]>
git.saurik.com Git - redis.git/log
antirez [Fri, 13 Apr 2012 13:12:16 +0000 (15:12 +0200)]
EVAL errors are more clear now.
antirez [Fri, 13 Apr 2012 12:54:49 +0000 (14:54 +0200)]
Use Lua tostring() before concatenation.
antirez [Fri, 13 Apr 2012 11:40:57 +0000 (13:40 +0200)]
Tests modified to match the new global protection implementation.
antirez [Fri, 13 Apr 2012 11:36:08 +0000 (13:36 +0200)]
mt.declared is no longer needed.
Lua global protection can now be simpified becuase we no longer have the
global() function. It's useless to occupy memory with this table, it is
also not faster because the metamethods we use are only called when a
global object does not exist or we are trying to create it from a
script.
antirez [Fri, 13 Apr 2012 11:26:59 +0000 (13:26 +0200)]
Stop access to global vars. Not configurable.
After considering the interaction between ability to delcare globals in
scripts using the 'global' function, and the complexities related to
hanlding replication and AOF in a sane way with globals AND ability to
turn protection On and Off, we reconsidered the design. The new design
makes clear that there is only one good way to write Redis scripts, that
is not using globals. In the rare cases state must be retained across
calls a Redis key can be used.
antirez [Fri, 13 Apr 2012 10:13:02 +0000 (12:13 +0200)]
Globals protection global() function modified for speed and correctness.
antirez [Fri, 13 Apr 2012 09:48:45 +0000 (11:48 +0200)]
Tests for lua globals protection.
antirez [Fri, 13 Apr 2012 09:16:50 +0000 (11:16 +0200)]
Scripting: globals protection can now be switched on/off.
antirez [Thu, 29 Mar 2012 10:02:28 +0000 (12:02 +0200)]
Protect globals access in Lua scripting.
antirez [Fri, 13 Apr 2012 14:13:56 +0000 (16:13 +0200)]
A few var names fixed in Makefile.
I modified it for error in a previous commit doing search & replace.
antirez [Thu, 12 Apr 2012 09:50:18 +0000 (11:50 +0200)]
Print arch bits with redis-server -v
antirez [Thu, 12 Apr 2012 09:49:52 +0000 (11:49 +0200)]
memtest.c: integer overflow fixed.
antirez [Thu, 12 Apr 2012 09:51:58 +0000 (11:51 +0200)]
Make gcov fixed.
antirez [Thu, 12 Apr 2012 09:09:38 +0000 (11:09 +0200)]
Makefile now introduces Redis-specific CFLAGS / LDFLAGS. Gcov target fixed. Added comments to describe how it works.
Erik Dubbelboer [Wed, 11 Apr 2012 15:04:31 +0000 (17:04 +0200)]
added explanation for the magic 511 backlog number
antirez [Wed, 11 Apr 2012 10:12:30 +0000 (12:12 +0200)]
Makefile.dep updated.
antirez [Wed, 11 Apr 2012 10:12:05 +0000 (12:12 +0200)]
make dep: redirect output to Makefile.dep.
antirez [Wed, 11 Apr 2012 09:58:32 +0000 (11:58 +0200)]
Make inline functions rioRead/Write/Tell static. This fixes issue #447.
antirez [Wed, 11 Apr 2012 09:32:22 +0000 (11:32 +0200)]
Macros ULONG_ONEZERO / ULONG_ZEROONE were inverted in #ifdef to test 32/64 bit arch.
Pieter Noordhuis [Sun, 25 Mar 2012 02:25:03 +0000 (19:25 -0700)]
Clean up Makefiles
Remove unused variables. Instead of overriding non-standard variables
such as ARCH and PROF, use standard variables CFLAGS and LDFLAGS to
override Makefile settings. Move dependencies generated by `make dep` to
a separate file.
