bharat42/credis

Automatically exported from code.google.com/p/credis

credis's Issues

freebsd compilation

it doesnt compile on freebsd..

just add these include on credis.c :

#include <netinet/in.h>

Original issue reported on code.google.com by [email protected] on 29 Jul 2010 at 10:05

Multibulk lost on publisher subscriber schema

When in a publisher/subscriber schema some published messages may be lost with 
curren implementation of cr_readln

I attach two test files, one is a publisher, and the oter is a subscriber. Just 
compile and launch the subscriber before publisher.

*Some frames get lost because of line 

  rc = cr_receivedata(rhnd->fd, rhnd->timeout, buf->data + buf->len, avail);

When filling the buffer, it can be populated with more than a message. Only 
first message gets parsed and other messages are lost.

If that line is modified to
  rc = cr_receivedata(rhnd->fd, rhnd->timeout, buf->data + buf->len, 1);

Recompile and run the same test, shows no message lost.

Some speed will be lost with this patch (i haven't bencharmked this, and 
probably won't do) but no message will be lost.

I think that is a trade off between speed and fiability.

Tests and patch are provided.

Any other implementation with no speed lost will be welcome.

Original issue reported on code.google.com by [email protected] on 24 Apr 2012 at 1:44

Attachments:

• Makefile.patch
• credis.patch
• credis-publisher-test.c
• credis-subscriber-test.c

Suggestion about using timed_connect instead sync connect

If you are using multiple servers (master + several slaves) sometimes some
of servers can be not available. Standard connect function on blocking
socket will take from 75 second to some minutes while attempt to create
connection with server which is unavailable. I suggest you use timed
connect (please see my patch in attach) with timeout from credis_connect
parameters. That will allow to report about connection problem within
user's provided timeout.

Regards, Dmitriy Lyfar.

Original issue reported on code.google.com by [email protected] on 5 Jun 2010 at 12:05

Attachments:

• timed_connect.patch

credis_keys on redis 2.x (issue 25)

hi romfelt,

there's an issue missing in your patch.
if you are using redis with security option (requirepass) you habe to
revalidate rhd->version after credis_auth otherwise rhd->version has no value.

redargs

kappe

Original issue reported on code.google.com by [email protected] on 9 Sep 2010 at 8:36

credis_keys returns -97

I'm using credis 0.2.3 on redis 2.0.0. my problem: credis_keys returns -97!

char **v;
int  ret;

ret = credis_keys( redis, "name:*", &v );

--> ret = -97

Original issue reported on code.google.com by [email protected] on 5 Sep 2010 at 7:20

add support for binary key and value

just a patch with cr_sendvandreceive function, which could send command 
base on a iovec array, and support binary key and value

int credis_getv(REDIS rhnd, const char *key, size_t key_len, char **val, size_t 
*val_len);
int credis_setv(REDIS rhnd, const char *key, size_t key_len, const char *val, 
size_t val_len);
int credis_existsv(REDIS rhnd, const char *key, size_t key_len);

Original issue reported on code.google.com by [email protected] on 9 May 2010 at 4:48

Bloat your makefile

Would you like to bloat your makefile a little?
It shouldn't break anything. I guess adding configure
would be overkill.
Cheers

Original issue reported on code.google.com by [email protected] on 9 Nov 2010 at 12:24

Attachments:

• bloat-makefile.patch

cr_push missing char length

What steps will reproduce the problem?
1. cr_push(redis, "mykey", "myvalue");

What is the expected output? What do you see instead?
Redis monitor shows:
RPUSH mykey 0

It should show:
RPUSH mykey 7
myvalue

What version of the product are you using? On what operating system?
SVN r13

Please provide any additional information below.
FIX:

In credis.c replace line 747 with this:

    int rc = cr_sendfandreceive(rhnd, CR_INT, "%s %s %d\r\n%s\r\n", 
                                left==1?"LPUSH":"RPUSH", key, strlen(val), val);

Original issue reported on code.google.com by [email protected] on 9 Feb 2010 at 2:12

Linking with c++

Linking with c++ sources is currently not possible. This should be added to
credis.h :

#if defined(__cplusplus)
#define __CREDIS_CLINKAGEBEGIN extern "C" {
#define __CREDIS_CLINKAGEEND }
#else
#define __CREDIS_CLINKAGEBEGIN
#define __CREDIS_CLINKAGEEND
#endif
__CREDIS_CLINKAGEBEGIN

[...]

