top_word结构中attr的怪现像 about scws HOT 7 CLOSED

每行下方的数字为attr的长度，用get_tops时和命令行调用一致，attr长度正常
电影 n 25.0200004577637(6)
1
创作 vn 14.8199996948242(3)
2
版权 n 14.7600002288818(3)
1
专有 vn 13.9399995803833(2)
2
陈凯歌 nr 11.8699998855591(1)
2

[root@localhost ~]# scws -r /etc/scws/rules.utf8.ini -d /usr/share/scws/dict.utf8.xdb -c utf8 -I -A -t 5 -a~v -i 110.txt
No. WordString Attr Weight(times)

1. 电影 n 25.02(6)
2. 创作 vn 14.82(3)
3. 版权 n 14.76(3)
4. 专有 vn 13.94(2)
5. 陈凯歌 nr 11.87(1)

from scws.

对scws中的get_words进行修改

•                           top = (scws_top_t) malloc(sizeof(struct scws_topword));
  

•                           top = (scws_top_t) pmalloc_z(xt->p, sizeof(struct scws_topword));�
  

就出现分段错误的提示

from scws.

对cli中的scws_cmd修改，显示get_words中attr的长度，确实发现有问题，补丁文件
//--- scws-1.2.3/cli/scws_cmd.c.orig 2013-01-06 13:39:51
//+++ scws-1.2.3/cli/scws_cmd.c 2022-12-22 20:42:38
//@@ -286,6 +286,25 @@
// fprintf(fout, "EMPTY records!\n");
// }
//
//+ fprintf(fout, "No. WordString Attr Weight(times)\n");
//+ fprintf(fout, "-------------------------------------------------\n");
//+ if ((top = xtop = scws_get_words(s, attr)) != NULL)
//+ {
//+ tlimit = 1;
//+ while (xtop != NULL)
//+ {
//+ fprintf(fout, "%02d. %-24.24s %-4.2s %.2f(%d) %4d-\n",
//+ tlimit, xtop->word, xtop->attr, xtop->weight, xtop->times, strlen(xtop->attr));
//+ xtop = xtop->next;
//+ tlimit++;
//+ }
//+ scws_free_tops(top);
//+ }
//+ else
//+ {
//+ fprintf(fout, "EMPTY records!\n");
//+ }
//+
// if (xmode & XMODE_STAT_FILE)
// free(str);
// }

显示结果最后一列为attr的长度

1. 研究 vn 4.45(1) 5-
2. 生命科学 n 7.37(1) 1-
3. 北京 ns 6.35(1) 5-
4. 大学生 n 4.70(1) 1-
5. 喝 vn 0.00(1) 5-
6. 进口 vn 4.87(1) 5-
7. 红酒 n 6.17(1) 1-

from scws.

from scws.

get_tops同样是复制2字节没有这情况，不明白get_words为啥会出现超过2个字节就不等长的情况

from scws.

from scws.

strlen只能处理以'\0'结尾的字符串，xtop->attr只是一个两元素的字符数组

from scws.