Pieter Noordhuis [Sun, 25 Mar 2012 02:33:08 +0000 (19:33 -0700)]
Everything x86 is little endian
antirez [Tue, 10 Apr 2012 14:46:29 +0000 (16:46 +0200)]
Minor MIGRATE implementation simplification about ttl handling.
antirez [Tue, 10 Apr 2012 14:48:28 +0000 (16:48 +0200)]
Check write(2) return value to avoid warnings, because in this context failing write is not critical.
antirez [Tue, 10 Apr 2012 14:34:51 +0000 (16:34 +0200)]
Version 2.5.6.
antirez [Sun, 8 Apr 2012 09:16:40 +0000 (11:16 +0200)]
Test for maxclients.
antirez [Mon, 9 Apr 2012 10:33:57 +0000 (12:33 +0200)]
dump/restore fixed to use the new crc64 API.
antirez [Tue, 10 Apr 2012 13:47:10 +0000 (15:47 +0200)]
It is now possible to enable/disable RDB checksum computation from redis.conf or via CONFIG SET/GET. Also CONFIG SET support added for rdbcompression as well.
antirez [Mon, 9 Apr 2012 20:40:41 +0000 (22:40 +0200)]
RDB files now embed a crc64 checksum. Version of RDB bumped to 5.
antirez [Mon, 9 Apr 2012 10:36:44 +0000 (12:36 +0200)]
Fixed compilation of new rio.c changes (typos and so forth.)
antirez [Mon, 9 Apr 2012 10:33:09 +0000 (12:33 +0200)]
Add checksum computation to rio.c
antirez [Mon, 9 Apr 2012 10:20:47 +0000 (12:20 +0200)]
crc64.c modified for incremental computation.
antirez [Mon, 9 Apr 2012 09:11:00 +0000 (11:11 +0200)]
rio.c file somewhat documented so that the casual reader can understand what's going on without reading the code.
antirez [Sat, 7 Apr 2012 11:22:04 +0000 (13:22 +0200)]
Added new test to check that "CONFIG appendonly no" actually kills the background AOF operation in progress if any.
antirez [Sat, 7 Apr 2012 10:11:23 +0000 (12:11 +0200)]
For coverage testing use exit() instead of _exit() when termiating saving children.
antirez [Sat, 7 Apr 2012 09:26:24 +0000 (11:26 +0200)]
Tests for MONITOR.
antirez [Sat, 7 Apr 2012 09:14:52 +0000 (11:14 +0200)]
New client info field added to CLIENT LIST output: multi, containing the length of the current pipeline. Test modified accordingly.
antirez [Sat, 7 Apr 2012 00:10:48 +0000 (02:10 +0200)]
Never used function stringObjectEqualsMs() removed.
antirez [Sat, 7 Apr 2012 00:03:29 +0000 (02:03 +0200)]
Removed dead code: function rdbSaveTime() is no longer used since RDB now saves expires in milliseconds.
antirez [Fri, 6 Apr 2012 21:52:53 +0000 (23:52 +0200)]
Two new tests for BGREWRTIEAOF.
Check for scheduled rewrite if a BGSAVAE is in progress.
Check for error if a rewrite is already in progress.
antirez [Fri, 6 Apr 2012 21:52:28 +0000 (23:52 +0200)]
redis.tcl: no longer leave unread replies if an error happens during a MULTI/EXEC block.
antirez [Fri, 6 Apr 2012 19:12:50 +0000 (21:12 +0200)]
New INFO field in persistence section: bgrewriteaof_scheduled.
antirez [Fri, 6 Apr 2012 10:27:31 +0000 (12:27 +0200)]
version bumped to 2.5.5
jokea [Fri, 6 Jan 2012 10:56:07 +0000 (18:56 +0800)]
implement aeWait using poll(2). Fixes issue #267.
antirez [Thu, 5 Apr 2012 13:52:08 +0000 (15:52 +0200)]
expireGenericCommand(): better variable names and a top-comment that describes the function's behavior.
Premysl Hruby [Wed, 28 Mar 2012 09:48:36 +0000 (11:48 +0200)]
for (p)expireat use absolute time, without double recomputation
Premysl Hruby [Wed, 28 Mar 2012 10:03:32 +0000 (12:03 +0200)]
fix mstime() ommited while comparing if key is already expired
Premysl Hruby [Wed, 4 Apr 2012 17:17:32 +0000 (19:17 +0200)]
add support for generation of lcov coverage reports
Premysl Hruby [Thu, 5 Apr 2012 06:25:22 +0000 (08:25 +0200)]
future-proof version comparison
antirez [Wed, 4 Apr 2012 17:54:23 +0000 (19:54 +0200)]
On slow computers, 10 seconds are not enough for this heavy replication test.