__CREDIS_CLINKAGEEND

Also, some function could take a const char*, instead of a char* (eg: 2nd
argument of credis_get). This would avoid using a const_cast in c++.

Original issue reported on code.google.com by [email protected] on 7 Dec 2009 at 6:36

User-friendly error from credis-test when Redis server is not present

credis-test.c will segfault if a connection to the redis server cannot be made. 
 There is no error handling, so this is confusing to the new user.  Here is a 
trivial patch to fix it:

--- credis-test.c.old   2010-05-19 16:50:09.000000000 -0400
+++ credis-test.c   2010-07-18 00:58:41.508511597 -0400
@@ -73,6 +73,10 @@

 int main(int argc, char **argv) {
   REDIS redis = credis_connect(NULL, 0, 10000);
+  if (redis == NULL) {
+    printf("Error connecting to Redis server.  Please start server to run 
tests.\n");
+    exit(1);
+  }
   REDIS_INFO info;
   char *val, **valv, lstr[50000];
   const char *keyv[] = {"kalle", "adam", "unknown", "bertil", "none"};

Original issue reported on code.google.com by aaron.hamid on 18 Jul 2010 at 5:01

credis_zincrby doesn't work

What steps will reproduce the problem?
1. credis_zincrby(rh, "mykey", 1.0, "mymember", &new_score)
2. ZRANGE mykey 0 -1 (in telnet connect to server)

What is the expected output? What do you see instead?

I would expect something like:
*1
$8
mymember

Instead I get:
*1
$1
8

What version of the product are you using? On what operating system?

credis 0.2.3
redis 2.2.2

Please provide any additional information below.

It appears that credis increases the score of strlen("mymember") rather than 
"mymember".

Original issue reported on code.google.com by [email protected] on 15 Mar 2011 at 5:19

undefined HDEL

HDEL is missing.

This patch adds hdel code to credis.c and credis.h


Please check compatibility with old redis API. Should be the same but i haven't 
tested against old redis versions.

Original issue reported on code.google.com by [email protected] on 8 May 2012 at 7:07

Attachments:

• credis.c.patch
• credis.h.patch

Wrong command format for redis-2.2.2

Everyting was going fine while I was using redis-2.0.4
But when I replaced it with new redis-2.2.2 I've encountered a problem

When I use:

credis_hset(rhnd, "hash", "field", "value_of_field");

I see following in monitoring:

+1300875825.517306 "HSET" "hash" "field" "value_of_field"   <-- in redis-2.0.4
1300875854.226830 "HSET" "hash" "field" "14"                <-- in redis-2.2.2

Looks like HSET command is not a bulk command anymore.

I've attached a small patch which fixes HSET command but looks like there is 
more commands with the same situation (see issue #37)

Original issue reported on code.google.com by [email protected] on 23 Mar 2011 at 10:39

Random memory errors if Redis returns (integer) 0

What steps will reproduce the problem?
1. int returnValue = credis_lrange(redis,"not_existing",0,-1);


What is the expected output? What do you see instead?
If the key does not exist returnValue is -97 and I get strange memory issues. 
Also the program 
hangs for a while.


What version of the product are you using? On what operating system?
SVN r13

Please provide any additional information below.
I think the problem is somewhere in the cr_receivedata() method.

