]>
git.saurik.com Git - redis.git/log
Pieter Noordhuis [Thu, 2 Sep 2010 17:18:55 +0000 (19:18 +0200)]
Changed reply buildup internals
Pieter Noordhuis [Thu, 2 Sep 2010 12:31:25 +0000 (14:31 +0200)]
Removed unneeded function
Pieter Noordhuis [Thu, 2 Sep 2010 12:30:56 +0000 (14:30 +0200)]
Use existing reply functions where possible
Pieter Noordhuis [Thu, 2 Sep 2010 12:19:15 +0000 (14:19 +0200)]
Free the sds in addReplySds when it cannot be added to the reply
Pieter Noordhuis [Thu, 2 Sep 2010 12:17:53 +0000 (14:17 +0200)]
Intialize bufpos in the fake client
Pieter Noordhuis [Thu, 2 Sep 2010 10:38:34 +0000 (12:38 +0200)]
Use specialized function to add multi bulk reply length
Pieter Noordhuis [Mon, 30 Aug 2010 14:51:39 +0000 (16:51 +0200)]
Don't build a reply when replaying the AOF
Pieter Noordhuis [Mon, 30 Aug 2010 14:02:06 +0000 (16:02 +0200)]
Wrapper for adding unknown multi bulk length to reply list
Pieter Noordhuis [Mon, 30 Aug 2010 12:44:34 +0000 (14:44 +0200)]
Refactor reply buildup for speed on large multi bulk replies
Pieter Noordhuis [Mon, 30 Aug 2010 09:25:02 +0000 (11:25 +0200)]
Show the current throughput while benchmarking
Pieter Noordhuis [Mon, 30 Aug 2010 09:14:54 +0000 (11:14 +0200)]
Refactor reply parsing code in redis-benchmark for efficiency
antirez [Mon, 30 Aug 2010 08:32:32 +0000 (10:32 +0200)]
Fix for a race in BGSAVE that may result in some data not being saved as soon as possible (when the configured saving triggers should fire). Also known as Issue 313, more details there in the google code issue.
antirez [Fri, 27 Aug 2010 15:06:36 +0000 (17:06 +0200)]
fix for the prev fix
antirez [Fri, 27 Aug 2010 15:04:26 +0000 (17:04 +0200)]
Fix for bug 312, yet to verify in a couple of minutes...
antirez [Fri, 27 Aug 2010 10:46:10 +0000 (12:46 +0200)]
Fixed another instace of the Issue 173
antirez [Fri, 27 Aug 2010 09:01:03 +0000 (11:01 +0200)]
Fixed segfault in freeMemoryIfNeeded due to the fact that keys are now sds strings and not objects in the main hash table, thanks to Anthony Lauzon for spotting the bug and providing a patch.
antirez [Thu, 26 Aug 2010 17:10:40 +0000 (19:10 +0200)]
intset loading bug fixed
antirez [Thu, 26 Aug 2010 16:47:03 +0000 (18:47 +0200)]
test for intset integer encodability test and some small refactoring
antirez [Thu, 26 Aug 2010 16:11:26 +0000 (18:11 +0200)]
translated a few long logn into int64_t for correctness and to avoid compilation warnings as well
antirez [Thu, 26 Aug 2010 15:52:02 +0000 (17:52 +0200)]
redis-cli tests commented out since there is a problem with the CI server and this tests, will be activated again once the problem is fixed
antirez [Thu, 26 Aug 2010 15:29:13 +0000 (17:29 +0200)]
SORT stress testing against bigger aggregate values
antirez [Thu, 26 Aug 2010 14:58:02 +0000 (16:58 +0200)]
Version is now 2.1.3
antirez [Thu, 26 Aug 2010 14:19:06 +0000 (16:19 +0200)]
Merge remote branch 'pietern/redis-cli-fix'
antirez [Thu, 26 Aug 2010 11:18:56 +0000 (13:18 +0200)]
forgot the traceleaks var in tests set to 1, reverted to the default, 0
Pieter Noordhuis [Thu, 26 Aug 2010 11:18:19 +0000 (13:18 +0200)]
Fix parenthesis error on decrementing *argc
antirez [Thu, 26 Aug 2010 11:18:24 +0000 (13:18 +0200)]
fixed a memory leak in the new Set code
antirez [Thu, 26 Aug 2010 10:28:53 +0000 (12:28 +0200)]
computeObjectSwappability is now able to compute swappability for intset encoded sets
Pieter Noordhuis [Thu, 26 Aug 2010 10:13:51 +0000 (12:13 +0200)]
Fix type that was not renamed and compiler warning
antirez [Thu, 26 Aug 2010 10:10:16 +0000 (12:10 +0200)]
removed a duplicated ERRNO checking that is useless at all
antirez [Thu, 26 Aug 2010 10:04:24 +0000 (12:04 +0200)]
Merge remote branch 'pietern/intset-split'
Pieter Noordhuis [Thu, 26 Aug 2010 09:22:58 +0000 (11:22 +0200)]
Make the function intsetUpgrade self-contained
