Maison >base de données >tutoriel mysql >Redis的Client设计
Redis的client设计如下: /* With multiplexing we need to take per-clinet state. * Clients are taken in a liked list. */ty
Redis的client设计如下:
/* With multiplexing we need to take per-clinet state.
* Clients are taken in a liked list. */
typedef struct redisClient {
int fd;
redisDb *db;
int dictid;
sds querybuf;
robj **argv, **mbargv;
int argc, mbargc;
int bulklen; /* bulk read len. -1 if not in bulk read mode */
int multibulk; /* multi bulk command format active */
list *reply;
int sentlen;
time_t lastinteraction; /* time of the last interaction, used for timeout */
int flags; /* REDIS_CLOSE | REDIS_SLAVE | REDIS_MONITOR */
int slaveseldb; /* slave selected db, if this client is a slave */
int authenticated; /* when requirepass is non-NULL */
int replstate; /* replication state if this is a slave */
int repldbfd; /* replication DB file descriptor */
long repldboff; /* replication DB file offset */
off_t repldbsize; /* replication DB file size */
} redisClient;
在accpet后,开始创建一个client。
代码如下:
static client createClient(void) {
client c = zmalloc(sizeof(struct _client));
char err[ANET_ERR_LEN];
c->fd = anetTcpNonBlockConnect(err,config.hostip,config.hostport);
if (c->fd == ANET_ERR) {
zfree(c);
fprintf(stderr,"Connect: %s\n",err);
return NULL;
}
anetTcpNoDelay(NULL,c->fd);
c->obuf = sdsempty();
c->ibuf = sdsempty();
c->mbulk = -1;
c->readlen = 0;
c->written = 0;
c->totreceived = 0;
c->state = CLIENT_CONNECTING;
aeCreateFileEvent(config.el, c->fd, AE_WRITABLE, writeHandler, c);
config.liveclients++;
listAddNodeTail(config.clients,c);
return c;
}
窃以为,以下代码写的有点那个。
listAddNodeTail(config.clients,c);
这个函数既然命名为createclient就该做create的事,,而把listaddnodetail放到外层函数里面实现。
既然在createclient函数里实现了add功能,又再返回一个c指针到外层函数实现超出链接数判断,显得有些畸形。
同学们,你们觉得呢?
Ubuntu 14.04下Redis安装及简单测试
Redis集群明细文档
Ubuntu 12.10下安装Redis(图文详解)+ Jedis连接Redis
Redis系列-安装部署维护篇
CentOS 6.3安装Redis
Redis安装部署学习笔记
Redis配置文件redis.conf 详解
Redis 的详细介绍:请点这里
Redis 的下载地址:请点这里
本文永久更新链接地址: