ANBF语法中的数值类型有3种：二进制、十进制和十六进制，可以是一个以点号分隔的数列，也可以是一个数值的范围。例如，%d11.22.33.44.55表示五个有次序的十进制数字“11、22、33、44、55”，而%x80-ff表示一个字节，这个字节的数值可以是在0x80至0xff之间。
我把以点号分隔的数列定义为NumVal，把范围类型的数值定义为RangedNumVal。这两个类实现了Element，其实我觉得应该定义一个接口NumVal（继承Element），然后一个SerialNumVal和一个RangedNumVal（实现NumVal），这样看起来更漂亮？作为一个完美主义者看到现在这个定义真是很蛋疼，有时间再重新考虑吧。
由于二进制、十进制和十六进制的构成都是很相似的，只是进制符号（b、d、x）以及数字符号（01、0123456789、0123456789abcdef）不同而已，为了避免重复地写三个很相像的方法，我投机取巧的定义了一个Matcher接口，这个接口是用来判断字符是否在预设的符号集里面的，没什么技术含量，看代码就明白了。
先来看看解析代码：
/*
This file is one of the component a Context-free Grammar Parser Generator,
which accept a piece of text as the input, and generates a parser
for the inputted context-free grammar.
Copyright (C) 2013, Junbiao Pan (Email: panjunbiao@gmail.com)
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
// bin-val = "b" 1*BIT
// [ 1*("." 1*BIT) / ("-" 1*BIT) ]
// BIT = "0" / "1"
// 二进制解析器
protected Element bin_val() throws IOException, MatchException {
// 真正的解析工作由val方法完成，只要把二进制数的符号集｛0、1｝通过Matcher实例传递给它就OK了。
return val('b', new Matcher() {
@Override
public boolean match(int value) {
// 如果符号是0或1就匹配
return value == '0' || value == '1';
}
@Override
public String expected() {
// 提示符号不在符号集内（仅用于异常情况）
return "['0', '1']";
}
});
}
// dec-val = "d" 1*DIGIT
// [ 1*("." 1*DIGIT) / ("-" 1*DIGIT) ]
protected Element dec_val() throws IOException, MatchException {
// 同上，把十进制的符号集0～9传递给val方法
return val('d', new Matcher() {
@Override
public boolean match(int value) {
// 直到写博客才发现这段代码错了，符号集不应该包含A～F的情形啊，居然单元测试已经通过了，尼玛这是什么测试质量！
// PS：单元测试代码也是我自己写的。。。
return (value >= 0x30 && value <= 0x39) || (value >= 'A' && value <= 'F') || (value >= 'a' && value <= 'f');
}
@Override
public String expected() {
// 错误代码，无语了。。。
return "['0'-'9', 'A'-'F', 'a'-'f']";
}
});
}
// hex-val = "x" 1*HEXDIG
// [ 1*("." 1*HEXDIG) / ("-" 1*HEXDIG) ]
protected Element hex_val() throws IOException, MatchException {
// 将十六进制的符号集通过Matcher实例传递给val方法进行解析
return val('x', new Matcher() {
@Override
public boolean match(int value) {
return (value >= 0x30 && value <= 0x39) || (value >= 'A' && value <= 'F') || (value >= 'a' && value <= 'f');
}
@Override
public String expected() {
return "['0'-'9', 'A'-'F', 'a'-'f']";
}
});
}
// 解析各个进制
protected Element val(char base, Matcher matcher) throws IOException, MatchException {
// 检查进制符号
assertMatch(is.peek(), base);
int baseValue = is.read();
String from = "";
String val = "";
// 进制符号之后的第一个字符，必须在Matcher定义的字符集内，否则异常
if (matcher.match(is.peek())) {
// 连续读入符合字符的字符，构成NumVal的第一个数值。
while (matcher.match(is.peek())) {
from += (char)is.read();
}
// 第一个数值后面如果是跟着点号，则是一个数列NumVal，如果是－破折号，则是一个范围型数值RangedNumVal，如果都不是，则是单一个数值
if (match(is.peek(), '.')) {
NumVal numval = new NumVal(String.valueOf((char)baseValue));
// 将刚才匹配到的数值作为第一个数值加到将要返回的NumVal中
numval.addValue(from);
// 如果后面跟着点号，则继续加入新的数值到NumVal中
while (match(is.peek(), '.')) {
int next = is.peek(1);
if (!(matcher.match(next))) {
break;
}
is.read();
val = "";
while (matcher.match(is.peek())) {
val += (char)is.read();
}
numval.addValue(val);
}
// 直到不能匹配到点号，数列结束，返回
return numval;
} else if (match(is.peek(), '-')) {
// 这里向前读取两个字符，因此即使破折号后面跟着的不是数字，也能返回单一个数字而且将破折号留给后面的分析程序
// 这是本程序里为数不多的能够具备回溯的代码段之一，嘿嘿。
int next = is.peek(1);
if (!(matcher.match(next))) {
// 如果破折号后面跟的不是数字，则破折号不读入，返回单一数值
NumVal numval = new NumVal(String.valueOf((char)baseValue));
numval.addValue(from);
return numval;
}
// 否则，破折号后面是数值，读取之，并返回RangedNumVal类型
is.read();
val ="";
val += (char)is.read();
while (matcher.match(is.peek())) {
val += (char)is.read();
}
return new RangedNumVal(String.valueOf((char)baseValue), from, val);
} else {
// 第一个数值之后跟的既不是点号，也不是破折号，则返回单一数值
NumVal numval = new NumVal(String.valueOf((char)baseValue));
numval.addValue(from);
return numval;
}
} else {
throw new MatchException(matcher.expected(), is.peek(), is.getPos(), is.getLine());
}
}
// num-val = "%" (bin-val / dec-val / hex-val)
// 解析num-val
protected Element num_val() throws IOException, MatchException {
String base = "", from ="", val ="";
// 百分号开头
assertMatch(is.peek(), '%');
is.read();
// 根据进制符号选择相应的解析方法（函数）
switch ((char)is.peek()) {
case 'b': case 'B': return bin_val();
case 'd': case 'D': return dec_val();
case 'x': case 'X': return hex_val();
default: throw new MatchException("['b', 'd', 'x']", is.peek(), is.getPos(), is.getLine());
}
}
接下来看看单元测试部分，不详细说了，其中有一句注释说明为什么上面有错误代码不能测试出来：
/*
This file is one of the component a Context-free Grammar Parser Generator,
which accept a piece of text as the input, and generates a parser
for the inputted context-free grammar.