Premysl Hruby [Tue, 3 Apr 2012 15:40:31 +0000 (17:40 +0200)]
allocate alternate signal stack, change of sigaction flags for sigterm
Premysl Hruby [Tue, 3 Apr 2012 12:18:35 +0000 (14:18 +0200)]
in kill_server send the signal once, then wait for up to 5sec before sending lethal SIGKILL
Premysl Hruby [Tue, 3 Apr 2012 11:32:49 +0000 (13:32 +0200)]
new option for choosing number of test clients to run
Premysl Hruby [Mon, 2 Apr 2012 09:56:03 +0000 (11:56 +0200)]
remove mentions of VM in comments
antirez [Wed, 4 Apr 2012 16:31:37 +0000 (18:31 +0200)]
Structure field controlling the INFO field master_link_down_since_seconds initialized correctly to avoid strange INFO output at startup when a slave has yet to connect to its master.
antirez [Wed, 4 Apr 2012 13:38:13 +0000 (15:38 +0200)]
New "os" field in INFO output providing information about the operating system.
antirez [Wed, 4 Apr 2012 13:11:30 +0000 (15:11 +0200)]
SLAVEOF is not a write command.
antirez [Wed, 4 Apr 2012 13:11:17 +0000 (15:11 +0200)]
Print milliseconds of the current second in log lines timestamps. Sometimes precise timing is very important for debugging.
antirez [Tue, 3 Apr 2012 13:29:47 +0000 (15:29 +0200)]
redis-cli help.h updated. Script to generate it updated as well.
antirez [Tue, 3 Apr 2012 13:10:51 +0000 (15:10 +0200)]
More MIGRATE tests.
antirez [Tue, 3 Apr 2012 13:10:42 +0000 (15:10 +0200)]
Another fix for MIGRATE.
antirez [Tue, 3 Apr 2012 10:17:40 +0000 (12:17 +0200)]
Two fixed for MIGRATE: fix TTL propagation and fix transferring of data bigger than 64k.
antirez [Tue, 3 Apr 2012 09:53:45 +0000 (11:53 +0200)]
When the user-provided 'maxclients' value is too big for the max number of files we can open, at least try to search the max the OS is allowing (in steps of 256 filedes).
antirez [Tue, 3 Apr 2012 08:31:51 +0000 (10:31 +0200)]
redis-trib.rb removed from 2.6 branch.
antirez [Mon, 2 Apr 2012 14:38:59 +0000 (16:38 +0200)]
MIGRATE test modified because the implementation changed.
antirez [Mon, 2 Apr 2012 14:38:24 +0000 (16:38 +0200)]
MIGRATE now let the client distinguish I/O errors and timeouts from other erros.
antirez [Mon, 2 Apr 2012 12:27:04 +0000 (14:27 +0200)]
Version bumped 2.4.5
antirez [Mon, 2 Apr 2012 09:44:25 +0000 (11:44 +0200)]
DUMP, RESTORE, MIGRATE tests.
antirez [Mon, 2 Apr 2012 12:21:59 +0000 (14:21 +0200)]
New verions of DUMP, RESTORE, MIGRATE back ported from unstable to 2.6
antirez [Mon, 2 Apr 2012 10:31:44 +0000 (12:31 +0200)]
CRC64 implementation added to Redis code base.
antirez [Sat, 31 Mar 2012 15:08:40 +0000 (17:08 +0200)]
Write RDB magic using a REDIS_RDB_VERSION define that is defined inside rdb.h
ThePicard [Sun, 1 Apr 2012 06:39:58 +0000 (23:39 -0700)]
Fixed a typo in install_server.sh
antirez [Sat, 31 Mar 2012 09:23:30 +0000 (11:23 +0200)]
syncio.c calls in replication.c fixed for the new millisecond timeout API.
antirez [Sat, 31 Mar 2012 09:21:45 +0000 (11:21 +0200)]
Better syncio.c with millisecond resolution.
antirez [Fri, 30 Mar 2012 08:39:34 +0000 (10:39 +0200)]
Purely aesthetic code change.
Joseph Jang [Thu, 29 Mar 2012 17:06:53 +0000 (02:06 +0900)]
Fixed a memory leak with replication
occurs when two or more dbs are replicated and at least one of them is >db10
antirez [Fri, 30 Mar 2012 08:26:07 +0000 (10:26 +0200)]
Regression test for issue 417 (memory leak when replicating to DB with id >= 10)
antirez [Thu, 29 Mar 2012 07:33:29 +0000 (09:33 +0200)]
Fixed typo in comment: "te" -> "the".
antirez [Thu, 29 Mar 2012 07:24:02 +0000 (09:24 +0200)]
Fix for slaves chains. Force resync of slaves (simply disconnecting them) when SLAVEOF turns a master into a slave.
antirez [Wed, 28 Mar 2012 18:47:50 +0000 (20:47 +0200)]
Test for redis.sha1hex().