Original issue reported on code.google.com by [email protected] on 9 Feb 2010 at 3:49

In debug mode most calls print "cr_readln() @ 253: need to read more data"

What steps will reproduce the problem?
1. credis_get(), credis_lrange(),...

SVN r13

Is this just a left over debug print or are there any problems with the 
framework / connection?

Original issue reported on code.google.com by [email protected] on 9 Feb 2010 at 1:09

credis_exists returns -1 for non existing key and 0 for existing key

What steps will reproduce the problem?
1. credis_exists(redis,"non_existing"); -> -1
2. credis_exists(redis,"existing"); -> 0

What is the expected output? What do you see instead?
Redis return 0 for non existing and 1 for existing keys. 

What version of the product are you using? On what operating system?
SVN r13

Please provide any additional information below.
It would be better to follow the Redis docs.

Original issue reported on code.google.com by [email protected] on 9 Feb 2010 at 4:11

credis_connect("localhost", 0, 2000) results in segfault

rh = credis_connect("localhost", 0, 2000) results in segfault

Original issue reported on code.google.com by [email protected] on 6 Oct 2010 at 8:07

In Redis 2.0 RPUSH and LPUSH returns integer reply not inline as in 1.2.6

RPUSH key string
LPUSH key string
Time complexity: O(1)

Add the string value to the head (LPUSH) or tail (RPUSH) of the list stored at 
key. If the key does not exist an empty list is created just before the append 
operation. If the key exists but is not a List an error is returned.
Return value

Integer reply, specifically, the number of elements inside the list after the 
push operation.

Original issue reported on code.google.com by [email protected] on 15 Sep 2010 at 8:47

64-bit signed integer

Hi,

The Redis documentation for INCR, INCRBY, DECR, DECRBY states that Redis treats 
the string as a 64-bit signed integer.  Would it make sense for the 
corresponding credis functions to return long rather than int?

Thanks,

Dean

Original issue reported on code.google.com by [email protected] on 5 Sep 2010 at 7:21

Problem with keys and rpush

What steps will reproduce the problem?

Using credis_keys and credis_rpush
to get CREDIS_ERR_PROTOCOL


What is the expected output? What do you see instead?

expect:
Get results as the redis-cli

getting:
Error in keys functions and rpush function:
CREDIS_ERR_PROTOCOL

What version of the product are you using? On what operating system?
redis-server: 1.3.8
credis: trunk

Please provide any additional information below.

1.3.8 corresponds to 'credis_keys function' CR_MULTIBULK and 1.3.8 corresponds 
to 'cr_push fuction' CR_INT 

Original issue reported on code.google.com by [email protected] on 2 Nov 2010 at 12:25

credis_set bug

What steps will reproduce the problem?
1.call credis_set
2.provide fh, key and value args to the function

What is the expected output? What do you see instead?
KEY should be set to VALUE
Instead it gets set to sizeof(VALUE) of the provided VALUE e.g. 5 in case of 
KEY "hello".

What version of the product are you using? On what operating system?
credis 0.2.3 with redis 2.4.17 on RHEL 6.3 x86_64

Please provide any additional information below.
The attached test-client.c is compiled with the following line:
gcc -g -O2 -Wall -DPRINTDEBUG credis.c test-client.c -o test-client

Uppon executing the client, debug shows:

credis_connect() @ 768: Connected to Redis version: 2.4.17

cr_sendandreceive() @ 602: Sending message: len=21, data=SET abra 7
kadabra

cr_readln() @ 423: received 37 bytes: +OK
-ERR unknown command 'kadabra'

I'm guessing it's a parse string issue, since debug shows:
data=SET abra <val lenght> kadabra
Bbut i'm no C expert.

Checking the result with the redis_cli tool:

./redis-cli GET abra
"7"

Original issue reported on code.google.com by [email protected] on 24 Sep 2012 at 7:18

Attachments:

• test-client.c

A software license for credis

It would be helpful if you could choose a license under which to release 
credis, e.g. is it BSD licensed like redis itself?

