apibuild.py

libxml,在UNIX/LINUX下非常重要的一个库,为XML相关应用提供方便.目前上载的是最新版本,若要取得最新版本,请参考里面的readme.

		    l = string.strip(l)
		    retdesc = retdesc + " " + l
		    del lines[0]
	    else:
	        desc = desc + " " + l
		del lines[0]
		     
	retdesc = string.strip(retdesc)
	desc = string.strip(desc)

	if quiet == 0:
	     #
	     # report missing comments
	     #
	    i = 0
	    while i < nbargs:
	        if args[i][2] == None and args[i][0] != "void" and args[i][1] != None:
		    self.warning("Function comment for %s lacks description of arg %s" % (name, args[i][1]))
		i = i + 1
	    if retdesc == "" and ret[0] != "void":
		self.warning("Function comment for %s lacks description of return value" % (name))
	    if desc == "":
	        self.warning("Function comment for %s lacks description of the function" % (name))


	return(((ret[0], retdesc), args, desc))

    def parsePreproc(self, token):
	if debug:
	    print "=> preproc ", token, self.lexer.tokens
        name = token[1]
	if name == "#include":
	    token = self.lexer.token()
	    if token == None:
	        return None
	    if token[0] == 'preproc':
		self.index_add(token[1], self.filename, not self.is_header,
		                "include")
		return self.lexer.token()
	    return token
	if name == "#define":
	    token = self.lexer.token()
	    if token == None:
	        return None
	    if token[0] == 'preproc':
	         # TODO macros with arguments
		name = token[1]
	        lst = []
		token = self.lexer.token()
		while token != None and token[0] == 'preproc' and \
		      token[1][0] != '#':
		    lst.append(token[1])
		    token = self.lexer.token()
                try:
		    name = string.split(name, '(') [0]
                except:
                    pass
                info = self.parseMacroComment(name, not self.is_header)
		self.index_add(name, self.filename, not self.is_header,
		                "macro", info)
		return token

	#
	# Processing of conditionals modified by Bill 1/1/05
	#
	# We process conditionals (i.e. tokens from #ifdef, #ifndef,
	# #if, #else and #endif) for headers and mainline code,
	# store the ones from the header in libxml2-api.xml, and later
	# (in the routine merge_public) verify that the two (header and
	# mainline code) agree.
	#
	# There is a small problem with processing the headers. Some of
	# the variables are not concerned with enabling / disabling of
	# library functions (e.g. '__XML_PARSER_H__'), and we don't want
	# them to be included in libxml2-api.xml, or involved in
	# the check between the header and the mainline code.  To
	# accomplish this, we ignore any conditional which doesn't include
	# the string 'ENABLED'
	#
	if name == "#ifdef":
	    apstr = self.lexer.tokens[0][1]
	    try:
	        self.defines.append(apstr)
		if string.find(apstr, 'ENABLED') != -1:
		    self.conditionals.append("defined(%s)" % apstr)    
	    except:
	        pass
	elif name == "#ifndef":
	    apstr = self.lexer.tokens[0][1]
	    try:
	        self.defines.append(apstr)
		if string.find(apstr, 'ENABLED') != -1:
		    self.conditionals.append("!defined(%s)" % apstr)    
	    except:
	        pass
	elif name == "#if":
	    apstr = ""
	    for tok in self.lexer.tokens:
	        if apstr != "":
		    apstr = apstr + " "
	        apstr = apstr + tok[1]
	    try:
	        self.defines.append(apstr)
		if string.find(apstr, 'ENABLED') != -1:
		    self.conditionals.append(apstr)
	    except:
	        pass
	elif name == "#else":
	    if self.conditionals != [] and \
	       string.find(self.defines[-1], 'ENABLED') != -1:
	        self.conditionals[-1] = "!(%s)" % self.conditionals[-1]
	elif name == "#endif":
	    if self.conditionals != [] and \
	       string.find(self.defines[-1], 'ENABLED') != -1:
	        self.conditionals = self.conditionals[:-1]
	    self.defines = self.defines[:-1]
	token = self.lexer.token()
	while token != None and token[0] == 'preproc' and \
	    token[1][0] != '#':
	    token = self.lexer.token()
	return token

