Redis—数据结构之sds

Redis是一个Key Value数据库。Redis有5种数据类型：字符串、列表、哈希、集合、有序集合。而字符串的底层实现方法之一就是使用sds。以下描述中请读者注意区分sds是指简单动态字符串这一数据结构(用大写表示)还是sdshdr头部中buf数组的起始地址(用小写表示)。

SDS源码

如下源码所示。

根据要保存的字符串长度选用不同的头部大小，从而节省内存，注意sdshdr5与其他不同，下面会有介绍。

SDS由两部分组成：sds、sdshdr。sds是一个char类型的指针，指向buf数组首元素，buf数组是存储字符串的实际位置；sdshdr是SDS的头部，为SDS加上一个头部的好处就是为了提高某些地方的效率，比如获取buf数组中字符串长度，用O（1）的复杂度从头部就能取得。buf数组是一个空数组，从而使得sdshdr是一个可变长度的结构体，用一个空数组的好处就是分配内存时，只用分配一次，而且头部所占用的内存和sds的内存是连续的，释放时也只用释放一次。

sdshdr结构体中各字段的介绍：len : 已存储的字符串长度；alloc : 能存储的字符串的最大容量，不包括SDS头部和结尾的NULL字符；flags : 标志位，低3位代表了sds头部类型，高5位未用；buf[] : 字符数组，存储字符串；注意sdshdr5没有len和alloc字段，其flags的低3位同样代表头部类型，但高5位代表保存的字符串长度。 __attribute__ ((__packed__)) : 使得编译器不会因为内存对齐而在结构体中填充字节，以保证内存的紧凑，这样sds - 1就可以得到flags字段，进而能够得到其头部类型。如果填充了字节，则就不能得到flags字段。

buf数组尾部隐含有一个'\0'，SDS是以len字段来判断是否到达字符串末尾，而不是以'\0'判断结尾。所以sds存储的字符串中间可以出现'\0'，即sds字符串是二进制安全的。

typedef char *sds;

struct __attribute__ ((__packed__)) sdshdr5 {
    unsigned char flags; /* 3 lsb of type, and 5 msb of string length */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr8 {
    uint8_t len; /* used */
    uint8_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr16 {
    uint16_t len; /* used */
    uint16_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr32 {
    uint32_t len; /* used */
    uint32_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr64 {
    uint64_t len; /* used */
    uint64_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};

 

既然有这么多类型的头部，一定会有类似宏定义之类能够标识头部，的确有，如下所示：

// flags的低三位代表不同类型的sds头部：
#define SDS_TYPE_5  0
#define SDS_TYPE_8  1
#define SDS_TYPE_16 2
#define SDS_TYPE_32 3
#define SDS_TYPE_64 4
#define SDS_TYPE_MASK 7
#define SDS_TYPE_BITS 3

 

SDS操作

因为sds和头部是内存连续的，所以当我们得到了一个sds，只要将它-1就可得到flags字段，减头部大小即可得到头部起始地址。SDS的很多操作就是利用了这一点，从而带来了极大的方便和快速。下面我们介绍几个SDS比较重要的几个操作

1. 获取头部起始地址

将sds减去头部大小即可。非常方便快速。

// 返回一个指向sds头部的起始地址的指针
#define SDS_HDR_VAR(T,s) struct sdshdr##T *sh = (void*)((s)-(sizeof(struct sdshdr##T)));
// 返回sds头部的起始地址
#define SDS_HDR(T,s) ((struct sdshdr##T *)((s)-(sizeof(struct sdshdr##T))))

2. 获取buf数组中sds存储的字符串长度

先后移1位，得到flags字段，再和掩码相与即可得到头部类型。

static inline size_t sdslen(const sds s) {
    unsigned char flags = s[-1]; // 内存空间连续，所以往后移1个字节，便是flags字段
    switch(flags&SDS_TYPE_MASK) { // 和flags低3位相与，得到sds头部类型
        case SDS_TYPE_5:
            return SDS_TYPE_5_LEN(flags);
        case SDS_TYPE_8:
            return SDS_HDR(8,s)->len; // 先移动到sds头部的起始地址，进而可以直接获取len字段的值。下同
        case SDS_TYPE_16:
            return SDS_HDR(16,s)->len;
        case SDS_TYPE_32:
            return SDS_HDR(32,s)->len;
        case SDS_TYPE_64:
            return SDS_HDR(64,s)->len;
    }
    return 0;
}

3. 获取buf数组中剩余可用的内存大小

static inline size_t sdsavail(const sds s) {
    unsigned char flags = s[-1]; // 后移1字节，得到flags字段
    switch(flags&SDS_TYPE_MASK) { // 得到sds头部类型
        case SDS_TYPE_5: {
            return 0;
        }
        case SDS_TYPE_8: {
            SDS_HDR_VAR(8,s);
            return sh->alloc - sh->len; // 总大小减去已使用大小
        }
        case SDS_TYPE_16: {
            SDS_HDR_VAR(16,s);
            return sh->alloc - sh->len;
        }
        case SDS_TYPE_32: {
            SDS_HDR_VAR(32,s);
            return sh->alloc - sh->len;
        }
        case SDS_TYPE_64: {
            SDS_HDR_VAR(64,s);
            return sh->alloc - sh->len;
        }
    }
    return 0;
}

4. 使用字符串初始化一个SDS

注意分配时，程序会自动为buf数组最后一个元素后面添加上'\0'，'\0'对外部完全是透明的，分配内存时自动多分配1个字节保存'\0'，buf数组最后自动添加'\0'。

// sds尾部隐含有一个'\0'；sds是以len字段来判断是否到达字符串末尾
// 所以sds存储的字符串中间可以出现'\0'，即sds字符串是二进制安全的

// 分配一个新sds，buf数组存储内容init
sds sdsnewlen(const void *init, size_t initlen) {
    void *sh;
    sds s;
    char type = sdsReqType(initlen); // 根据长度大小选择合适的sds头部
    /* Empty strings are usually created in order to append. Use type 8
     * since type 5 is not good at this. */
    if (type == SDS_TYPE_5 && initlen == 0) type = SDS_TYPE_8;
    int hdrlen = sdsHdrSize(type); // 获取sds头部大小
    unsigned char *fp; /* flags pointer. */

    // 为sds分配内存，总大小为：头部大小+存储字符串的长度+末尾隐含的空字符大小
    sh = s_malloc(hdrlen+initlen+1); 
    if (!init)
        memset(sh, 0, hdrlen+initlen+1); // 内存初始化为0
    if (sh == NULL) return NULL;
    s = (char*)sh+hdrlen; // buf数组的起始地址
    fp = ((unsigned char*)s)-1; // 指向flags字段
    // 初始化sds头部的len,alloc,flags字段
    switch(type) {
        case SDS_TYPE_5: {
            *fp = type | (initlen << SDS_TYPE_BITS);
            break;
        }
        case SDS_TYPE_8: {
            SDS_HDR_VAR(8,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
        case SDS_TYPE_16: {
            SDS_HDR_VAR(16,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
        case SDS_TYPE_32: {
            SDS_HDR_VAR(32,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
        case SDS_TYPE_64: {
            SDS_HDR_VAR(64,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
    }
    // 初始化buf数组
    if (initlen && init)
        memcpy(s, init, initlen); // 拷贝init到buf数组
    s[initlen] = '\0'; // 添加末尾的空字符
    return s;
}

5. 空间预分配

当需要将SDS的len增加addlen个字节时，如果SDS剩余空间足够，则什么都不用做。如果剩余空间不够，则会分配新的内存空间，并且采用预分配。新长度newlen为原len+addlen，若newlen小于1M，则为SDS分配新的内存大小为2*newlen；若newlen大于等于1M，则SDS分配新的内存大小为newlen  + 1M。