Pieter Noordhuis [Thu, 26 Aug 2010 09:06:30 +0000 (11:06 +0200)]
Expand macro's to functions for readability
Pieter Noordhuis [Wed, 25 Aug 2010 12:48:50 +0000 (14:48 +0200)]
Use fstat to detect if stdin was redirected
Pieter Noordhuis [Wed, 25 Aug 2010 12:15:41 +0000 (14:15 +0200)]
Make helper functions simpler
Pieter Noordhuis [Wed, 25 Aug 2010 12:08:32 +0000 (14:08 +0200)]
Comments in redis-cli tests
Pieter Noordhuis [Wed, 25 Aug 2010 11:39:11 +0000 (13:39 +0200)]
Add a newline to tty output after every reply
Pieter Noordhuis [Wed, 25 Aug 2010 08:05:50 +0000 (10:05 +0200)]
Re-introduce the interactive field so we can reconnect in interactive mode
Pieter Noordhuis [Wed, 25 Aug 2010 07:54:02 +0000 (09:54 +0200)]
Merge master and move argument splitting patch to sds.c
antirez [Tue, 24 Aug 2010 16:39:34 +0000 (18:39 +0200)]
redis-cli now supports automatically reconnection in interactive mode
antirez [Tue, 24 Aug 2010 16:08:09 +0000 (18:08 +0200)]
Issue 179 fixed, now redis-cli is able to parse correctly multi bulk replies with elements that are errors
antirez [Tue, 24 Aug 2010 15:09:25 +0000 (17:09 +0200)]
The pid file is now created only after the server is correctly initialied. It is also removed on sigterm and when the stack trace is produced after a sigbus or a sigsegv. This two changes should fix the Issue 175
antirez [Tue, 24 Aug 2010 14:25:00 +0000 (16:25 +0200)]
slave now detect lost connection during SYNC, fixing Issue 173
antirez [Tue, 24 Aug 2010 14:04:13 +0000 (16:04 +0200)]
slave with attached slaves now close the conection to all the slaves when the connection to the master is lost. Now a slave without a connected link to the master will refuse SYNC from other slaves. Enhanced the replication error reporting. All this will fix Issue 156
antirez [Tue, 24 Aug 2010 10:10:59 +0000 (12:10 +0200)]
BLPOPping clients are no longer subject to connection timeouts, fixing issues 155
antirez [Tue, 24 Aug 2010 09:49:05 +0000 (11:49 +0200)]
added tests for invalid bulk argument
antirez [Tue, 24 Aug 2010 09:45:05 +0000 (11:45 +0200)]
sanity check for the bulk argument in protocol parsing code, fixing issue 146
antirez [Tue, 24 Aug 2010 08:10:01 +0000 (10:10 +0200)]
changed the comments on top of redis-copy.rb to reflect what the program really does
antirez [Mon, 23 Aug 2010 15:06:38 +0000 (17:06 +0200)]
Fixed overflow detection in argument to long convertion function in general, and in expire/ttl pairs specifically, addressing issue 54
Pieter Noordhuis [Sat, 21 Aug 2010 09:25:13 +0000 (11:25 +0200)]
Rename iterator to setTypeIterator for consistency
Pieter Noordhuis [Sat, 21 Aug 2010 09:15:31 +0000 (11:15 +0200)]
Make SORT use the hybrid set accessors to allow sorting intsets
Pieter Noordhuis [Sat, 21 Aug 2010 09:03:56 +0000 (11:03 +0200)]
SORT tests with hash table encoded set as input
Pieter Noordhuis [Sat, 21 Aug 2010 09:02:22 +0000 (11:02 +0200)]
Make SORT tests use both ziplists and linked lists as input
Pieter Noordhuis [Sat, 21 Aug 2010 08:55:53 +0000 (10:55 +0200)]
Move SORT tests around
Pieter Noordhuis [Sat, 21 Aug 2010 08:54:31 +0000 (10:54 +0200)]
Don't abort test suite when the server block has a return value
Pieter Noordhuis [Fri, 20 Aug 2010 11:42:42 +0000 (13:42 +0200)]
Check if stroll return value was clamped
Pieter Noordhuis [Fri, 20 Aug 2010 10:40:29 +0000 (12:40 +0200)]
Merge branch 'master' into intset-split
Conflicts:
src/Makefile
src/t_set.c
antirez [Thu, 5 Aug 2010 09:36:39 +0000 (11:36 +0200)]
redis cli argument splitting is general and is now moved into the sds.c lib
Pieter Noordhuis [Wed, 4 Aug 2010 16:36:03 +0000 (18:36 +0200)]
Deprecate starting interactive mode using the -i flag
Pieter Noordhuis [Wed, 4 Aug 2010 16:16:39 +0000 (18:16 +0200)]
Set tty before going into interactive mode to get non-pretty output when
the commands are read from stdin.