Original issue reported on code.google.com by [email protected] on 11 Mar 2010 at 6:07

Supress "... no effect" warnings

What steps will reproduce the problem?
1. Checkout credis
2. Run "make"

What is the expected output? What do you see instead?

Expected: No warnings.
Actual result:
- "statement with no effect"
- "left-hand operand of comma expression has no effect"

What version of the product are you using? On what operating system?

svn rev8, OS X 10.6.2, gcc version 4.2.1

Please provide any additional information below.

A workarond for this would be:

=== modified file 'credis.c'
--- credis.c    2009-12-04 13:40:49 +0000
+++ credis.c    2010-01-01 06:16:35 +0000
@@ -92,7 +92,7 @@

 /* add -DPRINTDEBUG to CPPFLAGS in Makefile for debug outputs */
 #ifndef PRINTDEBUG
-#define DEBUG
+#define DEBUG(args...)
 #else
 #define DEBUG(args...) cr_debug(__FILE__, __FUNCTION__, __LINE__, ##args)

Original issue reported on code.google.com by [email protected] on 1 Jan 2010 at 6:18

credis can't compile successfully using msvc-9.0

1>------ Build started: Project: credis, Configuration: Debug Win32 ------
1>Compiling...
1>credis-test.c
1>e:\credis\credis-test.c(35) : fatal error C1083: Cannot open include file: 
'sys/time.h': No such file or directory
1>credis.c
1>e:\credis\credis.c(786) : error C2143: syntax error : missing ';' before 
'type'
1>e:\credis\credis.c(791) : error C2065: 'he' : undeclared identifier
1>e:\credis\credis.c(791) : warning C4047: '=' : 'int' differs in levels of 
indirection from 'hostent *'
1>e:\credis\credis.c(793) : error C2065: 'he' : undeclared identifier
1>e:\credis\credis.c(793) : warning C4047: '=' : 'int' differs in levels of 
indirection from 'hostent *'
1>e:\credis\credis.c(795) : error C2065: 'he' : undeclared identifier
1>e:\credis\credis.c(795) : warning C4047: '==' : 'int' differs in levels of 
indirection from 'void *'
1>e:\credis\credis.c(798) : error C2065: 'he' : undeclared identifier
1>e:\credis\credis.c(798) : error C2223: left of '->h_addr_list' must point to 
struct/union
1>e:\credis\credis.c(798) : warning C4022: 'memcpy' : pointer mismatch for 
actual parameter 2
1>e:\credis\credis.c(798) : error C2198: 'memcpy' : too few arguments for call
1>e:\credis\credis.c(817) : warning C4013: 'fcntl' undefined; assuming extern 
returning int
1>e:\credis\credis.c(817) : error C2065: 'F_GETFL' : undeclared identifier
1>e:\credis\credis.c(818) : error C2065: 'F_SETFL' : undeclared identifier
1>e:\credis\credis.c(818) : error C2065: 'O_NONBLOCK' : undeclared identifier
1>e:\credis\credis.c(823) : error C2065: 'EINPROGRESS' : undeclared identifier
1>e:\credis\credis.c(829) : warning C4133: 'function' : incompatible types - 
from 'int *' to 'char *'
1>e:\credis\credis.c(1860) : warning C4013: 'strcasecmp' undefined; assuming 
extern returning int
1>Generating Code...
1>Project : warning PRJ0018 : The following environment variables were not 
found:
1>$(hge)
1>Build log was saved at "file://e:\credis\Debug\BuildLog.htm"
1>credis - 12 error(s), 7 warning(s)
========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========

Original issue reported on code.google.com by [email protected] on 27 Nov 2011 at 8:52

[patch] credis_keys does not work

Key searches are returned as a space separated list of key names, in a single 
string. Therefore, it has to use CR_BULK instead of CR_MULTIBULK, and parse 
the string.