 // 为sds的len字段增加addlen个字节，剩余空间不足时会引起空间重新分配
sds sdsMakeRoomFor(sds s, size_t addlen) {
    void *sh, *newsh;
    size_t avail = sdsavail(s);
    size_t len, newlen;
    char type, oldtype = s[-1] & SDS_TYPE_MASK;
    int hdrlen;

    /* Return ASAP if there is enough space left. */
    if (avail >= addlen) return s; // sds剩余空间足够

    len = sdslen(s);
    sh = (char*)s-sdsHdrSize(oldtype);
    newlen = (len+addlen); // sds剩余空间不够，新的len为len+addlen
    
    // 下面两步实现空间预分配
    if (newlen < SDS_MAX_PREALLOC) // 新长度小于1M，则len设为2*(len+addlen)大小
        newlen *= 2;
    else
        newlen += SDS_MAX_PREALLOC; // 新长度大于1M，则len设为 len+1M 大小

    type = sdsReqType(newlen); // 新len对应的sds头部

    /* Don't use type 5: the user is appending to the string and type 5 is
     * not able to remember empty space, so sdsMakeRoomFor() must be called
     * at every appending operation. */
    if (type == SDS_TYPE_5) type = SDS_TYPE_8;

    hdrlen = sdsHdrSize(type);
    if (oldtype==type) {
        newsh = s_realloc(sh, hdrlen+newlen+1);
        if (newsh == NULL) return NULL;
        s = (char*)newsh+hdrlen;
    } else {
        /* Since the header size changes, need to move the string forward,
         * and can't use realloc */
        newsh = s_malloc(hdrlen+newlen+1);
        if (newsh == NULL) return NULL;
        memcpy((char*)newsh+hdrlen, s, len+1);
        s_free(sh);
        s = (char*)newsh+hdrlen;
        s[-1] = type;
        sdssetlen(s, len);
    }
    sdssetalloc(s, newlen);
    return s;
}

6. 惰性空间释放

当要清空一个SDS时，并不真正释放其内存，而是设置len字段为0即可，这样当之后再次使用到该SDS时，可避免重新分配内存，从而提高效率。

 // 清空sds内容，len字段清为0
 // 但之前的空间并未释放，可避免以后的重新分配内存。实现惰性空间释放
void sdsclear(sds s) {
    sdssetlen(s, 0);
    s[0] = '\0';
}

 

只要理解了sds和sdshdr，其操作函数便很容易理解。剩下的就不一一介绍了，我在阅读过程中也做了部分注释，下面附上源码及注释。SDS共两个文件：sds.h和sds.c

sds.h :

/* SDSLib 2.0 -- A C dynamic strings library
 * 简单动态字符串
 */

#ifndef __SDS_H
#define __SDS_H

#define SDS_MAX_PREALLOC (1024*1024) // 1M，空间预分配使用

#include <sys/types.h>
#include <stdarg.h>
#include <stdint.h>

// 指向存储数据的起始地址
typedef char *sds;

/* Note: sdshdr5 is never used, we just access the flags byte directly.
 * However is here to document the layout of type 5 SDS strings. */

// sds由两部分组成：sds头部(即下面的各种结构体)、真正存储字符串的字符数组
// 这两部分在内存上连续 
// len : 已存储的字符串长度
// alloc : 能存储的字符串的最大容量，不包括sds头部和结尾的NULL字符
// flags : 标志位，最低三位代表了sds头部类型
// buf[] : 字符数组，存储字符串
// __attribute__ ((__packed__)) : 
// 使得编译器不会因为内存对齐而在结构体中填充字节，以保证内存的紧凑，使得下面的s[-1]得到正确的地址
// char buf[] : 初始时不占用内存，而且使得头部内存和存储字符串的内存地址连续。
// sdshdr5比较特殊,flags字段低3位代表sds头部类型，高5位代表已存储的字符串长度
// 分为不同类型的头部，目的是为了存储不同长度的字符串使用不同类型，从而节省内存

 struct __attribute__ ((__packed__)) sdshdr5 {
    unsigned char flags; /* 3 lsb of type, and 5 msb of string length */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr8 {
    uint8_t len; /* used */
    uint8_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr16 {
    uint16_t len; /* used */
    uint16_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr32 {
    uint32_t len; /* used */
    uint32_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};
struct __attribute__ ((__packed__)) sdshdr64 {
    uint64_t len; /* used */
    uint64_t alloc; /* excluding the header and null terminator */
    unsigned char flags; /* 3 lsb of type, 5 unused bits */
    char buf[];
};

// flags的低三位代表不同类型的sds头部：
#define SDS_TYPE_5  0
#define SDS_TYPE_8  1
#define SDS_TYPE_16 2
#define SDS_TYPE_32 3
#define SDS_TYPE_64 4
#define SDS_TYPE_MASK 7
#define SDS_TYPE_BITS 3

// 返回一个指向sds头部的起始地址的指针
#define SDS_HDR_VAR(T,s) struct sdshdr##T *sh = (void*)((s)-(sizeof(struct sdshdr##T)));
// 返回sds头部的起始地址
#define SDS_HDR(T,s) ((struct sdshdr##T *)((s)-(sizeof(struct sdshdr##T))))
// 获取sdshdr5类型的sds存储的字符串长度
#define SDS_TYPE_5_LEN(f) ((f)>>SDS_TYPE_BITS)

// 获取buf数组中sds存储的字符串长度
static inline size_t sdslen(const sds s) {
    unsigned char flags = s[-1]; // 内存空间连续，所以往后移1个字节，便是flags字段
    switch(flags&SDS_TYPE_MASK) { // 和flags低3位相与，得到sds头部类型
        case SDS_TYPE_5:
            return SDS_TYPE_5_LEN(flags);
        case SDS_TYPE_8:
            return SDS_HDR(8,s)->len; // 先移动到sds头部的起始地址，进而可以直接获取len字段的值。下同
        case SDS_TYPE_16:
            return SDS_HDR(16,s)->len;
        case SDS_TYPE_32:
            return SDS_HDR(32,s)->len;
        case SDS_TYPE_64:
            return SDS_HDR(64,s)->len;
    }
    return 0;
}

// 获取buf数组中剩余可用的内存大小
static inline size_t sdsavail(const sds s) {
    unsigned char flags = s[-1]; // 后移1字节，得到flags字段
    switch(flags&SDS_TYPE_MASK) { // 得到sds头部类型
        case SDS_TYPE_5: {
            return 0;
        }
        case SDS_TYPE_8: {
            SDS_HDR_VAR(8,s);
            return sh->alloc - sh->len; // 总大小减去已使用大小
        }
        case SDS_TYPE_16: {
            SDS_HDR_VAR(16,s);
            return sh->alloc - sh->len;
        }
        case SDS_TYPE_32: {
            SDS_HDR_VAR(32,s);
            return sh->alloc - sh->len;
        }
        case SDS_TYPE_64: {
            SDS_HDR_VAR(64,s);
            return sh->alloc - sh->len;
        }
    }
    return 0;
}

// 设置sds头部的len字段
static inline void sdssetlen(sds s, size_t newlen) {
    unsigned char flags = s[-1];
    switch(flags&SDS_TYPE_MASK) {
        case SDS_TYPE_5:
            {
                // 对于sdshdr5，则是设置flags的高5位
                unsigned char *fp = ((unsigned char*)s)-1;
                *fp = SDS_TYPE_5 | (newlen << SDS_TYPE_BITS);
            }
            break;
        case SDS_TYPE_8:
            SDS_HDR(8,s)->len = newlen;
            break;
        case SDS_TYPE_16:
            SDS_HDR(16,s)->len = newlen;
            break;
        case SDS_TYPE_32:
            SDS_HDR(32,s)->len = newlen;
            break;
        case SDS_TYPE_64:
            SDS_HDR(64,s)->len = newlen;
            break;
    }
}

