summaryrefslogtreecommitdiffstats
path: root/lib/libc/db (follow)
Commit message (Collapse)AuthorAgeFilesLines
...
* Fix a mem leak and a potential NULL deref; from NetBSD's coverity analysis.otto2006-03-191-5/+9
| | | | ok millert@
* Fix a memory leak that could be trigged by a read errormillert2006-01-251-5/+11
| | | | | | | Update the count of currently cached pages in mpool_delete() Use CIRCLEQ_EMPTY in the loop that delets the lru cache Don't update the pageread statistic if there was a read error With aaron@, OK krw@
* Add missing break; in a switch case statement from r1.7 pread() changes. Fromaaron2006-01-241-1/+2
| | | | dhelder AT arbor.net. millert@ ok
* Use queue macros instead of directly accessing fields. ok millert@otto2005-10-121-8/+5
|
* Pass local var instead of mp->lqh.cqh_first to CIRCLEQ_REMOVE macro.otto2005-10-101-3/+3
| | | | | | The latter might fail, depending on the actual implementation of CIRCLEQ_REMOVE. Found by a strict queue.h I'm working on. ok pedro@ millert@
* zap remaining rcsid.espie2005-08-082-18/+2
| | | | | | Kill old files that are no longer compiled. okay theo
* zap rcsidespie2005-08-0528-247/+28
|
* - punctuation, macro, layout, wording improvementsjaredy2005-07-178-107/+120
| | | | | | | | | - avoid first person - document dbm_rdonly - update some error return values in the mpool(3) routines - sync header file excerpts ok jmc
* ansify. ok deraadt@ moritz@otto2005-03-2328-548/+232
|
* deregistermillert2005-01-036-75/+67
|
* Avoid namespace pollution by renaming index variables -> idx.millert2005-01-037-105/+105
| | | | Idea from NetBSD; OK deraadt@
* Make a u_int32_t-sized copy of data->size and key->size (which are size_t)millert2005-01-031-5/+7
| | | | | so when we memmove() it we get the correct 32 bits copied on sparc64. Adapted from FreeBSD. OK otto@
* Avoid overwriting the cursor page when the cursor page becomes theotto2004-10-011-3/+11
| | | | | LRU page. From pacman at manson dot clss dot net. ok deraadt@ millert@
* ansi and ARSUSEDderaadt2004-09-151-4/+4
|
* Working hcreate(3) et al from NetBSD (cgd) via ray at cyth dot net.millert2004-06-244-309/+4
| | | | Now passes the regress tests.
* another errno -> err in a structure. OK deraadt@marc2004-06-213-15/+15
|
* O_NOFOLLOW and O_SYNC are also valid flags to dbopen(3)millert2004-05-051-1/+3
|
* Make ndbm match POSIX. A datum consists of a void *dptr and a size_t dsize.millert2004-05-033-12/+15
| | | | Also make the open functions take a mode_t, not int. OK deraadt@
* datbase -> database; from Jared Yanovichjmc2003-10-211-2/+2
|
* Remove some double semicolons (hmm, do two semis equal a maxi?).millert2003-08-061-3/+3
| | | | I've skipped the GNU stuff for now. From Patrick Latifi.
* convert some complicated lists to displays;jmc2003-07-073-69/+31
|
* pull in headersderaadt2003-06-253-9/+14
|
* Sync with share/misc/license.template and add missing DARPA creditmillert2003-06-171-8/+8
| | | | where applicable.
* Use an ISC-tyle license for all my code; it is simpler and more permissive.millert2003-06-031-22/+11
|
* Remove the advertising clause in the UCB license which Berkeleymillert2003-06-0243-288/+116
| | | | rescinded 22 July 1999. Proofed by myself and Theo.