Pieter Noordhuis [Wed, 4 Aug 2010 15:46:56 +0000 (17:46 +0200)]
Change output format for non-tty redis-cli execution
Pieter Noordhuis [Wed, 4 Aug 2010 15:16:05 +0000 (17:16 +0200)]
Let the output mode depend on having a tty or not
Pieter Noordhuis [Wed, 4 Aug 2010 15:02:13 +0000 (17:02 +0200)]
Tests for redis-cli in non-interactive mode
Minor change in redis-cli output for the (multi-)bulk response but this
will be fixed in the next commit.
Pieter Noordhuis [Wed, 4 Aug 2010 13:29:28 +0000 (15:29 +0200)]
Add tests for quotation in an interactive redis-cli session
Patched redis-cli to abort on unexpected quotation. This caused
redis-cli to get into an infinite, memory-consuming loop.
Pieter Noordhuis [Wed, 4 Aug 2010 12:15:52 +0000 (14:15 +0200)]
Tests for the interactive mode of redis-cli
Changed redis-cli to output the raw response for a bulk reply when it is
run in interactive mode instead of checking isatty.
antirez [Tue, 3 Aug 2010 12:25:22 +0000 (14:25 +0200)]
PERSIST: a fix and some basic test
antirez [Tue, 3 Aug 2010 12:19:20 +0000 (14:19 +0200)]
PERSIST command implemented
antirez [Tue, 3 Aug 2010 11:38:39 +0000 (13:38 +0200)]
replication test with expires
antirez [Tue, 3 Aug 2010 11:09:27 +0000 (13:09 +0200)]
Merge remote branch 'pietern/test-encoding-fix' into writeonexpire
antirez [Tue, 3 Aug 2010 11:08:32 +0000 (13:08 +0200)]
no longer passing tests due to the new write-on-volatile semantics modified/removed
antirez [Tue, 3 Aug 2010 10:26:30 +0000 (12:26 +0200)]
allow to set a new EXPIRE of an existing volatile key
antirez [Mon, 2 Aug 2010 19:37:39 +0000 (21:37 +0200)]
memory leak removed from expire propagation code
antirez [Mon, 2 Aug 2010 16:13:39 +0000 (18:13 +0200)]
support for write operations against expiring keys, by master-controlled expiring in replication and AOF synthesizing DEL operations
Pieter Noordhuis [Sun, 1 Aug 2010 09:20:26 +0000 (11:20 +0200)]
Fix assertion function on value encoding
antirez [Sat, 31 Jul 2010 11:55:06 +0000 (13:55 +0200)]
Merge remote branch 'pietern/list-test'
Pieter Noordhuis [Thu, 29 Jul 2010 20:13:31 +0000 (22:13 +0200)]
Change getDoubleFromObject to fail on NaN.
Return an error when the resulting value is not a number (NaN). Fix
ZUNIONSTORE/ZINTERSTORE to clean up when a weight argument is not a
double value.
Pieter Noordhuis [Thu, 29 Jul 2010 19:31:58 +0000 (21:31 +0200)]
Fix ZUNIONSTORE/ZINTERSTORE to never store a NaN score.
When +inf and -inf are added, the result is NaN. We don't want NaN
scores in a sorted set, so agreed on the result of this operation being
zero.
Pieter Noordhuis [Thu, 29 Jul 2010 11:53:52 +0000 (13:53 +0200)]
Use a large value to consistently trigger a list encoding,
even when the list is swapped out and in again.