// 将sds头部的len字段增加inc
static inline void sdsinclen(sds s, size_t inc) {
    unsigned char flags = s[-1];
    switch(flags&SDS_TYPE_MASK) {
        case SDS_TYPE_5:
            {
                // 对于sdshdr5，则是设置flags的高5位
                unsigned char *fp = ((unsigned char*)s)-1;
                unsigned char newlen = SDS_TYPE_5_LEN(flags)+inc;
                *fp = SDS_TYPE_5 | (newlen << SDS_TYPE_BITS);
            }
            break;
        case SDS_TYPE_8:
            SDS_HDR(8,s)->len += inc;
            break;
        case SDS_TYPE_16:
            SDS_HDR(16,s)->len += inc;
            break;
        case SDS_TYPE_32:
            SDS_HDR(32,s)->len += inc;
            break;
        case SDS_TYPE_64:
            SDS_HDR(64,s)->len += inc;
            break;
    }
}

/* sdsalloc() = sdsavail() + sdslen() */
// 获取sds的buf数组总的大小
static inline size_t sdsalloc(const sds s) {
    unsigned char flags = s[-1];
    switch(flags&SDS_TYPE_MASK) {
        case SDS_TYPE_5:
            return SDS_TYPE_5_LEN(flags);
        case SDS_TYPE_8:
            return SDS_HDR(8,s)->alloc;
        case SDS_TYPE_16:
            return SDS_HDR(16,s)->alloc;
        case SDS_TYPE_32:
            return SDS_HDR(32,s)->alloc;
        case SDS_TYPE_64:
            return SDS_HDR(64,s)->alloc;
    }
    return 0;
}

// 设置sds的buf数组总的大小
static inline void sdssetalloc(sds s, size_t newlen) {
    unsigned char flags = s[-1];
    switch(flags&SDS_TYPE_MASK) {
        case SDS_TYPE_5:
            /* Nothing to do, this type has no total allocation info. */
            break;
        case SDS_TYPE_8:
            SDS_HDR(8,s)->alloc = newlen;
            break;
        case SDS_TYPE_16:
            SDS_HDR(16,s)->alloc = newlen;
            break;
        case SDS_TYPE_32:
            SDS_HDR(32,s)->alloc = newlen;
            break;
        case SDS_TYPE_64:
            SDS_HDR(64,s)->alloc = newlen;
            break;
    }
}

sds sdsnewlen(const void *init, size_t initlen);
sds sdsnew(const char *init);
sds sdsempty(void);
sds sdsdup(const sds s);
void sdsfree(sds s);
sds sdsgrowzero(sds s, size_t len);
sds sdscatlen(sds s, const void *t, size_t len);
sds sdscat(sds s, const char *t);
sds sdscatsds(sds s, const sds t);
sds sdscpylen(sds s, const char *t, size_t len);
sds sdscpy(sds s, const char *t);

sds sdscatvprintf(sds s, const char *fmt, va_list ap);
#ifdef __GNUC__
sds sdscatprintf(sds s, const char *fmt, ...)
    __attribute__((format(printf, 2, 3)));
#else
sds sdscatprintf(sds s, const char *fmt, ...);
#endif

sds sdscatfmt(sds s, char const *fmt, ...);
sds sdstrim(sds s, const char *cset);
void sdsrange(sds s, int start, int end);
void sdsupdatelen(sds s);
void sdsclear(sds s);
int sdscmp(const sds s1, const sds s2);
sds *sdssplitlen(const char *s, int len, const char *sep, int seplen, int *count);
void sdsfreesplitres(sds *tokens, int count);
void sdstolower(sds s);
void sdstoupper(sds s);
sds sdsfromlonglong(long long value);
sds sdscatrepr(sds s, const char *p, size_t len);
sds *sdssplitargs(const char *line, int *argc);
sds sdsmapchars(sds s, const char *from, const char *to, size_t setlen);
sds sdsjoin(char **argv, int argc, char *sep);
sds sdsjoinsds(sds *argv, int argc, const char *sep, size_t seplen);

/* Low level functions exposed to the user API */
sds sdsMakeRoomFor(sds s, size_t addlen);
void sdsIncrLen(sds s, int incr);
sds sdsRemoveFreeSpace(sds s);
size_t sdsAllocSize(sds s);
void *sdsAllocPtr(sds s);

/* Export the allocator used by SDS to the program using SDS.
 * Sometimes the program SDS is linked to, may use a different set of
 * allocators, but may want to allocate or free things that SDS will
 * respectively free or allocate. */
void *sds_malloc(size_t size);
void *sds_realloc(void *ptr, size_t size);
void sds_free(void *ptr);

#ifdef REDIS_TEST
int sdsTest(int argc, char *argv[]);
#endif

#endif

sds.c :

/* SDSLib 2.0 -- A C dynamic strings library
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <assert.h>
#include "sds.h"
#include "sdsalloc.h"

// 获取type类型的sds对应的头部类型大小
static inline int sdsHdrSize(char type) {
    switch(type&SDS_TYPE_MASK) {
        case SDS_TYPE_5:
            return sizeof(struct sdshdr5);
        case SDS_TYPE_8:
            return sizeof(struct sdshdr8);
        case SDS_TYPE_16:
            return sizeof(struct sdshdr16);
        case SDS_TYPE_32:
            return sizeof(struct sdshdr32);
        case SDS_TYPE_64:
            return sizeof(struct sdshdr64);
    }
    return 0;
}

// 根据不同大小选用不同类型的sds头部
static inline char sdsReqType(size_t string_size) {
    if (string_size < 1<<5) // 0~2^5-1
        return SDS_TYPE_5;
    if (string_size < 1<<8) // 2^5~2^8-1
        return SDS_TYPE_8;
    if (string_size < 1<<16) // 2^8~2^16-1
        return SDS_TYPE_16;
    if (string_size < 1<<32) // 2^16~2^32-1
        return SDS_TYPE_32;
    return SDS_TYPE_64; // 2^32~
}

/* Create a new sds string with the content specified by the 'init' pointer
 * and 'initlen'.
 * If NULL is used for 'init' the string is initialized with zero bytes.
 *
 * The string is always null-termined (all the sds strings are, always) so
 * even if you create an sds string with:
 *
 * mystring = sdsnewlen("abc",3);
 *
 * You can print the string with printf() as there is an implicit \0 at the
 * end of the string. However the string is binary safe and can contain
 * \0 characters in the middle, as the length is stored in the sds header. */
 
// sds尾部隐含有一个'\0'；sds是以len字段来判断是否到达字符串末尾
// 所以sds存储的字符串中间可以出现'\0'，即sds字符串是二进制安全的

// 分配一个新sds，buf数组存储内容init
sds sdsnewlen(const void *init, size_t initlen) {
    void *sh;
    sds s;
    char type = sdsReqType(initlen); // 根据长度大小选择合适的sds头部
    /* Empty strings are usually created in order to append. Use type 8
     * since type 5 is not good at this. */
    if (type == SDS_TYPE_5 && initlen == 0) type = SDS_TYPE_8;
    int hdrlen = sdsHdrSize(type); // 获取sds头部大小
    unsigned char *fp; /* flags pointer. */