     #
     # token acquisition on top of the lexer, it handle internally
     # preprocessor and comments since they are logically not part of
     # the program structure.
     #
    def token(self):
        global ignored_words

        token = self.lexer.token()
	while token != None:
	    if token[0] == 'comment':
		token = self.parseComment(token)
		continue
	    elif token[0] == 'preproc':
		token = self.parsePreproc(token)
		continue
	    elif token[0] == "name" and ignored_words.has_key(token[1]):
	        (n, info) = ignored_words[token[1]]
		i = 0
		while i < n:
		    token = self.lexer.token()
		    i = i + 1
		token = self.lexer.token()
		continue
	    else:
	        if debug:
		    print "=> ", token
	        return token
	return None

     #
     # Parse a typedef, it records the type and its name.
     #
    def parseTypedef(self, token):
        if token == None:
	    return None
	token = self.parseType(token)
	if token == None:
	    self.error("parsing typedef")
	    return None
	base_type = self.type
	type = base_type
	 #self.debug("end typedef type", token)
	while token != None:
	    if token[0] == "name":
		name = token[1]
		signature = self.signature
		if signature != None:
		    type = string.split(type, '(')[0]
		    d = self.mergeFunctionComment(name,
			    ((type, None), signature), 1)
		    self.index_add(name, self.filename, not self.is_header,
				    "functype", d)
		else:
		    if base_type == "struct":
			self.index_add(name, self.filename, not self.is_header,
					"struct", type)
			base_type = "struct " + name
	            else:
			self.index_add(name, self.filename, not self.is_header,
		                    "typedef", type)
		token = self.token()
	    else:
		self.error("parsing typedef: expecting a name")
		return token
	     #self.debug("end typedef", token)
	    if token != None and token[0] == 'sep' and token[1] == ',':
	        type = base_type
	        token = self.token()
		while token != None and token[0] == "op":
		    type = type + token[1]
		    token = self.token()
	    elif token != None and token[0] == 'sep' and token[1] == ';':
	        break;
	    elif token != None and token[0] == 'name':
	        type = base_type
	        continue;
	    else:
		self.error("parsing typedef: expecting ';'", token)
		return token
	token = self.token()
	return token
	     
     #
     # Parse a C code block, used for functions it parse till
     # the balancing } included
     #
    def parseBlock(self, token):
        while token != None:
	    if token[0] == "sep" and token[1] == "{":
	        token = self.token()
		token = self.parseBlock(token)
	    elif token[0] == "sep" and token[1] == "}":
	        self.comment = None
	        token = self.token()
		return token
	    else:
	        if self.collect_ref == 1:
		    oldtok = token
		    token = self.token()
		    if oldtok[0] == "name" and oldtok[1][0:3] == "xml":
		        if token[0] == "sep" and token[1] == "(":
			    self.index_add_ref(oldtok[1], self.filename,
			                        0, "function")
			    token = self.token()
			elif token[0] == "name":
			    token = self.token()
			    if token[0] == "sep" and (token[1] == ";" or
			       token[1] == "," or token[1] == "="):
				self.index_add_ref(oldtok[1], self.filename,
						    0, "type")
		    elif oldtok[0] == "name" and oldtok[1][0:4] == "XML_":
			self.index_add_ref(oldtok[1], self.filename,
					    0, "typedef")
		    elif oldtok[0] == "name" and oldtok[1][0:7] == "LIBXML_":
			self.index_add_ref(oldtok[1], self.filename,
					    0, "typedef")
			 
