:13 +0000 (11:57 +0100)]
Comment MDB_node, mdb_page_unspill, mdb_update_key
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:13 +0000 (11:57 +0100)]
Fix mdb_reader_list() and its spec.
It and the MDB_msg_func can now return >= 0 for success.
Always return any MDB_msg_func() error result.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:13 +0000 (11:57 +0100)]
Support dynamic max MDB keysize.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:13 +0000 (11:57 +0100)]
Raise safe max MDB_MAXKEYSIZE.
Use a sub-DB for DUPSORT item #1/#2 per key if needed: Not a sub-
page too big for a node, nor an overflow page (which not all DUPSORT
code checks for). Move "insert" code, to avoid non-loop goto upwards.
(This is the commit which needs the change to xdata.mv_size in
commit
9d6e4a916367e85ffdf37b1cb7b5cdb7ac0e15b5 "page sizes".)
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
Relax MDB_MAXKEYSIZE for existing data.
Handle keys stored by a liblmdb with a bigger MDB_MAXKEYSIZE.
mdb_get/mdb_del(absent key bigger than our MDB_MAXKEYSIZE)
now return MDB_NOTFOUND instead of MDB_BAD_VALSIZE.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
ITS#7589 Fix me_nodemax and related node sizes.
Change me_nodemax to not count the mp_ptrs[] entry. That's mostly
how it was used. Compare node sizes ">" me_nodemax instead of ">=".
The ">=" was a workaround for confusing sizes with and without the
mp_ptrs[] entry, but broke for nodes with size (old me_nodemax-1).
Explicitly make me_nodemax even. An odd value could break the
comparisons. It was even anyway because MDB_MINKEYS == 2.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
ITS#7589 Fix delete(uneven-sized MDB_DUPFIXED data).
Do not give uneven sizes to the sub-page and the node it lives in.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
put(MDB_MULTIPLE): Skip existing item cleanly.
Don't set dkey.mv_size if mdb won't clear it before next iteration.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
Add EVEN(): Micro-optimize "up to multiple of 2".
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
Warning cleanup.
Drop unused computed/assigned values.
mdb_page_alloc(): Make "mop_len>0 implies mop!=NULL" provable.
Hallvard Furuseth [Wed, 11 Dec 2013 10:57:12 +0000 (11:57 +0100)]
Clean up mdb_copy.
Support MDB_NOSUBDIR. Catch more errors. Write messages
to stderr, not stdout which the data too may be piped to.
Howard Chu [Thu, 5 Dec 2013 02:42:49 +0000 (18:42 -0800)]
Fix GET_BOTH_RANGE
On near match, return the matched data.
Howard Chu [Sat, 30 Nov 2013 01:03:49 +0000 (17:03 -0800)]
ITS#7756 keysizes were wrong in mdb_update_key
Howard Chu [Tue, 12 Nov 2013 16:10:33 +0000 (08:10 -0800)]
Doc tweak
Measured perf cost of meminit was 7% worst case on an Intel Core2
system. YMMV.
Howard Chu [Tue, 12 Nov 2013 04:15:54 +0000 (20:15 -0800)]
Bump to 0.9.10
Howard Chu [Tue, 12 Nov 2013 04:08:26 +0000 (20:08 -0800)]
Windows/Win64/MSVC tweaks
Silence warnings, fix broken MinGW64 definitions, etc.
Howard Chu [Tue, 12 Nov 2013 00:35:17 +0000 (16:35 -0800)]
Clarify WRITEMAP / NOMEMINIT
Howard Chu [Mon, 11 Nov 2013 23:09:32 +0000 (15:09 -0800)]
Change default mem init behavior
Init malloc'd memory by default, use #MDB_NOMEMINIT to disable.
Bump version to 0.9.10
Hallvard Furuseth [Thu, 7 Nov 2013 22:41:30 +0000 (23:41 +0100)]
Add MDB_CLEANMEM option.
Initialize unused portions of DB pages, for security or to silence
checkers like Purify and Valgrind. Like BDB configure --enable-umrw.
Hallvard Furuseth [Thu, 7 Nov 2013 22:38:15 +0000 (23:38 +0100)]
Fix MDB_MAXKEYSIZE doc: Account for MDB_DUPSORT.
A node may need room for [key, subpage with 2 items] = 3 keys.
Hallvard Furuseth [Thu, 7 Nov 2013 22:38:15 +0000 (23:38 +0100)]
ITS#7713 Handle bigger OS page sizes
Use DB page size = min(32k, OS pagesize). Previous limit was 8k
(MDB_MINKEYS*MDB_PAGESIZE).
Handle DB pagesize < OS pagesize. That's an I/O pessimization,
but transactions remain atomic: Only writing the MDB_meta must
be atomic, and it fits in one OS page.
Don't truncate desired subpage size: Asssign it to a size_t
(mv_size), not an uint16_t (mp_upper).
Hallvard Furuseth [Thu, 7 Nov 2013 22:38:15 +0000 (23:38 +0100)]
mdb_cursor_put(): Factor out code. Kill a label.
Howard Chu [Thu, 7 Nov 2013 17:37:48 +0000 (09:37 -0800)]
Fix prev commit, cursor depth comparison
Howard Chu [Thu, 7 Nov 2013 00:49:25 +0000 (16:49 -0800)]
Fix cursor_del0 cursor fixup
Skip cursors that aren't as deep as current cursor
Howard Chu [Tue, 5 Nov 2013 04:00:55 +0000 (20:00 -0800)]
More for #7589
Howard Chu [Fri, 1 Nov 2013 00:31:15 +0000 (17:31 -0700)]
ITS#7589 another mdb_page_split fix
Tweak split_indx check, go one slot further.
Howard Chu [Thu, 31 Oct 2013 23:01:04 +0000 (16:01 -0700)]
ITS#7736 fix regression in ITS#7733 patch
In
d8eccb353b1be7f07e74149dde9a41bb199d74e8
Howard Chu [Mon, 28 Oct 2013 14:25:54 +0000 (07:25 -0700)]
ITS#7733 check cursor index before cursor_del
David Wilson [Mon, 28 Oct 2013 13:58:04 +0000 (06:58 -0700)]
ITS#7734 Add MDB_NORDAHEAD to CHANGELESS.
Howard Chu [Thu, 24 Oct 2013 22:57:17 +0000 (15:57 -0700)]
Bump to 0.9.9
Howard Chu [Mon, 21 Oct 2013 17:29:38 +0000 (10:29 -0700)]
Add mdb_env_get_fd()
Return the environment's filedescriptor. Useful when the caller
is doing their own locking.
Howard Chu [Mon, 14 Oct 2013 16:47:26 +0000 (09:47 -0700)]
Howard Chu [Sat, 12 Oct 2013 23:20:04 +0000 (16:20 -0700)]
ITS#7725 add MDB_NORDAHEAD flag for env_open