    // 为sds分配内存，总大小为：头部大小+存储字符串的长度+末尾隐含的空字符大小
    sh = s_malloc(hdrlen+initlen+1); 
    if (!init)
        memset(sh, 0, hdrlen+initlen+1); // 内存初始化为0
    if (sh == NULL) return NULL;
    s = (char*)sh+hdrlen; // buf数组的起始地址
    fp = ((unsigned char*)s)-1; // 指向flags字段
    // 初始化sds头部的len,alloc,flags字段
    switch(type) {
        case SDS_TYPE_5: {
            *fp = type | (initlen << SDS_TYPE_BITS);
            break;
        }
        case SDS_TYPE_8: {
            SDS_HDR_VAR(8,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
        case SDS_TYPE_16: {
            SDS_HDR_VAR(16,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
        case SDS_TYPE_32: {
            SDS_HDR_VAR(32,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
        case SDS_TYPE_64: {
            SDS_HDR_VAR(64,s);
            sh->len = initlen;
            sh->alloc = initlen;
            *fp = type;
            break;
        }
    }
    // 初始化buf数组
    if (initlen && init)
        memcpy(s, init, initlen); // 拷贝init到buf数组
    s[initlen] = '\0'; // 添加末尾的空字符
    return s;
}

/* Create an empty (zero length) sds string. Even in this case the string
 * always has an implicit null term. */
// 分配一个空的sds
sds sdsempty(void) {
    return sdsnewlen("",0);
}

/* Create a new sds string starting from a null terminated C string. */
// 分配一个新sds，以c字符串初始化其buf数组
sds sdsnew(const char *init) {
    size_t initlen = (init == NULL) ? 0 : strlen(init);
    return sdsnewlen(init, initlen);
}

/* Duplicate an sds string. */
// 复制一个sds
sds sdsdup(const sds s) {
    return sdsnewlen(s, sdslen(s));
}

/* Free an sds string. No operation is performed if 's' is NULL. */
// 释放一个sds
void sdsfree(sds s) {
    if (s == NULL) return;
    s_free((char*)s-sdsHdrSize(s[-1])); // 先移动到sds头部起始地址，再释放
}

/* Set the sds string length to the length as obtained with strlen(), so
 * considering as content only up to the first null term character.
 *
 * This function is useful when the sds string is hacked manually in some
 * way, like in the following example:
 *
 * s = sdsnew("foobar");
 * s[2] = '\0';
 * sdsupdatelen(s);
 * printf("%d\n", sdslen(s));
 *
 * The output will be "2", but if we comment out the call to sdsupdatelen()
 * the output will be "6" as the string was modified but the logical length
 * remains 6 bytes. */
 
 // 以第1个'\0'位置为sds长度，重新设置sds长度
 // 这样做目的是将第一个'\0'后面的部分全部截断
void sdsupdatelen(sds s) {
    int reallen = strlen(s); // 以strlen()来计算长度，到第1个'\0'截止
    sdssetlen(s, reallen);
}

/* Modify an sds string in-place to make it empty (zero length).
 * However all the existing buffer is not discarded but set as free space
 * so that next append operations will not require allocations up to the
 * number of bytes previously available. */
 
 // 清空sds内容，len字段清为0
 // 但之前的空间并未释放，可避免以后的重新分配内存。实现惰性空间释放
void sdsclear(sds s) {
    sdssetlen(s, 0);
    s[0] = '\0';
}

/* Enlarge the free space at the end of the sds string so that the caller
 * is sure that after calling this function can overwrite up to addlen
 * bytes after the end of the string, plus one more byte for nul term.
 *
 * Note: this does not change the *length* of the sds string as returned
 * by sdslen(), but only the free buffer space we have. */
 
 // 为sds的len字段增加addlen个字节，剩余空间不足时会引起空间重新分配
sds sdsMakeRoomFor(sds s, size_t addlen) {
    void *sh, *newsh;
    size_t avail = sdsavail(s);
    size_t len, newlen;
    char type, oldtype = s[-1] & SDS_TYPE_MASK;
    int hdrlen;

    /* Return ASAP if there is enough space left. */
    if (avail >= addlen) return s; // sds剩余空间足够

    len = sdslen(s);
    sh = (char*)s-sdsHdrSize(oldtype);
    newlen = (len+addlen); // sds剩余空间不够，新的len为len+addlen
    
    // 下面两步实现空间预分配
    if (newlen < SDS_MAX_PREALLOC) // 新长度小于1M，则len设为2*(len+addlen)大小
        newlen *= 2;
    else
        newlen += SDS_MAX_PREALLOC; // 新长度大于1M，则len设为 len+1M 大小

    type = sdsReqType(newlen); // 新len对应的sds头部

    /* Don't use type 5: the user is appending to the string and type 5 is
     * not able to remember empty space, so sdsMakeRoomFor() must be called
     * at every appending operation. */
    if (type == SDS_TYPE_5) type = SDS_TYPE_8;

    hdrlen = sdsHdrSize(type);
    if (oldtype==type) {
        newsh = s_realloc(sh, hdrlen+newlen+1);
        if (newsh == NULL) return NULL;
        s = (char*)newsh+hdrlen;
    } else {
        /* Since the header size changes, need to move the string forward,
         * and can't use realloc */
        newsh = s_malloc(hdrlen+newlen+1);
        if (newsh == NULL) return NULL;
        memcpy((char*)newsh+hdrlen, s, len+1);
        s_free(sh);
        s = (char*)newsh+hdrlen;
        s[-1] = type;
        sdssetlen(s, len);
    }
    sdssetalloc(s, newlen);
    return s;
}

/* Reallocate the sds string so that it has no free space at the end. The
 * contained string remains not altered, but next concatenation operations
 * will require a reallocation.
 *
 * After the call, the passed sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call. */
sds sdsRemoveFreeSpace(sds s) {
    void *sh, *newsh;
    char type, oldtype = s[-1] & SDS_TYPE_MASK;
    int hdrlen;
    size_t len = sdslen(s);
    sh = (char*)s-sdsHdrSize(oldtype);

    type = sdsReqType(len);
    hdrlen = sdsHdrSize(type);
    if (oldtype==type) {
        newsh = s_realloc(sh, hdrlen+len+1);
        if (newsh == NULL) return NULL;
        s = (char*)newsh+hdrlen;
    } else {
        newsh = s_malloc(hdrlen+len+1);
        if (newsh == NULL) return NULL;
        memcpy((char*)newsh+hdrlen, s, len+1);
        s_free(sh);
        s = (char*)newsh+hdrlen;
        s[-1] = type;
        sdssetlen(s, len);
    }
    sdssetalloc(s, len);
    return s;
}

/* Return the total size of the allocation of the specifed sds string,
 * including:
 * 1) The sds header before the pointer.
 * 2) The string.
 * 3) The free buffer at the end if any.
 * 4) The implicit null term.
 */
 // 返回sds的总长度：sds头部+alloc字段+结尾的空字符
size_t sdsAllocSize(sds s) {
    size_t alloc = sdsalloc(s);
    return sdsHdrSize(s[-1])+alloc+1;
}

/* Return the pointer of the actual SDS allocation (normally SDS strings
 * are referenced by the start of the string buffer). */
 // 返回sds的起始地址
void *sdsAllocPtr(sds s) {
    return (void*) (s-sdsHdrSize(s[-1]));
}

/* Increment the sds length and decrements the left free space at the
 * end of the string according to 'incr'. Also set the null term
 * in the new end of the string.
 *
 * This function is used in order to fix the string length after the
 * user calls sdsMakeRoomFor(), writes something after the end of
 * the current string, and finally needs to set the new length.
 *
 * Note: it is possible to use a negative increment in order to
 * right-trim the string.
 *
 * Usage example:
 *
 * Using sdsIncrLen() and sdsMakeRoomFor() it is possible to mount the
 * following schema, to cat bytes coming from the kernel to the end of an
 * sds string without copying into an intermediate buffer:
 *
 * oldlen = sdslen(s);
 * s = sdsMakeRoomFor(s, BUFFER_SIZE);
 * nread = read(fd, s+oldlen, BUFFER_SIZE);
 * ... check for nread <= 0 and handle it ...
 * sdsIncrLen(s, nread);
 */
 
