Hello

The str_strcmp function does not do a correct lexicographic comparison between the received strings:

<verbatim> /** * \brief Compare two str's case sensitive * \param str1 first str * \param str2 second str * \return 0 if both are equal, positive if str1 is greater, negative if str2 is greater, -2 on errors */ static inline int str_strcmp(const str *str1, const str *str2) { if(str1==NULL || str2==NULL || str1->s ==NULL || str2->s==NULL || str1->len<0 || str2->len<0) { LM_ERR("bad parameters\n"); return -2; }

if (str1->len < str2->len) return -1; else if (str1->len > str2->len) return 1; else return strncmp(str1->s, str2->s, str1->len); } </verbatim>

in this case str1={"z",1} will be "smaller" then str2={"aa", 2} because it's length is smaller . Method is used is various tree implementations... (same from str_casestrcmp())

Is this "normal" ?

Marius