Flex开发解析器示例(1)

  

通用单据处理支持主表修改,对于复杂数据类型,当仅主表数据的部分数据项改变时只同步这些已变化的字段,减少不必要的开销。
接口表记录变更的方式是记录每个变更字段的当前值。
每个字段的变更描述格式为：字段名=[值长度,属性,值]。
多个字段之间用逗号(",")分隔。

此格式可以避免转义问题(如值内容中出现"],"的情况).
属性目前用于区分是否是base64编码,这可用于binary类型字段的支持，可扩展。

字段类型在同步到目标库时根据数据库的结构信息决定,不需要在此指定。

 

以下为用Flex开发的解析变更字段表达式的解析器代码。

 

/* name: fld_list.y */ %option noyywrap %x S_FIELDNAME %x S_EQUEL %x S_DATA_LEN %x S_DATA_LEN2 %x S_DATA_PROP %x S_DATA_PROP2 %x S_FIELDVAL %x S_END %{ #include <string.h> #include <stdlib.h> char *fld_name; /*字段名称 */ unsigned int data_len = 0; /* 字段值长度*/ int data_prop = 0; /*字段属性*/ char *data_buffer; /* 字段值 */ #define MAX_FIELD_NAME_LEN  64 struct stFieldInfo {  char *name;  unsigned int len;  int prop;  char *buffer;    struct stFieldInfo *next; }; struct stFieldInfo  *field_list = 0; struct stFieldInfo  *last_field = 0; int add_field(struct stFieldInfo *fld); %} %% [A-Za-z\_]+([A-Za-z\_]|[0-9])* {  fld_name = strdup(yytext);  BEGIN S_FIELDNAME; } <S_FIELDNAME>([" "\t])*"="([" "\t])*\[([" "\t])* {  BEGIN S_EQUEL; } <S_EQUEL>[0-9]* {  data_len = atoi(yytext);  BEGIN S_DATA_LEN; } <S_DATA_LEN>([" "\t])*","([" "\t])* {  BEGIN S_DATA_LEN2;  } <S_DATA_LEN2>[0-9]* {  data_prop = atoi(yytext);  BEGIN S_DATA_PROP; } <S_DATA_PROP>([" "\t])*","([" "\t])* {  BEGIN S_DATA_PROP2;  } <S_DATA_PROP2>. {  int c;  int len = 0;  char *p;  data_buffer = malloc(data_len+1);  data_buffer[0] = yytext[0];  p = data_buffer;  p++;  len++;  while(len<data_len&&(c=input())&&c!=EOF) {   *p = c;   p++;   len++;  }  if (c==EOF||len<data_len)   yyterminate();  data_buffer[len] = '\0';  BEGIN S_FIELDVAL; } <S_FIELDVAL>([" "\t])*\]([" "\t])*[,]? {  struct stFieldInfo *fld = malloc(sizeof(struct stFieldInfo));  fld->name = fld_name;  fld->len = data_len;  fld->prop = data_prop;  fld->buffer = data_buffer;  fld->next = 0;  add_field(fld);    BEGIN INITIAL; } . { ECHO; } %% int add_field(struct stFieldInfo *fld) {   if (field_list==0) {    field_list = fld;    last_field = fld;    return 0;   }   last_field->next = fld;   last_field = fld;   return 0;   } int print_fld_list() {  int fld_num = 0;  struct stFieldInfo *fld = field_list;  while(fld!=0) {   printf("name:%s, len=%d,prop=%d,data=%s.\n",fld->name,fld->len,fld->prop,fld->buffer);   fld_num++;   fld = fld->next;  }  printf("field_num:%d.\n",fld_num);  return 0; } int main() {  char *str_input = "f123=[3,3,abc],f23=[5,1,xxxxx]";  field_list = 0;  last_field = 0;  yy_scan_string(str_input);  yylex();  print_fld_list();  getch();  return 0; }

 

 

此示例的输出为:

name:f123, len=3,prop=3,data=abc. name:f23, len=5,prop=1,data=xxxxx. field_num:2.