 // 更新sds的len字段，incr可以为负，表示减少len
void sdsIncrLen(sds s, int incr) {
    unsigned char flags = s[-1];
    size_t len;
    switch(flags&SDS_TYPE_MASK) {
        case SDS_TYPE_5: {
            unsigned char *fp = ((unsigned char*)s)-1;
            unsigned char oldlen = SDS_TYPE_5_LEN(flags);
            assert((incr > 0 && oldlen+incr < 32) || (incr < 0 && oldlen >= (unsigned int)(-incr)));
            *fp = SDS_TYPE_5 | ((oldlen+incr) << SDS_TYPE_BITS);
            len = oldlen+incr;
            break;
        }
        case SDS_TYPE_8: {
            SDS_HDR_VAR(8,s);
            assert((incr >= 0 && sh->alloc-sh->len >= incr) || (incr < 0 && sh->len >= (unsigned int)(-incr)));
            len = (sh->len += incr);
            break;
        }
        case SDS_TYPE_16: {
            SDS_HDR_VAR(16,s);
            assert((incr >= 0 && sh->alloc-sh->len >= incr) || (incr < 0 && sh->len >= (unsigned int)(-incr)));
            len = (sh->len += incr);
            break;
        }
        case SDS_TYPE_32: {
            SDS_HDR_VAR(32,s);
            assert((incr >= 0 && sh->alloc-sh->len >= (unsigned int)incr) || (incr < 0 && sh->len >= (unsigned int)(-incr)));
            len = (sh->len += incr);
            break;
        }
        case SDS_TYPE_64: {
            SDS_HDR_VAR(64,s);
            assert((incr >= 0 && sh->alloc-sh->len >= (uint64_t)incr) || (incr < 0 && sh->len >= (uint64_t)(-incr)));
            len = (sh->len += incr);
            break;
        }
        default: len = 0; /* Just to avoid compilation warnings. */
    }
    s[len] = '\0';
}

/* Grow the sds to have the specified length. Bytes that were not part of
 * the original length of the sds will be set to zero.
 *
 * if the specified length is smaller than the current length, no operation
 * is performed. */
 
 // 为sds分配len大小的空间，len小于目前的len字段，则什么都不做
sds sdsgrowzero(sds s, size_t len) {
    size_t curlen = sdslen(s);

    if (len <= curlen) return s;
    s = sdsMakeRoomFor(s,len-curlen);
    if (s == NULL) return NULL;

    /* Make sure added region doesn't contain garbage */
    memset(s+curlen,0,(len-curlen+1)); /* also set trailing \0 byte */
    sdssetlen(s, len);
    return s;
}

/* Append the specified binary-safe string pointed by 't' of 'len' bytes to the
 * end of the specified sds string 's'.
 *
 * After the call, the passed sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call. */
 
 // 在原sds后追加内容*t
sds sdscatlen(sds s, const void *t, size_t len) {
    size_t curlen = sdslen(s);

    s = sdsMakeRoomFor(s,len);
    if (s == NULL) return NULL;
    memcpy(s+curlen, t, len);
    sdssetlen(s, curlen+len);
    s[curlen+len] = '\0';
    return s;
}

/* Append the specified null termianted C string to the sds string 's'.
 *
 * After the call, the passed sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call. */
 
 // 在原sds后追加字符串
sds sdscat(sds s, const char *t) {
    return sdscatlen(s, t, strlen(t));
}

/* Append the specified sds 't' to the existing sds 's'.
 *
 * After the call, the modified sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call. */
 
  // 在原sds后追加sds
sds sdscatsds(sds s, const sds t) {
    return sdscatlen(s, t, sdslen(t));
}

/* Destructively modify the sds string 's' to hold the specified binary
 * safe string pointed by 't' of length 'len' bytes. */
 
 // 将长度为len的字符串*t拷贝到sds
sds sdscpylen(sds s, const char *t, size_t len) {
    if (sdsalloc(s) < len) {
        s = sdsMakeRoomFor(s,len-sdslen(s));
        if (s == NULL) return NULL;
    }
    memcpy(s, t, len);
    s[len] = '\0';
    sdssetlen(s, len);
    return s;
}

/* Like sdscpylen() but 't' must be a null-termined string so that the length
 * of the string is obtained with strlen(). */
  // 将字符串*t拷贝到sds
sds sdscpy(sds s, const char *t) {
    return sdscpylen(s, t, strlen(t));
}

/* Helper for sdscatlonglong() doing the actual number -> string
 * conversion. 's' must point to a string with room for at least
 * SDS_LLSTR_SIZE bytes.
 *
 * The function returns the length of the null-terminated string
 * representation stored at 's'. */
#define SDS_LLSTR_SIZE 21

// 将一个long long类型的值转换为一个字符串
int sdsll2str(char *s, long long value) {
    char *p, aux;
    unsigned long long v;
    size_t l;

    /* Generate the string representation, this method produces
     * an reversed string. */
    v = (value < 0) ? -value : value;
    p = s;
    do {
        *p++ = '0'+(v%10);
        v /= 10;
    } while(v);
    if (value < 0) *p++ = '-';

    /* Compute length and add null term. */
    l = p-s;
    *p = '\0';

    /* Reverse the string. */
    p--;
    while(s < p) {
        aux = *s;
        *s = *p;
        *p = aux;
        s++;
        p--;
    }
    return l;
}

/* Identical sdsll2str(), but for unsigned long long type. */

// 将一个unsigned long long类型的值转换为一个字符串
int sdsull2str(char *s, unsigned long long v) {
    char *p, aux;
    size_t l;

    /* Generate the string representation, this method produces
     * an reversed string. */
    p = s;
    do {
        *p++ = '0'+(v%10);
        v /= 10;
    } while(v);

    /* Compute length and add null term. */
    l = p-s;
    *p = '\0';

    /* Reverse the string. */
    p--;
    while(s < p) {
        aux = *s;
        *s = *p;
        *p = aux;
        s++;
        p--;
    }
    return l;
}

/* Create an sds string from a long long value. It is much faster than:
 *
 * sdscatprintf(sdsempty(),"%lld\n", value);
 */
 
 // 将一个long long类型的值以字符串的形式保存为sds
sds sdsfromlonglong(long long value) {
    char buf[SDS_LLSTR_SIZE];
    int len = sdsll2str(buf,value);

    return sdsnewlen(buf,len);
}

/* Like sdscatprintf() but gets va_list instead of being variadic. */

// 打印不定参数的内容到sds
sds sdscatvprintf(sds s, const char *fmt, va_list ap) {
    va_list cpy;
    char staticbuf[1024], *buf = staticbuf, *t;
    size_t buflen = strlen(fmt)*2;

    /* We try to start using a static buffer for speed.
     * If not possible we revert to heap allocation. */
    if (buflen > sizeof(staticbuf)) {
        buf = s_malloc(buflen);
        if (buf == NULL) return NULL;
    } else {
        buflen = sizeof(staticbuf);
    }

    /* Try with buffers two times bigger every time we fail to
     * fit the string in the current buffer size. */
    while(1) {
        buf[buflen-2] = '\0';
        va_copy(cpy,ap);
        vsnprintf(buf, buflen, fmt, cpy);
        va_end(cpy);
        if (buf[buflen-2] != '\0') {
            if (buf != staticbuf) s_free(buf);
            buflen *= 2;
            buf = s_malloc(buflen);
            if (buf == NULL) return NULL;
            continue;
        }
        break;
    }

    /* Finally concat the obtained string to the SDS string and return it. */
    t = sdscat(s, buf);
    if (buf != staticbuf) s_free(buf);
    return t;
}