* section reorder;jmc2003-05-301-3/+3
|
* typos;jmc2003-05-072-6/+10
|
* -Wall cleanup, make rcsid variables constavsm2003-05-0131-61/+61
| | | | millert@ ok
* snprintfderaadt2002-06-051-4/+3
|
* use int32_t, not int, for on-disk format structure.itojun2002-05-311-18/+18
|
* o) start new sentence on a new line;mpech2002-05-012-7/+12
| | | | | | | o) always close .Bl tags; o) fix usage of .Xr; millert@ ok
* Fix temporary buffer memory leak with big hash keys and data;millert2002-03-251-2/+6
| | | | Christopher L. Cousins
* Cosmetic changes such as correct printf format specifiers on LP64millert2002-02-2510-75/+75
| | | | machines and whitespace nits. Some from NetBSD.
* We have search.h in /usr/include so we don't need this private one.millert2002-02-242-58/+1
|
* Manual cleanup of remaining userland __P use (excluding packages maintained outside the tree)millert2002-02-172-6/+6
|
* Part one of userland __P removal. Done with a simple regexp with some minor hand editing to make comments line up correctly. Another pass is forthcoming that handles the cases that could not be done automatically.millert2002-02-1622-172/+172
|
* Allow O_SYNC and O_NOFOLLOW flags in dbopen()millert2002-02-091-4/+4
|
* Only squeeze a short key/value pair onto a page with other complete key/valuemillert2002-02-011-5/+11
| | | | pairs, not onto a page containing the end of a big pair; mycroft@netbsd.org
* Return a useful errno value in the overflow case. Also, print amillert2002-02-011-3/+9
| | | | message to stderr in one more case. mycroft@netbsd.org
* If (keysize+datasize)%(bsize=14)==0, insertion of a `big key' would cause anmillert2002-02-011-7/+19
| | | | | | | | | | | | invariant (actually, an ugly hack) to fail, and all Hell would break loose. When deleting a big key, the offset of an empty page should be bsize, not bsize-1; otherwise an insertion into the empty page will cause the new key to be elongated by 1 byte. Make the packing more dense in a couple of cases. From mycroft@netbsd.org
* After writing a page out to disk, run the page back through the user'smillert2002-02-011-2/+11
| | | | | | | | | | | input filter to ensure we don't end up with a cached copy of the page in the wrong byte-order for the host cpu. This fixes a fatal bug which bites when the DB 'lorder' is different to the cpu's, and a cached page is accessed soon after it was flushed to disk. From scw@netbsd.org
* Fix a condition where the hole would be inserted in the wrong place during amillert2002-02-011-4/+4
| | | | split; mycroft@netbsd.org
* Fill in missing (default) mmap(2) flag MAP_FILE.millert2002-01-311-3/+3
| | | | mycroft@netbsd.org
* Use pread(2) and pwrite(2) instead of lseek(2) + read(2) / write(2).millert2002-01-313-24/+20
| | | | Based on changes from NetBSD (thorpej).
* return 0, not NULL, the function returns an intart2001-09-161-3/+3
|
* o) We always close .Bl and .Bd tags;mpech2001-08-063-6/+4
| | | | | | | | | | o) .Sh AUTHOR -> .Sh AUTHORS; o) We don't like .Pp before/after .Sh; o) We don't like .Pp before/after .Rs/.Re; o) NetBSD -> .Nx; o) OpenBSD -> .Ox; millert@ ok
* Move decision whether to make a new hash table or not to be aftermillert2001-08-041-13/+8
| | | | | | | | the open(). That way if the user opened with O_EXLOCK and another process has the .db locked we don't decide to zero hash before the locking process has had a chance to write to the database. This also makes the logic simpler since we can just fstat the fd and check for size == 0 && !O_RDONLY. Closes PR 1934.
* more .%A handling; bk@rt.fmderaadt2001-06-231-2/+3
|
* ok, tmac is now fixedderaadt2001-06-231-2/+3
|
* join .%A entries; most by bk@rt.fmderaadt2001-06-231-3/+2
|
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.