Original issue reported on code.google.com by mathieu.rene on 13 Mar 2010 at 9:14

Attachments:

• credis.diff

credis_lrem() malformed request bug - fix provided

credis.cc:credis_lrem() - this request is malformed, and the request hangs 
forever. The following is a correct implementation.

int credis_lrem(Redis rhnd, const char *key, int count, const char *val)
{
  return cr_sendfandreceive(rhnd, CR_INT, "LREM %s %d %d\r\n%s\r\n", 
                                     key, count, (int)strlen(val), val);
}

Original issue reported on code.google.com by [email protected] on 18 Mar 2010 at 12:48

vsnprintf in windows returns -1 when the buffer is too small, instead of the needed size

By the way, there's a fix needed for WIN32, because vsnprintf in windows 
returns -1 when the buffer is too small, instead of the needed size. So in 
cr_sendfandreceive you should replace

 va_start(ap, format);
 rc = vsnprintf(buf->data, buf->size, format, ap);
 va_end(ap);

 if (rc < 0)
   return -1;

 if (rc >= buf->size) {
   DEBUG("truncated, get more memory and try again");
   if (cr_moremem(buf, rc - buf->size + 1))
     return CREDIS_ERR_NOMEM;

   va_start(ap, format);
   rc = vsnprintf(buf->data, buf->size, format, ap);
   va_end(ap);
 }

 buf->len = rc;


with 

  va_start(ap, format);
  rc = vsnprintf(buf->data, buf->size, format, ap);
  va_end(ap);

  if (rc < 0) {
#ifdef WIN32
    rc = buf->size * 2;
#else
    return -1;
#endif
  }

  while (rc >= buf->size) {
    DEBUG("truncated, get more memory and try again");
    if (cr_moremem(buf, rc - buf->size + 1))
      return CREDIS_ERR_NOMEM;

    va_start(ap, format);
    rc = vsnprintf(buf->data, buf->size, format, ap);
    va_end(ap);

    if (rc < 0) {
#ifdef WIN32
        rc = buf->size * 2;
#else
        return -1;
#endif
    }
  }

  buf->len = rc;

Regards,
Fernickk

Original issue reported on code.google.com by [email protected] on 9 Sep 2010 at 8:43

After using credis_type(), the value of the previous key is altered

What steps will reproduce the problem?
1. Make a simple program to get a key, then get the type of the key. Then 
output both.
2.
3.

What is the expected output? What do you see instead?
I expect to see the true value of the key. Instead, the key returns something 
different, namely 
'ing'.

What version of the product are you using? On what operating system?
0.2.1
mac os x 10.6


Please provide any additional information below.

#include <stdio.h>
#include "credis.h"

int main(int argc, char **argv)
{
  REDIS redis = credis_connect(NULL, 0, 10000);

  char *val;
  credis_get(redis, "name", &val);
  int type = credis_type(redis, "name");
  printf("type: %d\n", type);
  printf("val: %s\n", val);

  credis_close(redis);
  return 0;
}

Original issue reported on code.google.com by [email protected] on 12 May 2010 at 2:45

HSET returns error when value is modified and shouldn't

When trying to modify a previously set value with hset the condition


  if (rc == 0 && rhnd->reply.integer == 0)
    rc = -1;

  return rc;

Will return error. 

This is a wrong behaviour in credis lib as long as i know. 

Redis return values for hset specifies that

    rhnd->reply.integer value is 1 for new set values, and 0 for modified values.

This patch avoid that hset return error when a value is modified by hset 
function.

I have noticed same behaviour is present in zadd. Same patch can be applied to 
that function too. 

Maybe other functions are affected too.

