- Aug 01, 2011
-
-
Lubos Slovak authored
refs #1087
-
Lubos Slovak authored
Conflicts: src/knot/server/xfr-in.c
-
Lubos Slovak authored
- Renamed to either deep_copy or shallow_copy. - Removed zonedb_copy() (not used).
-
-
Marek Vavrusa authored
-
Jan Kadlec authored
Conflicts: src/knot/conf/cf-lex.l src/zcompile/zcompile.c src/zcompile/zparser.y
-
Lubos Slovak authored
- Was not increasing reference count. refs #993
-
Lubos Slovak authored
-
Lubos Slovak authored
+ changed RRSet comparing function to compare also TTLs fixes #1030 @ 45m
-
- Jul 31, 2011
-
-
Marek Vavrusa authored
-
Marek Vavrusa authored
-
Marek Vavrusa authored
* Node and rrset now must use new API for changing owner (*_set_owner()) * Every structure that stores pointer to dname must retain reference to it * Destructors in rrset now release all stored dnames, obsoletes parameters: free_owner * Destructor in rdata_deep_free has obsolete parameter: free_all_dnames * Fixed some direct data manipulation * Some code readability improvements * To do: * Reference counting in zparser is disabled for now, implement correctly. * Remove obsolete parameters. * Destructors should have no extra parameters. * Tested with server startup, querying and AXFR client/server Commit refs #993.
-
- Jul 29, 2011
-
-
Jan Kadlec authored
-
Jan Kadlec authored
Refs #1068 @3h
-
Marek Vavrusa authored
-
- Jul 28, 2011
-
-
Marek Vavrusa authored
No leaks so far with server execution. * Newly created dnames have refcount=1, caller is responsible for releasing them. * Copied dnames have also refcount=1 (as they are new). To do: * Carefully implement more uses of dname not protected by refcounting. * Consider renaming dnslib_dname_copy() do _duplicate() as it creates duplicate. * Consider shallow copying of dname, as it's possible with refcounting. * Test and review. Commit refs #993.
-
Marek Vavrusa authored
GCC builtins are used for atomic add/sub and fetch. May be extended with assembly for other compilers later. Commit refs #993.
-
-
Lubos Slovak authored
refs #643 @10m
-
Lubos Slovak authored
- Fixed rehashing based on unittest output. TODO: there is still some condition on which it crashes on segfault; hard to replicate. - Small fixes in zone, zone tests. refs #643 @30m
-
Lubos Slovak authored
refs #641 @1h30m
-
Jan Kadlec authored
Added setting of previous node field. Removed assert that should not be there and was coushing abort with cz. zone (I removed that one before, too :D) Closes #1069
-
Lubos Slovak authored
-
Lubos Slovak authored
- They were causing the compiler to crash.
-
Lubos Slovak authored
- Zonedb always uses zone->name as the key. - Removed assert from dnslib_node-parent(), replaced by check. (Probably should do the same with other such asserts there.) - Reenabled adjust_rrsets() function (don't know why it was disabled). - Fixed wrong assert in zone-contents
-
Marek Vavrusa authored
* Fixed XFR client termination * Updated to new APIs Conflicts: src/dnslib/zone.c src/knot/server/dthreads.c src/knot/server/xfr-handler.c src/knot/server/xfr-handler.h
-
Lubos Slovak authored
-
- Jul 27, 2011
-
-
Lubos Slovak authored
- When the search in AVL tree went to the leftmost node, it did not return any 'previous' node. This was fixed in the AVL implementation by distinguishing this case. It may have impact on the performance as there is now one more condition jump for each encountered node during search. But it seems there is no better way around it. - Fixed the usages of the tree-finding functions. fixes #1065 @2h
-
Marek Vavrusa authored
-
Marek Vavrusa authored
Commit refs #1063.
-
Lubos Slovak authored
- Setting zone pointer to NSEC3 nodes inserted into the zone. - Fixed searching for NSEC3 node for some name (find_nsec3_for_name()) - there was a typo (name instead of nsec3_name, and not freeing the nsec3_name resulted in leak). - Small fixes in debug output.
-
- Jul 26, 2011
-
-
Lubos Slovak authored
refs #1056 @5m
-
Lubos Slovak authored
refs #1056 @5m
-
Lubos Slovak authored
refs #1056 @25m
-
Lubos Slovak authored
-
Lubos Slovak authored
- Some usages still remain, they would require more systematic changes. refs #1056 @2h
-
- Jul 25, 2011
-
-
Lubos Slovak authored
closes #1059 @10m
-
Lubos Slovak authored
refs #1058 @15m
-
Lubos Slovak authored
- Modified dname-table API a bit. - Just switching old node pointer to new node pointer. - Should not generate any problems. closes #1058 @30m
-
Lubos Slovak authored
-