Copyright (C) 2013, Junbiao Pan (Email: panjunbiao@gmail.com)
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
// bin-val = "b" 1*BIT
// [ 1*("." 1*BIT) / ("-" 1*BIT) ]
// BIT = "0" / "1"
// 测试二进制数的解析
@Test
public void testBin_val() throws Exception {
Tester<String> tester = new Tester<String>() {
@Override
public String test(AbnfParser parser) throws MatchException, IOException {
return parser.bin_val().toString();
}
};
String input;
input = "b1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
input = "b1010";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
input = "B1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
input = "b1.1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
input = "b0101.1111";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
input = "b0000-1111";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+".00").bin_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"-1234").bin_val().toString());
input = "b00.11.00.01.10.00.11.00.11";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).bin_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+".").bin_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"..").bin_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+".bb").bin_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"-00").bin_val().toString());
Assertion.assertMatchException("", tester, 1, 1);
Assertion.assertMatchException("b", tester, 2,1);
Assertion.assertMatchException("bg", tester, 2, 1);
Assertion.assertMatchException("b.", tester, 2, 1);
Assertion.assertMatchException("b-", tester, 2, 1);
}
// dec-val = "d" 1*DIGIT
// [ 1*("." 1*DIGIT) / ("-" 1*DIGIT) ]
// 测试十进制数的解析
@Test
public void testDec_val() throws Exception {
Tester<String> tester = new Tester<String>() {
@Override
public String test(AbnfParser parser) throws MatchException, IOException {
return parser.dec_val().toString();
}
};
String input;
input = "d1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
input = "d1234";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
input = "D1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
input = "d1.2";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
input = "d1234.5678";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
input = "d1234-5678";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+".00").dec_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"-1234").dec_val().toString());
input = "d12.34.56.78.9a.bc.de.f0";
// 看看这里，就明白为什么单元测试测不出十进制数带有a~f符号的问题了，竟然有这样错误的测试用例！！！
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).dec_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+".").dec_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"..").dec_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+".##").dec_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"-00").dec_val().toString());
Assertion.assertMatchException("", tester, 1, 1);
Assertion.assertMatchException("d", tester, 2, 1);
Assertion.assertMatchException("dg", tester, 2, 1);
Assertion.assertMatchException("d.", tester, 2, 1);
Assertion.assertMatchException("d-", tester, 2, 1);
}
// hex-val = "x" 1*HEXDIG
// [ 1*("." 1*HEXDIG) / ("-" 1*HEXDIG) ]
// 测试十六进制数的解析
@Test
public void testHex_val() throws Exception {
Tester<String> tester = new Tester<String>() {
@Override
public String test(AbnfParser parser) throws MatchException, IOException {
return parser.hex_val().toString();
}
};
String input;
input = "x1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "x1234";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "X1";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "x1.2";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "x1234.5678";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "xabcd.ef";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "xA1.2B";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
input = "x1234-abCD";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input+"-").hex_val().toString());
input = "x12.34.56.78.9a.bc.de.f0";
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input).hex_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input + ".").hex_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input + ".g0").hex_val().toString());
Assert.assertEquals("%" + input, AbnfParserFactory.newInstance(input + "-00").hex_val().toString());
Assertion.assertMatchException("", tester, 1, 1);
Assertion.assertMatchException("x", tester, 2, 1);
Assertion.assertMatchException("xg", tester, 2, 1);
Assertion.assertMatchException("x.", tester, 2, 1);
Assertion.assertMatchException("x-", tester, 2, 1);
}
// num-val = "%" (bin-val / dec-val / hex-val)
// 综合情况测试
@Test
public void testNum_val() throws Exception {
String input;
input = "%b0101";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%b0101.1010.1111";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%b0101-1111";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%d1234";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%d0123.4567.8901";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%d12345-67890";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%x0123";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%x0123.4567.89ab.CDEF";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
input = "%x0123456789-ABCDEFabcdef09";
Assert.assertEquals(input, AbnfParserFactory.newInstance(input).num_val().toString());
}
本系列文章索引：基于预测的ABNF文法分析器