Original issue reported on code.google.com by [email protected] on 8 May 2012 at 7:18

Attachments:

• credis.c.patch

int rc = cr_sendfandreceive(rhnd, CR_INT, "DELETE %s\r\n", key);

What steps will reproduce the problem?
1. calling credis_del()

The redis protocoll defines delete as:

DEL key delete a key

So credis.c line 644:
int rc = cr_sendfandreceive(rhnd, CR_INT, "DELETE %s\r\n", key);
must be:
int rc = cr_sendfandreceive(rhnd, CR_INT, "DEL %s\r\n", key);

Original issue reported on code.google.com by [email protected] on 8 Feb 2010 at 9:53

[patch] PERSIST, HMSET, HINCRBY and HVALS realization

There's realization of PERSIST, HMSET, HINCRBY and HVALS commands attached.
It works for redis-2.2.2 and probably works for the earlier versions.

Original issue reported on code.google.com by [email protected] on 21 Apr 2011 at 7:55

Attachments:

• credis.diff

make install says not implemented

What steps will reproduce the problem?
1. wget
2. tar xfv
3. make install

What is the expected output? What do you see instead?

I expected it to say install undefined or actually work

What version of the product are you using? On what operating system?

0.2.3, Ubuntu

Please provide any additional information below.

Please either complete the install routines or remove them.  You are doing a 
disservice otherwise.

Original issue reported on code.google.com by smlikens on 9 Mar 2012 at 9:58

Adopt the new unified request protocol

Adopt the new unified request protocol: 
http://code.google.com/p/redis/wiki/ProtocolSpecification

Original issue reported on code.google.com by [email protected] on 30 Sep 2010 at 7:32

Add sipport for Pipelining

A Request/Response server can be implemented so that it is able to process new 
requests even if the client didn't already read the old responses. This way it 
is possible to send multiple commands to the server without waiting for the 
replies at all, and finally read the replies in a single step.
Redis supports pipelining since the very early days, so whatever version you 
are running, you can use pipelining with Redis.

There should be something like the batch request, which remembers the order of 
commands and places the results in accordance with this order.

Original issue reported on code.google.com by [email protected] on 14 Jan 2012 at 3:55

fixed bug in cr_push

What steps will reproduce the problem?
1. rc = credis_rpush(redis, "test", "123456789");

What is the expected output? What do you see instead?
expected output:
 redis> lpop test
 "123456789"
error output:
 redis> lpop test
 "9"

What version of the product are you using? On what operating system?
  redis-2.2.2

solution: 
static int cr_push(REDIS rhnd, int left, const char *key, const char *val)
{
  int rc;

  if (rhnd->version.major >= 2) {
    // rc = cr_sendfandreceive(rhnd, CR_INT, "%s %s %zu\r\n%s\r\n", 
    //                        left==1?"LPUSH":"RPUSH", key, strlen(val), val);
    // fixed :
    rc = cr_sendfandreceive(rhnd, CR_INT, "*3\r\n$5\r\n%s\r\n$%zu\r\n%s\r\n$%zu\r\n%s\r\n", 
                            left==1?"LPUSH":"RPUSH", strlen(key), key, strlen(val), val);

Original issue reported on code.google.com by [email protected] on 5 Apr 2011 at 1:44

credis_quit is declared but not implemented

What steps will reproduce the problem?
1. #include "credis.h"
2. write code calling credis_quit()
3. link!

What is the expected output? What do you see instead?

It should link and send QUIT command to server!

What version of the product are you using? On what operating system?

I'm on revision r66

Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 23 Apr 2012 at 2:50

EXC_BAD_ACCESS when trying to get key type before any access

What steps will reproduce the problem?
1. Make a simple C program to get the type of a key using credis_type(rh, 
"mykey")

What is the expected output? What do you see instead?

I expect to retrieve the type of the key that I'm querying. Instead, I get 
EXC_BAD_ACCESS. 
However, if I credis_get() a string key first, I am able to get the type 
successfully.