/* Append to the sds string 's' a string obtained using printf-alike format
 * specifier.
 *
 * After the call, the modified sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call.
 *
 * Example:
 *
 * s = sdsnew("Sum is: ");
 * s = sdscatprintf(s,"%d+%d = %d",a,b,a+b).
 *
 * Often you need to create a string from scratch with the printf-alike
 * format. When this is the need, just use sdsempty() as the target string:
 *
 * s = sdscatprintf(sdsempty(), "... your format ...", args);
 */
 
// 打印不定参数的内容到sds
sds sdscatprintf(sds s, const char *fmt, ...) {
    va_list ap;
    char *t;
    va_start(ap, fmt);
    t = sdscatvprintf(s,fmt,ap);
    va_end(ap);
    return t;
}

/* This function is similar to sdscatprintf, but much faster as it does
 * not rely on sprintf() family functions implemented by the libc that
 * are often very slow. Moreover directly handling the sds string as
 * new data is concatenated provides a performance improvement.
 *
 * However this function only handles an incompatible subset of printf-alike
 * format specifiers:
 *
 * %s - C String
 * %S - SDS string
 * %i - signed int
 * %I - 64 bit signed integer (long long, int64_t)
 * %u - unsigned int
 * %U - 64 bit unsigned integer (unsigned long long, uint64_t)
 * %% - Verbatim "%" character.
 */
 
 // 格式化输出字符串到sds
sds sdscatfmt(sds s, char const *fmt, ...) {
    size_t initlen = sdslen(s);
    const char *f = fmt;
    int i;
    va_list ap;

    va_start(ap,fmt);
    f = fmt;    /* Next format specifier byte to process. */
    i = initlen; /* Position of the next byte to write to dest str. */
    while(*f) {
        char next, *str;
        size_t l;
        long long num;
        unsigned long long unum;

        /* Make sure there is always space for at least 1 char. */
        if (sdsavail(s)==0) {
            s = sdsMakeRoomFor(s,1);
        }

        switch(*f) {
        case '%':
            next = *(f+1);
            f++;
            switch(next) {
            case 's':
            case 'S':
                str = va_arg(ap,char*);
                l = (next == 's') ? strlen(str) : sdslen(str);
                if (sdsavail(s) < l) {
                    s = sdsMakeRoomFor(s,l);
                }
                memcpy(s+i,str,l);
                sdsinclen(s,l);
                i += l;
                break;
            case 'i':
            case 'I':
                if (next == 'i')
                    num = va_arg(ap,int);
                else
                    num = va_arg(ap,long long);
                {
                    char buf[SDS_LLSTR_SIZE];
                    l = sdsll2str(buf,num);
                    if (sdsavail(s) < l) {
                        s = sdsMakeRoomFor(s,l);
                    }
                    memcpy(s+i,buf,l);
                    sdsinclen(s,l);
                    i += l;
                }
                break;
            case 'u':
            case 'U':
                if (next == 'u')
                    unum = va_arg(ap,unsigned int);
                else
                    unum = va_arg(ap,unsigned long long);
                {
                    char buf[SDS_LLSTR_SIZE];
                    l = sdsull2str(buf,unum);
                    if (sdsavail(s) < l) {
                        s = sdsMakeRoomFor(s,l);
                    }
                    memcpy(s+i,buf,l);
                    sdsinclen(s,l);
                    i += l;
                }
                break;
            default: /* Handle %% and generally %<unknown>. */
                s[i++] = next;
                sdsinclen(s,1);
                break;
            }
            break;
        default:
            s[i++] = *f;
            sdsinclen(s,1);
            break;
        }
        f++;
    }
    va_end(ap);

    /* Add null-term */
    s[i] = '\0';
    return s;
}

/* Remove the part of the string from left and from right composed just of
 * contiguous characters found in 'cset', that is a null terminted C string.
 *
 * After the call, the modified sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call.
 *
 * Example:
 *
 * s = sdsnew("AA...AA.a.aa.aHelloWorld     :::");
 * s = sdstrim(s,"Aa. :");
 * printf("%s\n", s);
 *
 * Output will be just "Hello World".
 */
sds sdstrim(sds s, const char *cset) {
    char *start, *end, *sp, *ep;
    size_t len;

    sp = start = s;
    ep = end = s+sdslen(s)-1;
    while(sp <= end && strchr(cset, *sp)) sp++;
    while(ep > sp && strchr(cset, *ep)) ep--;
    len = (sp > ep) ? 0 : ((ep-sp)+1);
    if (s != sp) memmove(s, sp, len);
    s[len] = '\0';
    sdssetlen(s,len);
    return s;
}

/* Turn the string into a smaller (or equal) string containing only the
 * substring specified by the 'start' and 'end' indexes.
 *
 * start and end can be negative, where -1 means the last character of the
 * string, -2 the penultimate character, and so forth.
 *
 * The interval is inclusive, so the start and end characters will be part
 * of the resulting string.
 *
 * The string is modified in-place.
 *
 * Example:
 *
 * s = sdsnew("Hello World");
 * sdsrange(s,1,-1); => "ello World"
 */
void sdsrange(sds s, int start, int end) {
    size_t newlen, len = sdslen(s);

    if (len == 0) return;
    if (start < 0) {
        start = len+start;
        if (start < 0) start = 0;
    }
    if (end < 0) {
        end = len+end;
        if (end < 0) end = 0;
    }
    newlen = (start > end) ? 0 : (end-start)+1;
    if (newlen != 0) {
        if (start >= (signed)len) {
            newlen = 0;
        } else if (end >= (signed)len) {
            end = len-1;
            newlen = (start > end) ? 0 : (end-start)+1;
        }
    } else {
        start = 0;
    }
    if (start && newlen) memmove(s, s+start, newlen);
    s[newlen] = 0;
    sdssetlen(s,newlen);
}

/* Apply tolower() to every character of the sds string 's'. */
void sdstolower(sds s) {
    int len = sdslen(s), j;

    for (j = 0; j < len; j++) s[j] = tolower(s[j]);
}

/* Apply toupper() to every character of the sds string 's'. */
void sdstoupper(sds s) {
    int len = sdslen(s), j;

    for (j = 0; j < len; j++) s[j] = toupper(s[j]);
}

/* Compare two sds strings s1 and s2 with memcmp().
 *
 * Return value:
 *
 *     positive if s1 > s2.
 *     negative if s1 < s2.
 *     0 if s1 and s2 are exactly the same binary string.
 *
 * If two strings share exactly the same prefix, but one of the two has
 * additional characters, the longer string is considered to be greater than
 * the smaller one. */
int sdscmp(const sds s1, const sds s2) {
    size_t l1, l2, minlen;
    int cmp;

    l1 = sdslen(s1);
    l2 = sdslen(s2);
    minlen = (l1 < l2) ? l1 : l2;
    cmp = memcmp(s1,s2,minlen);
    if (cmp == 0) return l1-l2;
    return cmp;
}