Pieter Noordhuis [Thu, 29 Jul 2010 11:31:24 +0000 (13:31 +0200)]
ensure the value is swapped in before testing its encoding
antirez [Wed, 28 Jul 2010 16:56:52 +0000 (18:56 +0200)]
fix of the fix for the replication bug
antirez [Wed, 28 Jul 2010 16:42:02 +0000 (18:42 +0200)]
Fixed a replication bug in ZINTERSTORE.
In order to trigger the bug what's needed is to call ZINTERSTORE
resulting into an empty set created, bug against a key that already
existed. The command was not propagated, so the replica ended with the
key that the master removed. Sequence of command to reproduce:
redis-cli hset 446 34 905
redis-cli hset 446 393 911
redis-cli zadd 966 0.
085412045980529885 652
redis-cli zadd 645 0.
25081839284432045 280
redis-cli zinterstore 446 2 966 645
antirez [Wed, 28 Jul 2010 12:08:46 +0000 (14:08 +0200)]
better random dataset creation function in test. master-slave replication test now is able to save the two datasets in CSV when an inconsistency is detected.
antirez [Tue, 27 Jul 2010 13:26:08 +0000 (15:26 +0200)]
fixed a ziplist bug about encoding of integer values overflowing 64 bit
antirez [Tue, 27 Jul 2010 12:46:39 +0000 (14:46 +0200)]
removed test code having bad effects...
antirez [Tue, 27 Jul 2010 12:42:11 +0000 (14:42 +0200)]
Consistency test improved
antirez [Tue, 27 Jul 2010 12:30:02 +0000 (14:30 +0200)]
malloc definition with deprecated attribute was duplicated, one removed
antirez [Tue, 27 Jul 2010 08:09:26 +0000 (10:09 +0200)]
STRLEN command implemented
antirez [Tue, 27 Jul 2010 08:00:38 +0000 (10:00 +0200)]
hash table example commented out in dict.c
antirez [Tue, 27 Jul 2010 07:36:42 +0000 (09:36 +0200)]
use the function deprecated attribute if compiling with GCC to get warnings for malloc/free usages. We always want to use our zmalloc/zfree versions for memory usage tracking
Benjamin Kramer [Sat, 24 Jul 2010 21:20:00 +0000 (23:20 +0200)]
Add zcalloc and use it where appropriate
calloc is more effecient than malloc+memset when the system uses mmap to
allocate memory. mmap always returns zeroed memory so the memset can be
avoided. The threshold to use mmap is 16k in osx libc and 128k in bsd
libc and glibc. The kernel can lazily allocate the pages, this reduces
memory usage when we have a page table or hash table that is mostly
empty.
This change is most visible when you start a new redis instance with vm
enabled. You'll see no increased memory usage no matter how big your
page table is.
Benjamin Kramer [Sat, 24 Jul 2010 21:10:42 +0000 (23:10 +0200)]
Remove _dictAlloc and friends
zmalloc calls abort() so _dictPanic will never be called.
Benjamin Kramer [Sat, 24 Jul 2010 20:37:01 +0000 (22:37 +0200)]
Reduce code duplication
Pieter Noordhuis [Fri, 23 Jul 2010 11:08:35 +0000 (13:08 +0200)]
exit with non-zero status when there are failed tests
antirez [Thu, 22 Jul 2010 21:31:40 +0000 (23:31 +0200)]
don't open/close log file if log level is not matched
Pieter Noordhuis [Thu, 22 Jul 2010 14:06:27 +0000 (16:06 +0200)]
fix rare condition where 'key' would already be destroyed while is was needed later on
antirez [Thu, 22 Jul 2010 13:48:57 +0000 (15:48 +0200)]
defensive programming: set o->ptr to NULL before freeing objects
antirez [Thu, 22 Jul 2010 12:48:45 +0000 (14:48 +0200)]
other shared objects where created in the I/O thread in createStringObjectFromLongLong. Fixed as well.
antirez [Thu, 22 Jul 2010 11:12:24 +0000 (13:12 +0200)]
minor typo fixed in a comment
antirez [Thu, 22 Jul 2010 11:08:02 +0000 (13:08 +0200)]
don't use object sharing inside I/O threads, as a fix for a well known instability of VM introduced with the new object sharing code
antirez [Wed, 21 Jul 2010 11:16:26 +0000 (13:16 +0200)]
vm_blocked_clients count fixed in INFO, thanks to Pietern Noordhuis
antirez [Fri, 16 Jul 2010 21:56:18 +0000 (23:56 +0200)]
TODO list modified, trivial change to source code