What version of the product are you using? On what operating system?
credis 0.2.1
Mac OS X 10.6


Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 11 May 2010 at 3:30

add support for lua scripting

http://redis.io/commands/eval

Original issue reported on code.google.com by [email protected] on 14 Jan 2012 at 4:01

comments for credis_sadd and credis_srem are wrong

In credis.h
 /* returns -1 if the given member was already a member of the set */
 int credis_sadd(REDIS rhnd, const char *key, const char *member);

 /* returns -1 if the given member is not a member of the set */
 int credis_srem(REDIS rhnd, const char *key, const char *member);

but when the given member was already a member of the set, credis_sadd 
return 1, so as the credis_srem.

vsn:credis-0.2.1

Original issue reported on code.google.com by [email protected] on 16 Apr 2010 at 7:50

credis_lrange mess up results on large lists

I have attached a demo script of the problems with larger lists.

I have two main problems:

1) credis_lrange() doesn't return the list entries correctly. It seems that the 
entries get mixed.
2) if CR_BUFFER_SIZE ist set too low, you will get a Program received signal:  
“EXC_BAD_ACCESS” 
error.

It would be awesome if you could fix that. This is really a blocker for me now.

Original issue reported on code.google.com by [email protected] on 9 Feb 2010 at 7:50

Attachments:

• credis_lists_debug.cpp

credis_connect() calls gethostbyname(), which is not reentrant

Line 715 in credis_connect() calls gethostbyname() which is not reentrant.  
gethostbyname_r() is reentrant.  Also note that both of the gethostbyname() 
calls are deprecated, replaced with getnameinfo().

Original issue reported on code.google.com by [email protected] on 15 Sep 2010 at 12:00

add keyspace_hits and keyspace_misses to credis_info command

Add keyspace_hits and keyspace_misses  to credis_info

They keep track of how many times a key is found (useful for caches)

Original issue reported on code.google.com by halkeye on 9 Apr 2012 at 11:34

Attachments:

• credis_info_keyspace.patch

sigfault (bus error) if connection parameters error

some patch in attachment
so, add function credis_isconnect()

see patch file.

Original issue reported on code.google.com by akalend on 10 Jun 2010 at 6:28

Attachments:

• credis.patch

credis does not build on OS X

What steps will reproduce the problem?

Checkout credis and build it

What is the expected output? What do you see instead?

Expected: built succesfully
Actual output:

cc -shared -Wl,-soname,libcredis.so -o libcredis.so credis.o
ld: unknown option: -soname
collect2: ld returned 1 exit status
gmake: *** [libcredis.so] Error 1

What version of the product are you using? On what operating system?

svn r8, OS X 10.6.2, gcc version 4.2.1

Please provide any additional information below.

"""
=== modified file 'Makefile'
--- Makefile    2009-12-03 21:30:48 +0000
+++ Makefile    2010-01-01 06:33:35 +0000
@@ -2,6 +2,13 @@
 LDFLAGS =
 CPPFLAGS =

+OS = $(shell uname -s)
+ifeq ($(OS),Darwin)
+   SHAREDLIB_LINK_OPTIONS=-dynamiclib -Wl,-install_name -Wl,
+else
+   SHAREDLIB_LINK_OPTIONS=-shared -Wl,-soname,
+endif
+
 # targets to build with 'make all'
 TARGETS = credis-test libcredis.a libcredis.so

@@ -14,7 +21,8 @@
    $(AR) -cvq $@ $^

 libcredis.so: credis.o
-   $(CC) -shared -Wl,-soname,$@ -o $@ $^
+   $(CC) $(SHAREDLIB_LINK_OPTIONS)$@ -o $@ $^
+

 credis.o: credis.c credis.h
    $(CC) -c -fPIC $(CFLAGS) $(CPPFLAGS) -o $@ credis.c

"""
See also.
http://lists.ozlabs.org/pipermail/devicetree-discuss/2009-May/000818.html
http://www.finkproject.org/doc/porting/porting.en.html

Original issue reported on code.google.com by [email protected] on 1 Jan 2010 at 6:34

support the append command

the latest redis  provide 'append' command to maintain a string value, here 
is a simple patch to support it

http://code.google.com/p/rediscache/source/detail?r=7

Original issue reported on code.google.com by [email protected] on 9 May 2010 at 8:14

credis-test get wrong "info" result under redis 1.2.3 / cygwin

What steps will reproduce the problem?
1.start redis-server under cygwin
2../credis-test.exe

What is the expected output? What do you see instead?