/* Split 's' with separator in 'sep'. An array
 * of sds strings is returned. *count will be set
 * by reference to the number of tokens returned.
 *
 * On out of memory, zero length string, zero length
 * separator, NULL is returned.
 *
 * Note that 'sep' is able to split a string using
 * a multi-character separator. For example
 * sdssplit("foo_-_bar","_-_"); will return two
 * elements "foo" and "bar".
 *
 * This version of the function is binary-safe but
 * requires length arguments. sdssplit() is just the
 * same function but for zero-terminated strings.
 */
sds *sdssplitlen(const char *s, int len, const char *sep, int seplen, int *count) {
    int elements = 0, slots = 5, start = 0, j;
    sds *tokens;

    if (seplen < 1 || len < 0) return NULL;

    tokens = s_malloc(sizeof(sds)*slots);
    if (tokens == NULL) return NULL;

    if (len == 0) {
        *count = 0;
        return tokens;
    }
    for (j = 0; j < (len-(seplen-1)); j++) {
        /* make sure there is room for the next element and the final one */
        if (slots < elements+2) {
            sds *newtokens;

            slots *= 2;
            newtokens = s_realloc(tokens,sizeof(sds)*slots);
            if (newtokens == NULL) goto cleanup;
            tokens = newtokens;
        }
        /* search the separator */
        if ((seplen == 1 && *(s+j) == sep[0]) || (memcmp(s+j,sep,seplen) == 0)) {
            tokens[elements] = sdsnewlen(s+start,j-start);
            if (tokens[elements] == NULL) goto cleanup;
            elements++;
            start = j+seplen;
            j = j+seplen-1; /* skip the separator */
        }
    }
    /* Add the final element. We are sure there is room in the tokens array. */
    tokens[elements] = sdsnewlen(s+start,len-start);
    if (tokens[elements] == NULL) goto cleanup;
    elements++;
    *count = elements;
    return tokens;

cleanup:
    {
        int i;
        for (i = 0; i < elements; i++) sdsfree(tokens[i]);
        s_free(tokens);
        *count = 0;
        return NULL;
    }
}

/* Free the result returned by sdssplitlen(), or do nothing if 'tokens' is NULL. */
void sdsfreesplitres(sds *tokens, int count) {
    if (!tokens) return;
    while(count--)
        sdsfree(tokens[count]);
    s_free(tokens);
}

/* Append to the sds string "s" an escaped string representation where
 * all the non-printable characters (tested with isprint()) are turned into
 * escapes in the form "\n\r\a...." or "\x<hex-number>".
 *
 * After the call, the modified sds string is no longer valid and all the
 * references must be substituted with the new pointer returned by the call. */
sds sdscatrepr(sds s, const char *p, size_t len) {
    s = sdscatlen(s,"\"",1);
    while(len--) {
        switch(*p) {
        case '\\':
        case '"':
            s = sdscatprintf(s,"\\%c",*p);
            break;
        case '\n': s = sdscatlen(s,"\\n",2); break;
        case '\r': s = sdscatlen(s,"\\r",2); break;
        case '\t': s = sdscatlen(s,"\\t",2); break;
        case '\a': s = sdscatlen(s,"\\a",2); break;
        case '\b': s = sdscatlen(s,"\\b",2); break;
        default:
            if (isprint(*p))
                s = sdscatprintf(s,"%c",*p);
            else
                s = sdscatprintf(s,"\\x%02x",(unsigned char)*p);
            break;
        }
        p++;
    }
    return sdscatlen(s,"\"",1);
}

/* Helper function for sdssplitargs() that returns non zero if 'c'
 * is a valid hex digit. */
int is_hex_digit(char c) {
    return (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') ||
           (c >= 'A' && c <= 'F');
}

/* Helper function for sdssplitargs() that converts a hex digit into an
 * integer from 0 to 15 */
int hex_digit_to_int(char c) {
    switch(c) {
    case '0': return 0;
    case '1': return 1;
    case '2': return 2;
    case '3': return 3;
    case '4': return 4;
    case '5': return 5;
    case '6': return 6;
    case '7': return 7;
    case '8': return 8;
    case '9': return 9;
    case 'a': case 'A': return 10;
    case 'b': case 'B': return 11;
    case 'c': case 'C': return 12;
    case 'd': case 'D': return 13;
    case 'e': case 'E': return 14;
    case 'f': case 'F': return 15;
    default: return 0;
    }
}

/* Split a line into arguments, where every argument can be in the
 * following programming-language REPL-alike form:
 *
 * foo bar "newline are supported\n" and "\xff\x00otherstuff"
 *
 * The number of arguments is stored into *argc, and an array
 * of sds is returned.
 *
 * The caller should free the resulting array of sds strings with
 * sdsfreesplitres().
 *
 * Note that sdscatrepr() is able to convert back a string into
 * a quoted string in the same format sdssplitargs() is able to parse.
 *
 * The function returns the allocated tokens on success, even when the
 * input string is empty, or NULL if the input contains unbalanced
 * quotes or closed quotes followed by non space characters
 * as in: "foo"bar or "foo'
 */
sds *sdssplitargs(const char *line, int *argc) {
    const char *p = line;
    char *current = NULL;
    char **vector = NULL;

    *argc = 0;
    while(1) {
        /* skip blanks */
        while(*p && isspace(*p)) p++;
        if (*p) {
            /* get a token */
            int inq=0;  /* set to 1 if we are in "quotes" */
            int insq=0; /* set to 1 if we are in 'single quotes' */
            int done=0;

            if (current == NULL) current = sdsempty();
            while(!done) {
                if (inq) {
                    if (*p == '\\' && *(p+1) == 'x' &&
                                             is_hex_digit(*(p+2)) &&
                                             is_hex_digit(*(p+3)))
                    {
                        unsigned char byte;

                        byte = (hex_digit_to_int(*(p+2))*16)+
                                hex_digit_to_int(*(p+3));
                        current = sdscatlen(current,(char*)&byte,1);
                        p += 3;
                    } else if (*p == '\\' && *(p+1)) {
                        char c;

                        p++;
                        switch(*p) {
                        case 'n': c = '\n'; break;
                        case 'r': c = '\r'; break;
                        case 't': c = '\t'; break;
                        case 'b': c = '\b'; break;
                        case 'a': c = '\a'; break;
                        default: c = *p; break;
                        }
                        current = sdscatlen(current,&c,1);
                    } else if (*p == '"') {
                        /* closing quote must be followed by a space or
                         * nothing at all. */
                        if (*(p+1) && !isspace(*(p+1))) goto err;
                        done=1;
                    } else if (!*p) {
                        /* unterminated quotes */
                        goto err;
                    } else {
                        current = sdscatlen(current,p,1);
                    }
                } else if (insq) {
                    if (*p == '\\' && *(p+1) == '\'') {
                        p++;
                        current = sdscatlen(current,"'",1);
                    } else if (*p == '\'') {
                        /* closing quote must be followed by a space or
                         * nothing at all. */
                        if (*(p+1) && !isspace(*(p+1))) goto err;
                        done=1;
                    } else if (!*p) {
                        /* unterminated quotes */
                        goto err;
                    } else {
                        current = sdscatlen(current,p,1);
                    }
                } else {
                    switch(*p) {
                    case ' ':
                    case '\n':
                    case '\r':
                    case '\t':
                    case '\0':
                        done=1;
                        break;
                    case '"':
                        inq=1;
                        break;
                    case '\'':
                        insq=1;
                        break;
                    default:
                        current = sdscatlen(current,p,1);
                        break;
                    }
                }
                if (*p) p++;
            }
            /* add the token to the vector */
            vector = s_realloc(vector,((*argc)+1)*sizeof(char*));
            vector[*argc] = current;
            (*argc)++;
            current = NULL;
        } else {
            /* Even on empty input string return something not NULL. */
            if (vector == NULL) vector = s_malloc(sizeof(void*));
            return vector;
        }
    }

err:
    while((*argc)--)
        sdsfree(vector[*argc]);
    s_free(vector);
    if (current) sdsfree(current);
    *argc = 0;
    return NULL;
}

/* Modify the string substituting all the occurrences of the set of
 * characters specified in the 'from' string to the corresponding character
 * in the 'to' array.
 *
 * For instance: sdsmapchars(mystring, "ho", "01", 2)
 * will have the effect of turning the string "hello" into "0ell1".
 *
 * The function returns the sds string pointer, that is always the same
 * as the input pointer since no resize is needed. */
sds sdsmapchars(sds s, const char *from, const char *to, size_t setlen) {
    size_t j, i, l = sdslen(s);

    for (j = 0; j < l; j++) {
        for (i = 0; i < setlen; i++) {
            if (s[j] == from[i]) {
                s[j] = to[i];
                break;
            }
        }
    }
    return s;
}