		else:
		    token = self.token()
	return token

     #
     # Parse a C struct definition till the balancing }
     #
    def parseStruct(self, token):
        fields = []
	 #self.debug("start parseStruct", token)
        while token != None:
	    if token[0] == "sep" and token[1] == "{":
	        token = self.token()
		token = self.parseTypeBlock(token)
	    elif token[0] == "sep" and token[1] == "}":
		self.struct_fields = fields
		 #self.debug("end parseStruct", token)
		 #print fields
	        token = self.token()
		return token
	    else:
	        base_type = self.type
		 #self.debug("before parseType", token)
		token = self.parseType(token)
		 #self.debug("after parseType", token)
		if token != None and token[0] == "name":
		    fname = token[1]
		    token = self.token()
		    if token[0] == "sep" and token[1] == ";":
		        self.comment = None
		        token = self.token()
			fields.append((self.type, fname, self.comment))
			self.comment = None
		    else:
		        self.error("parseStruct: expecting ;", token)
		elif token != None and token[0] == "sep" and token[1] == "{":
		    token = self.token()
		    token = self.parseTypeBlock(token)
		    if token != None and token[0] == "name":
			token = self.token()
		    if token != None and token[0] == "sep" and token[1] == ";":
			token = self.token()
		    else:
		        self.error("parseStruct: expecting ;", token)
		else:
		    self.error("parseStruct: name", token)
		    token = self.token()
		self.type = base_type;
        self.struct_fields = fields
	 #self.debug("end parseStruct", token)
	 #print fields
	return token

     #
     # Parse a C enum block, parse till the balancing }
     #
    def parseEnumBlock(self, token):
        self.enums = []
	name = None
	self.comment = None
	comment = ""
	value = "0"
        while token != None:
	    if token[0] == "sep" and token[1] == "{":
	        token = self.token()
		token = self.parseTypeBlock(token)
	    elif token[0] == "sep" and token[1] == "}":
		if name != None:
		    if self.comment != None:
			comment = self.comment
			self.comment = None
		    self.enums.append((name, value, comment))
	        token = self.token()
		return token
	    elif token[0] == "name":
		    if name != None:
			if self.comment != None:
			    comment = string.strip(self.comment)
			    self.comment = None
			self.enums.append((name, value, comment))
		    name = token[1]
		    comment = ""
		    token = self.token()
		    if token[0] == "op" and token[1][0] == "=":
		        value = ""
		        if len(token[1]) > 1:
			    value = token[1][1:]
		        token = self.token()
		        while token[0] != "sep" or (token[1] != ',' and
			      token[1] != '}'):
			    value = value + token[1]
			    token = self.token()
		    else:
		        try:
			    value = "%d" % (int(value) + 1)
			except:
			    self.warning("Failed to compute value of enum %s" % (name))
			    value=""
		    if token[0] == "sep" and token[1] == ",":
			token = self.token()
	    else:
	        token = self.token()
	return token

     #
     # Parse a C definition block, used for structs it parse till
     # the balancing }
     #
    def parseTypeBlock(self, token):
        while token != None:
	    if token[0] == "sep" and token[1] == "{":
	        token = self.token()
		token = self.parseTypeBlock(token)
	    elif token[0] == "sep" and token[1] == "}":
	        token = self.token()
		return token
	    else:
	        token = self.token()
	return token

     #
     # Parse a type: the fact that the type name can either occur after
     #    the definition or within the definition makes it a little harder
     #    if inside, the name token is pushed back before returning
     #
    def parseType(self, token):
        self.type = ""
	self.struct_fields = []
        self.signature = None
	if token == None:
	    return token

	while token[0] == "name" and ( 
	      token[1] == "const" or token[1] == "unsigned"):
	    if self.type == "":
	        self.type = token[1]
	    else:
	        self.type = self.type + " " + token[1]
	    token = self.token()

        if token[0] == "name" and (token[1] == "long" or token[1] == "short"):
	    if self.type == "":
	        self.type = token[1]
	    else:
	        self.type = self.type + " " + token[1]
	    if token[0] == "name" and token[1] == "int":
		if self.type == "":
		    self.type = tmp[1]
		else:
		    self.type = self.type + " " + tmp[1]
	     
        elif token[0] == "name" and token[1] == "struct":
	    if self.type == "":
	        self.type = token[1]
	    else:
	        self.type = self.type + " " + token[1]
	    token = self.token()
	    nametok = None
	    if token[0] == "name":
	        nametok = token
		token = self.token()
	    if token != None and token[0] == "sep" and token[1] == "{":
		token = self.token()
		token = self.parseStruct(token)
	    elif token != None and token[0] == "op" and token[1] == "*":
	        self.type = self.type + " " + nametok[1] + " *"
		token = self.token()
		while token != None and token[0] == "op" and token[1] == "*":
		    self.type = self.type + " *"
		    token = self.token()
		if token[0] == "name":