 

上述代码是不可重入的。

 

修改代码以支持多线程环境使用。修改后的代码如下（FLEX 2.5.35版本）：

/* name: fld_list.y */   %option reentrant %option noyywrap %option extra-type="struct stWorkData*"   %x S_FIELDNAME %x S_EQUEL %x S_DATA_LEN %x S_DATA_LEN2 %x S_DATA_PROP %x S_DATA_PROP2 %x S_FIELDVAL %x S_END   %{   #define YY_NO_UNISTD_H #include <string.h> #include <stdlib.h>     #define MAX_FIELD_NAME_LEN  64   struct stFieldInfo {        char *name;        unsigned int len;        int prop;        char *buffer;               struct stFieldInfo *next; };   struct stWorkData {        char *fld_name; /*字段名称 */        unsigned int data_len; /* 字段值长度*/        int data_prop; /*字段属性*/        char *data_buffer; /* 字段值 */        struct stFieldInfo  *field_list;        struct stFieldInfo  *last_field; };   int add_field(struct stWorkData *wkdata,struct stFieldInfo *fld); %}   FLDNAME [A-Za-z\_]+([A-Za-z\_]|[0-9])* DELIM ([" "\t])* NUMBERS [0-9]* %% {FLDNAME} {        yyextra->fld_name = strdup(yyget_text(yyscanner));        BEGIN S_FIELDNAME; }   <S_FIELDNAME>{DELIM}"="{DELIM}\[{DELIM} {        BEGIN S_EQUEL; }   <S_EQUEL>{NUMBERS} {        yyextra->data_len = atoi(yyget_text(yyscanner));        BEGIN S_DATA_LEN; } <S_DATA_LEN>{DELIM}","{DELIM} {        BEGIN S_DATA_LEN2;       }   <S_DATA_LEN2>{NUMBERS}  {        yyextra->data_prop = atoi(yyget_text(yyscanner));        BEGIN S_DATA_PROP; }   <S_DATA_PROP>{DELIM}","{DELIM} {        BEGIN S_DATA_PROP2;     }   <S_DATA_PROP2>. {        int c;        unsigned int len = 0;        char *p;        yyextra->data_buffer = malloc(yyextra->data_len+1);        yyextra->data_buffer[0] = yyget_text(yyscanner)[0];        p = yyextra->data_buffer;        p++;        len++;        while(len<yyextra->data_len&&(c=input(yyscanner))&&c!=EOF) {               *p = c;               p++;               len++;        }        if (c==EOF||len<yyextra->data_len)               yyterminate();        yyextra->data_buffer[len] = '\0';          BEGIN S_FIELDVAL; }   <S_FIELDVAL>{DELIM}\]{DELIM}[,]? {        struct stFieldInfo *fld = malloc(sizeof(struct stFieldInfo));        fld->name = yyextra->fld_name;        fld->len = yyextra->data_len;        fld->prop = yyextra->data_prop;        fld->buffer = yyextra->data_buffer;        fld->next = 0;        add_field(yyextra,fld);               BEGIN INITIAL; } . { ECHO; }   %%   int add_field(struct stWorkData *wkdata,struct stFieldInfo *fld) {               if (wkdata->field_list==0) {                      wkdata->field_list = fld;                      wkdata->last_field = fld;                      return 0;               }               wkdata->last_field->next = fld;               wkdata->last_field = fld;                 return 0;              }   int print_fld_list(struct stWorkData *wkdata) {        int fld_num = 0;        struct stFieldInfo *fld = wkdata->field_list;        while(fld!=0) {               printf("name:%s, len=%d,prop=%d,data=%s.\n",fld->name,fld->len,fld->prop,fld->buffer);               fld_num++;                 fld = fld->next;        }        printf("field_num:%d.\n",fld_num);          return 0; }   int main() {     yyscan_t scanner;        struct stWorkData wkdata;          char *str_input = "f123=[3,3,abc],f23=[5,1,xxxxx]";          wkdata.data_len = 0;        wkdata.data_prop = 0;        wkdata.field_list = 0;        wkdata.last_field = 0;       yylex_init ( &scanner );        yylex_init_extra( &wkdata, &scanner );        yy_scan_string(str_input,scanner);       yylex ( scanner );     yylex_destroy ( scanner );          print_fld_list(&wkdata);          getch();          return 0; }

   

•   FLEX&Bison参考资料：

Lexical Analysis With Flex

http://flex.sourceforge.net/manual/

 

Bison Quick Tutorial

http://gps.nju.edu.cn:88/mediawiki/index.php/Bison_Quick_Tutorial

 

 

Flex 2.5.35 and Win32 / Visual C++

http://www.thebinaryidiot.com/archives/2011/01/29/flex-2-5-35-and-win32-visual-studio-cpp/