For "info" result, the expected output is:

$ telnet 127.0.0.1 6379
Trying 127.0.0.1...
Connected to 127.0.0.1.
Escape character is '^]'.
info
$390
redis_version:1.2.3
arch_bits:32
multiplexing_api:select
uptime_in_seconds:744
uptime_in_days:0
connected_clients:1
connected_slaves:0
used_memory:311100
used_memory_human:303.81K
changes_since_last_save:0
bgsave_in_progress:0
last_save_time:1267772385
bgrewriteaof_in_progress:0
total_connections_received:5
total_commands_processed:25561
role:master
db0:keys=6,expires=0

However, the credist-test giving the result:

$ ./credis-test.exe
Testing a number of credis functions. To perform a simplistic
benchmark, run: `./credis-test <num>' where <num> is the numbe
of set-commands to send.



************* misc info ************************************
ping returned: 0
lastsave returned: 1267772385
info returned -97
>redis_version: 1.2.3
>uptime_in_seconds: 1628878521
>uptime_in_days: 2281060
>connected_clients: 0
>connected_slaves: 2280632
>used_memory: 1627480722
>changes_since_last_save: 5922451296
>bgsave_in_progress: 1628867192
>last_save_time: 0
>total_connections_received: 8589934592
>total_commands_processed: 6989927294192242072
>role: 2280680

What version of the product are you using? On what operating system?
credis-0.2 / cygwin

Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 5 Mar 2010 at 7:12

[patch] SETEX, HKEYS, HLEN and HMGET realization

I suggest you realization of some redis commands that I need in my project.
There are SETEX, HKEYS, HLEN and HMGET commands.

Original issue reported on code.google.com by [email protected] on 11 Mar 2011 at 2:51

Attachments:

• credis.diff

Compilation error on FreeBSD 8.1 amd64

What steps will reproduce the problem?
  1. Download source credis-0.2.3.tar.gz & untar
  2. call `gmake`

What is the expected output? What do you see instead?
  gcc don't known where u_int32_t is defined,
  so <sys/types.h> must be included before <netinet/tcp.h>.
  Patch in attachment.

What version of the product are you using? On what operating system?
  credis-0.2.3 on FreeBSD 8.1 amd64

Original issue reported on code.google.com by [email protected] on 21 Nov 2010 at 7:22

Attachments:

• credis.diff

credis_lrange returns CREDIS_ERR_PROTOCOL / -97 if list-key ist not set

What steps will reproduce the problem?
1. credis_lrange(redis,"not_existing",0,-1,&elements);

What is the expected output? What do you see instead?
I would expect nil, null or 0 as return value

What version of the product are you using? On what operating system?
SVN rev. r13

Original issue reported on code.google.com by [email protected] on 8 Feb 2010 at 10:34

Random errors (-95, -97,more data needed) using credis_lrange on large lists

What steps will reproduce the problem?
1. using credis_lrange(redis,"myList",0,-1,&elements); on large lists

What is the expected output? What do you see instead?
It works most of the time but on every 3./4. call the function returns -95, -97 
or displays: more 
data needed. 

What version of the product are you using? On what operating system?
credis 0.1

Please provide any additional information below.
To receive any data I had to set:

#define CR_BUFFER_SIZE 2048
#define CR_MULTIBULK_SIZE 2048

Original issue reported on code.google.com by [email protected] on 8 Feb 2010 at 9:46