/* Join an array of C strings using the specified separator (also a C string).
 * Returns the result as an sds string. */
sds sdsjoin(char **argv, int argc, char *sep) {
    sds join = sdsempty();
    int j;

    for (j = 0; j < argc; j++) {
        join = sdscat(join, argv[j]);
        if (j != argc-1) join = sdscat(join,sep);
    }
    return join;
}

/* Like sdsjoin, but joins an array of SDS strings. */
sds sdsjoinsds(sds *argv, int argc, const char *sep, size_t seplen) {
    sds join = sdsempty();
    int j;

    for (j = 0; j < argc; j++) {
        join = sdscatsds(join, argv[j]);
        if (j != argc-1) join = sdscatlen(join,sep,seplen);
    }
    return join;
}

/* Wrappers to the allocators used by SDS. Note that SDS will actually
 * just use the macros defined into sdsalloc.h in order to avoid to pay
 * the overhead of function calls. Here we define these wrappers only for
 * the programs SDS is linked to, if they want to touch the SDS internals
 * even if they use a different allocator. */
void *sds_malloc(size_t size) { return s_malloc(size); }
void *sds_realloc(void *ptr, size_t size) { return s_realloc(ptr,size); }
void sds_free(void *ptr) { s_free(ptr); }

#if defined(SDS_TEST_MAIN)
#include <stdio.h>
#include "testhelp.h"
#include "limits.h"

#define UNUSED(x) (void)(x)
int sdsTest(void) {
    {
        sds x = sdsnew("foo"), y;

        test_cond("Create a string and obtain the length",
            sdslen(x) == 3 && memcmp(x,"foo\0",4) == 0)

        sdsfree(x);
        x = sdsnewlen("foo",2);
        test_cond("Create a string with specified length",
            sdslen(x) == 2 && memcmp(x,"fo\0",3) == 0)

        x = sdscat(x,"bar");
        test_cond("Strings concatenation",
            sdslen(x) == 5 && memcmp(x,"fobar\0",6) == 0);

        x = sdscpy(x,"a");
        test_cond("sdscpy() against an originally longer string",
            sdslen(x) == 1 && memcmp(x,"a\0",2) == 0)

        x = sdscpy(x,"xyzxxxxxxxxxxyyyyyyyyyykkkkkkkkkk");
        test_cond("sdscpy() against an originally shorter string",
            sdslen(x) == 33 &&
            memcmp(x,"xyzxxxxxxxxxxyyyyyyyyyykkkkkkkkkk\0",33) == 0)

        sdsfree(x);
        x = sdscatprintf(sdsempty(),"%d",123);
        test_cond("sdscatprintf() seems working in the base case",
            sdslen(x) == 3 && memcmp(x,"123\0",4) == 0)

        sdsfree(x);
        x = sdsnew("--");
        x = sdscatfmt(x, "Hello %s World %I,%I--", "Hi!", LLONG_MIN,LLONG_MAX);
        test_cond("sdscatfmt() seems working in the base case",
            sdslen(x) == 60 &&
            memcmp(x,"--Hello Hi! World -9223372036854775808,"
                     "9223372036854775807--",60) == 0)
        printf("[%s]\n",x);

        sdsfree(x);
        x = sdsnew("--");
        x = sdscatfmt(x, "%u,%U--", UINT_MAX, ULLONG_MAX);
        test_cond("sdscatfmt() seems working with unsigned numbers",
            sdslen(x) == 35 &&
            memcmp(x,"--4294967295,18446744073709551615--",35) == 0)

        sdsfree(x);
        x = sdsnew(" x ");
        sdstrim(x," x");
        test_cond("sdstrim() works when all chars match",
            sdslen(x) == 0)

        sdsfree(x);
        x = sdsnew(" x ");
        sdstrim(x," ");
        test_cond("sdstrim() works when a single char remains",
            sdslen(x) == 1 && x[0] == 'x')

        sdsfree(x);
        x = sdsnew("xxciaoyyy");
        sdstrim(x,"xy");
        test_cond("sdstrim() correctly trims characters",
            sdslen(x) == 4 && memcmp(x,"ciao\0",5) == 0)

        y = sdsdup(x);
        sdsrange(y,1,1);
        test_cond("sdsrange(...,1,1)",
            sdslen(y) == 1 && memcmp(y,"i\0",2) == 0)

        sdsfree(y);
        y = sdsdup(x);
        sdsrange(y,1,-1);
        test_cond("sdsrange(...,1,-1)",
            sdslen(y) == 3 && memcmp(y,"iao\0",4) == 0)

        sdsfree(y);
        y = sdsdup(x);
        sdsrange(y,-2,-1);
        test_cond("sdsrange(...,-2,-1)",
            sdslen(y) == 2 && memcmp(y,"ao\0",3) == 0)

        sdsfree(y);
        y = sdsdup(x);
        sdsrange(y,2,1);
        test_cond("sdsrange(...,2,1)",
            sdslen(y) == 0 && memcmp(y,"\0",1) == 0)

        sdsfree(y);
        y = sdsdup(x);
        sdsrange(y,1,100);
        test_cond("sdsrange(...,1,100)",
            sdslen(y) == 3 && memcmp(y,"iao\0",4) == 0)

        sdsfree(y);
        y = sdsdup(x);
        sdsrange(y,100,100);
        test_cond("sdsrange(...,100,100)",
            sdslen(y) == 0 && memcmp(y,"\0",1) == 0)

        sdsfree(y);
        sdsfree(x);
        x = sdsnew("foo");
        y = sdsnew("foa");
        test_cond("sdscmp(foo,foa)", sdscmp(x,y) > 0)

        sdsfree(y);
        sdsfree(x);
        x = sdsnew("bar");
        y = sdsnew("bar");
        test_cond("sdscmp(bar,bar)", sdscmp(x,y) == 0)

        sdsfree(y);
        sdsfree(x);
        x = sdsnew("aar");
        y = sdsnew("bar");
        test_cond("sdscmp(bar,bar)", sdscmp(x,y) < 0)

        sdsfree(y);
        sdsfree(x);
        x = sdsnewlen("\a\n\0foo\r",7);
        y = sdscatrepr(sdsempty(),x,sdslen(x));
        test_cond("sdscatrepr(...data...)",
            memcmp(y,"\"\\a\\n\\x00foo\\r\"",15) == 0)

        {
            unsigned int oldfree;
            char *p;
            int step = 10, j, i;

            sdsfree(x);
            sdsfree(y);
            x = sdsnew("0");
            test_cond("sdsnew() free/len buffers", sdslen(x) == 1 && sdsavail(x) == 0);

            /* Run the test a few times in order to hit the first two
             * SDS header types. */
            for (i = 0; i < 10; i++) {
                int oldlen = sdslen(x);
                x = sdsMakeRoomFor(x,step);
                int type = x[-1]&SDS_TYPE_MASK;

                test_cond("sdsMakeRoomFor() len", sdslen(x) == oldlen);
                if (type != SDS_TYPE_5) {
                    test_cond("sdsMakeRoomFor() free", sdsavail(x) >= step);
                    oldfree = sdsavail(x);
                }
                p = x+oldlen;
                for (j = 0; j < step; j++) {
                    p[j] = 'A'+j;
                }
                sdsIncrLen(x,step);
            }
            test_cond("sdsMakeRoomFor() content",
                memcmp("0ABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJABCDEFGHIJ",x,101) == 0);
            test_cond("sdsMakeRoomFor() final length",sdslen(x)==101);

            sdsfree(x);
        }
    }
    test_report()
    return 0;
}
#endif

#ifdef SDS_TEST_MAIN
int main(void) {
    return sdsTest();
}
#endif