local mod_pointer prefix_lookup (char* name,int l)
/* finds module name given a prefix */
{ mod_pointer p=root,* loc=&root; /* current node and where it comes from */
mod_pointer match=NULL; /* the first matching node, if any */
mod_pointer saved=NULL; /* another subtree that might have matches */
while (p!=NULL)
{ int l0=p->key_length; char* key=name_begin(p);
switch (mod_name_cmp(name,l,key,l0))
{ case less: p=*(loc=&p->llink); break;
case greater: p=*(loc=&p->rlink); break;
case equal: return p; /* a match, and no other matches are possible */
case extension:
{ enum mod_comparison cmp=
mod_name_cmp(name+l0,l-l0,key+l0,(int)strlen(key+l0));
switch(cmp)
{ case less: case greater:
err_print("! Incompatible module name");
print("\nName inconsistently extends <%.*s...>.\n",l0,key);
return NULL;
case prefix: case equal: return p;
case extension:
if (complete_name(p))
{ err_print("! Incompatible module name");
print("\nPrefix exists: <%s>.\n",key); return NULL;
}
{
free(key-1);
if ((key=(char*)malloc(l+2))==NULL) fatal("Out of dynamic memory!");
*key++='\1'; /* ensure that |complete_name(p)| is false afterwards */
strncpy(key,name,l); key[l]='\0'; /* store the incomplete name */
name_begin(p)=key; /* install new name in node |p| */
}
return p;
}
}
case prefix:
if (match!=NULL)
{ err_print("! Ambiguous prefix"); return NULL; }
match=p; saved=p->rlink; p=p->llink; /* |loc| is irrelevant now */
}
if (p==NULL && match!=NULL)
p=saved, saved=NULL; /* search other subtree */
}
if (match==NULL)
{ char* key=(char*)malloc(l+2);
if (key==NULL) fatal("Out of dynamic memory!");
*key++='\1'; /* ensure that |complete_name(p)| is false afterwards */
strncpy(key,name,l); key[l]='\0'; /* store the incomplete name */
(p=make_mod_node(key))->key_length=l; /* prepare new node */
return *loc=p; /* install new node into tree */
}
match->key_length=l; /* |name| is a shorter prefix than used before */
return match;
}
local mod_pointer make_mod_node (char* name)
{ mod_pointer node=mod_ptr; /* allocate new node */
if (mod_ptr++>=mod_table_end) overflow ("module name");
name_begin(node)=name;
node->llink=NULL; node->rlink=NULL;
init_module_name(node); /* initialise new node */
return node;
}
local mod_pointer mod_name_lookup (char* name, int l)
/* finds complete module name */
{ mod_pointer p; /* current node of the search tree */
mod_pointer* loc=&root; /* |p| will come from this location */
while ((p=*loc)!=NULL)
{ int l0=p->key_length; char* key=name_begin(p);
switch (mod_name_cmp(name,l,key,l0))
{ case less: loc=&p->llink; break;
case greater: loc=&p->rlink; break;
case equal: case extension:
{ enum mod_comparison cmp=
mod_name_cmp(name+l0,l-l0,key+l0,(int)strlen(key+l0));
switch(cmp)
{ case less: case greater:
err_print("! Incompatible module name");
print("\nName inconsistently extends <%.*s...>.\n",l0,key);
return NULL;
case extension: case equal:
if (complete_name(p))
if (cmp==equal) return p;
else
{ err_print("! Incompatible module name");
print("\nPrefix exists: <%s>.\n",key); return NULL;
}
name_begin(p)=store_string(name,l);
/* install |name| in place of |key| */
free(key-1);
return p;
}
}
case prefix:
err_print("! Incompatible module name");
print("\nName is a prefix of <%s%s>.\n"
,key, complete_name(p) ? "" : "...");
return NULL; /* dummy module name */
}
}
{ (p=make_mod_node(store_string(name,l)))->key_length=l; /* prepare new node */
return *loc=p; /* install new node into tree */
}
}
void htmlcxx::HTML::ParserSax::parseHtmlTag(_Iterator b, _Iterator c)
{
_Iterator name_begin(b); //申请一个新的游标name_begin游标，初始为b
++name_begin;//将name_begin后移一位
bool is_end_tag = (*name_begin == '/'); //判断name_begin是否是’/’，如果是则说明是结束型tag标签
if (is_end_tag) ++name_begin;//如果是结束型的tag标签，则再将name_begin前进一位，跳过’/’符号
_Iterator name_end(name_begin); //申请一个新的name_end游标，记录一下name_begin标签。
while (name_end != c && isalnum(*name_end)) //将name_end后移，直到遇到了文档末尾或者普通字符(字母或者数字)为止。
{
++name_end;
}
std::string name(name_begin, name_end);//取name_begin和name_end之间的这段文本认为是tagName并赋值给一个string型的变量name
//DEBUGP("Found %s tag %s\n", is_end_tag ? "closing" : "opening", name.c_str());
/*
之后，我们要对这个Name做一些判断，判断其是否是特殊类型的tag标签，
即是否是前面literal_mode_elem[]数组中的某个标签。
*/
if (!is_end_tag) //首先得判断这不是一个结束型的标签
{
std::string::size_type tag_len = name.length();
for (int i = 0; literal_mode_elem[i].len; ++i)
{
if (tag_len == literal_mode_elem[i].len)
{
#if defined(WIN32) && !defined(__MINGW32__)
if (!_stricmp(name.c_str(), literal_mode_elem[i].str))
#else
if (!strcmp(name.c_str(), literal_mode_elem[i].str)) //--modified by wujun 2013.5.7 --orgi: [ if (!strcasecmp(name.c_str(), literal_mode_elem[i].str)) ]
#endif
{
mpLiteral = literal_mode_elem[i].str;
break;
}
}
}
}
//之后，我们就可以将其作为Node节点来做一些存储的工作了。
htmlcxx::HTML::Node tag_node;
//by now, length is just the size of the tag
std::string text(b, c);
tag_node.length(static_cast<unsigned int>(text.length()));
tag_node.tagName(name);
tag_node.text(text);
tag_node.offset(mCurrentOffset);
tag_node.isTag(true);
tag_node.isComment(false);
mCurrentOffset += tag_node.length();//将当前的解析位置后移
this->foundTag(tag_node, is_end_tag);//调用方法将其加入到已有的Html树中
}
id_pointer id_lookup (char* first,char* last,int ilk)
/* look up an identifier */
{ int l,h; /* length and hash code of the given identifier */
if (last==NULL) last=first+(l=(int)strlen(first)); /* null-terminated string */
else l=(int)(last-first); /* compute the length */
{ char* p=first;
h=*p; while (++p<last) h=((h<<1)+*p)%hash_size;
}
{ id_pointer p=hash[h]; /* the head of the hash list */
while (p!=NULL && !names_match(p,first,l,ilk)) p=p->hash_link;
if (p==NULL) /* we haven't seen this identifier before */
{ p=id_ptr; /* this is where the new name entry will be created */
if (id_ptr++>=id_table_end) overflow ("identifier");
name_begin(p)=store_string(first,l);
if (program==cweave) init_id_name(p,ilk);
p->hash_link=hash[h]; hash[h]=p; /* insert |p| at beginning of hash list */
}
return p;
}
}