Nathan Fritz [Wed, 28 Mar 2012 18:10:24 +0000 (11:10 -0700)]
added redis.sha1hex(string) as lua scripting function.
(The original implementation was modified by @antirez to conform Redis
coding standards.)
antirez [Wed, 28 Mar 2012 11:51:23 +0000 (13:51 +0200)]
Fixes for redisLogFromHandler().
antirez [Wed, 28 Mar 2012 11:45:39 +0000 (13:45 +0200)]
Log from signal handlers is now safer.
antirez [Wed, 28 Mar 2012 08:55:17 +0000 (10:55 +0200)]
Redis test: regexp to check if valgrind reported errors modified. Now we no longer look at the total count because this includes "possibly lost" bytes that are not interesting for Redis (tons of false positives because of how sds.c works).
antirez [Tue, 27 Mar 2012 14:54:53 +0000 (16:54 +0200)]
define zlibc_free() in a way that is not shadowed by jemalloc.
antirez [Tue, 27 Mar 2012 13:24:33 +0000 (15:24 +0200)]
Produce the watchlog warning log in a way that is safer from a signal handler. Fix a memory leak in the backtrace generation function.
antirez [Tue, 27 Mar 2012 11:48:53 +0000 (13:48 +0200)]
Mask SIGALRM everything but in the main thread.
This is required to ensure that the signal will be delivered to the main
thread when the watchdog timer expires.
antirez [Tue, 27 Mar 2012 10:11:37 +0000 (12:11 +0200)]
Correctly set the SIGARLM timer for the software watchdog.
antirez [Tue, 27 Mar 2012 09:47:51 +0000 (11:47 +0200)]
Redis software watchdog.
huangz1990 [Thu, 15 Mar 2012 06:27:14 +0000 (14:27 +0800)]
fix typo
Premysl Hruby [Tue, 27 Mar 2012 15:19:54 +0000 (17:19 +0200)]
fix time() instead of mstime() in expireIfNeeded
Premysl Hruby [Tue, 27 Mar 2012 15:39:58 +0000 (17:39 +0200)]
use server.unixtime instead of time(NULL) where possible (cluster.c not checked though)
Premysl Hruby [Tue, 27 Mar 2012 16:18:57 +0000 (18:18 +0200)]
declare hashDictType as external too
Premysl Hruby [Tue, 27 Mar 2012 16:46:51 +0000 (18:46 +0200)]
remove disk-store related comments
antirez [Tue, 27 Mar 2012 08:33:45 +0000 (10:33 +0200)]
SIGSEGV handler refactored so that we can reuse stack trace and current client logging functionalities in other contexts.
antirez [Sun, 25 Mar 2012 09:43:19 +0000 (11:43 +0200)]
CONFIG RESETSTAT resets two more fields.
antirez [Sun, 25 Mar 2012 09:27:35 +0000 (11:27 +0200)]
New INFO field aof_delayed_fsync introduced.
This new field counts all the times Redis is configured with AOF enabled and
fsync policy 'everysec', but the previous fsync performed by the
background thread was not able to complete within two seconds, forcing
Redis to perform a write against the AOF file while the fsync is still
in progress (likely a blocking operation).
antirez [Sun, 25 Mar 2012 09:18:24 +0000 (11:18 +0200)]
Version is now 2.5.3.
antirez [Sun, 25 Mar 2012 08:57:34 +0000 (10:57 +0200)]
convert-zipmap-hash-on-load false positive fixed.
Apparently because the sample RDB file was not copied before every test
Redis had a chance to replace it with a newly written one, so that the
next test could fail.
antirez [Sat, 24 Mar 2012 11:06:56 +0000 (12:06 +0100)]
When running the test in valgrind mode, pass the right flags to show memory leaks stack traces but only including the "definitely lost" items.
antirez [Sat, 24 Mar 2012 10:52:56 +0000 (11:52 +0100)]
Add used allocator in redis-server -v output.