程序随机的产生1000个子串(长度3-16)的情况,如果不输出结果会更快些, 不区分大小写,
对1M以内的文件测试用时基本都小于20毫秒( PM1.4G 慢是慢了点 ), 速度应该没问题...
1M个子串在理论上速度也应该差不多, 八过偶用的 lex 貌似不支持超过 32K 的状态数, 得改改.
状态表是在程序中调用 flex 2.5.33 生成的...
代码太长, 就贴匹配的一段 :
int do1 ( unsigned char* __beg , unsigned char* end )
{
int count = 0;
int curr , st_bk , yyec , action;
unsigned char *beg = __beg , *bkup , *text;
while( beg < end )
{
curr = 1;
text = beg ;
do {
yyec = *beg ? yy_ec[*beg] : 1;
if( yy_accept[curr] ) {
st_bk = curr;
bkup = beg ;
}
while( yy_chk[ yy_base[curr] + yyec ] != curr )
{
if( (curr=yy_def[curr]) >= last_dfa )
yyec = yy_meta[ yyec ];
}
curr = yy_nxt[ yy_base[curr] + yyec ];
++beg ;
} while( yy_base[curr] != jambase );
if( !( action = yy_accept[curr] ) )
{
beg = bkup; curr = st_bk;
action = yy_accept[curr];
}
if( action != rule_def )
{
++count;
#if 1
{
unsigned char xx = *beg; *beg = 0;
printf( "find %s at pos %d\n" , text , text - __beg );
*beg = xx;
}
#endif
}
}
return count;
}
$ gcc -oxff g_substr.cpp g_table.c main.c slove.c -lstdc++ && \
time xff substr -N1000 && \
time xff gentbl && \
time xff slove * > ./result/r1
0.05user 0.04system 0:00.08elapsed 107%CPU (0avgtext+0avgdata 13520maxresident)k
0inputs+0outputs (856major+0minor)pagefaults 0swaps
call flex2.5.33(flex) return 0 , rules count 1000 , xff_tbl_i.inl
0.49user 0.32system 0:01.24elapsed 65%CPU (0avgtext+0avgdata 174976maxresident)k
0inputs+0outputs (11104major+0minor)pagefaults 0swaps
process end : Debug , file size 0 , find 0 substr , time used : 0 clock
process end : ff.h , file size 627 , find 0 substr , time used : 0 clock
process end : filter.dsp , file size 4104 , find 5 substr , time used : 0 clock
process end : filter.dsw , file size 535 , find 1 substr , time used : 0 clock
open file filter.ncb failed
process end : filter.plg , file size 1241 , find 1 substr , time used : 0 clock
process end : g_substr.cpp , file size 1500 , find 5 substr , time used : 0 clock
process end : g_table.c , file size 1628 , find 0 substr , time used : 0 clock
process end : g_table.c.bak , file size 1626 , find 0 substr , time used : 0 clock
process end : gen_tbl.c , file size 1703 , find 0 substr , time used : 0 clock
process end : main.c , file size 666 , find 0 substr , time used : 0 clock
process end : result , file size 0 , find 0 substr , time used : 0 clock
process end : slove.c , file size 3926 , find 5 substr , time used : 0 clock
process end : slove.c.bak , file size 3847 , find 4 substr , time used : 0 clock
process end : test , file size 158 , find 0 substr , time used : 0 clock
process end : test.bak , file size 160 , find 0 substr , time used : 0 clock
process end : xff.exe , file size 430912 , find 45 substr , time used : 30 clock
process end : xff.s , file size 10142 , find 1000 substr , time used : 0 clock
process end : xff.t , file size 210408 , find 0 substr , time used : 0 clock
process end : xff.temp.file.lexer.script.xx , file size 32230 , find 1000 substr , time used : 10 cl
ock
process end : xff.temp.file.lexer.script.xx.bak , file size 32076 , find 30 substr , time used : 0 c
lock
process end : xff_gen_table.exe , file size 168735 , find 1 substr , time used : 10 clock
process end : xff_tbl_i.inl , file size 256185 , find 0 substr , time used : 10 clock
Command exited with non-zero status 1
0.10user 0.03system 0:00.12elapsed 100%CPU (0avgtext+0avgdata 15568maxresident)k
0inputs+0outputs (979major+0minor)pagefaults 0swaps
