单元测试的重要性是不言而喻的,对于ABNF的例子来说,通过单元测试除了可以发现程序的BUG之外,还可以发现预测解析器能够支持哪些情况下的文法,以及那些情况下解析器无能为力(所谓FEATURE,嘿嘿)。
我在这个项目中使用JUnit来做单元测试,先来看一段最简单的测试代码:
/*
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/>.
*/
// SP = %x20
protected String SP() throws IOException, MatchException {
assertMatch(is.peek(), 0x20);
int value = is.read();
return String.valueOf((char)value);
}
@Test
public void testSP() throws Exception {
Tester<String> tester = new Tester<String>() {
@Override
public String test(AbnfParser parser) throws MatchException, IOException {
return parser.SP();
}
};
// 测试用例1:字符串输入流为两个空格的情况(结果应该是只匹配第一个空格)
Assert.assertEquals(String.valueOf((char)0x20), AbnfParserFactory.newInstance(new char[] {0x20, 0x20}).SP());
// 测试用例2:字符串输入流为空的情况
Assertion.assertMatchException("", tester, 1, 1);
// 测试用例3:字符串输入流为回车(0x0D)的情况
Assertion.assertMatchException("" + (char)0x0D, tester, 1, 1);
}
为图方便我把被测代码和测试代码都放到一起了,testSP()函数用来测试SP()函数的功能是否满足。SP()是用来解析空格输入的函数,即当输入流的下一个字符是空格时,返回这个空格,否则抛出一个匹配异常MatchException。
单元测试的关键要素(输入参数)是测试用例(比如字符串输入流)和期望结果,在每一个测试用例中,都要调用一次被测函数,然后检查输出结果是否与预期结果一直。这里面有一部分代码是要重复使用的,为了偷懒(以免要大段大段的复制粘贴代码),抽象出Tester测试器接口和Assertion断言类。
Tester测试器接口的代码如下:
public interface Tester<E> {
public abstract E test(AbnfParser parser) throws MatchException, IOException, CollisionException;
}
Tester接口只有一个test函数。例如在测试testSP函数中,定义了一个Tester的匿名类,在这个类的test函数中调用被测函数。有了Tester接口,我们在不同的测试用例下只需要指定一个Tester的实例就行,而不需要每次都调用被测函数。
再来看看Assertion断言类,这也是一个提供方便的类,这个类提供了若干个方法,支持不同情况下调用Tester实例,并将测试结果与期望结果相比较,如果不一致则抛出异常。
/*
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/>.
*/
import junit.framework.Assert;
import java.io.IOException;
public class Assertion {
// 断言匹配
public static void assertMatch(String input, Tester tester, Object expectedOutput, int expectedPos, int expectedLine) throws MatchException, IOException, CollisionException {
// 根据字符串输入流创建一个AbnfParser实例
AbnfParser parser = AbnfParserFactory.newInstance(input);
// 调用Tester接口实例进行测试
Object output = tester.test(parser);
// 判断测试结果是否与期望结果一致
if (output == null && expectedOutput != null) Assert.fail();
if (output != null && expectedOutput == null) Assert.fail();
if (output != null && expectedOutput != null) Assert.assertEquals(expectedOutput, output);
// 判断输入流的指针位置是否与期望一致
Assert.assertEquals(expectedPos, parser.getInputStream().getPos());
Assert.assertEquals(expectedLine, parser.getInputStream().getLine());
}
public static void assertMatch(String input, Tester tester, int expectedPos, int expectedLine) throws MatchException, IOException, CollisionException {
AbnfParser parser = AbnfParserFactory.newInstance(input);
tester.test(parser);
Assert.assertEquals(expectedPos, parser.getInputStream().getPos());
Assert.assertEquals(expectedLine, parser.getInputStream().getLine());
}
// 断言要抛出匹配异常
public static void assertMatchException(String input, Tester tester, int expectedPos, int expectedLine) {
// 根据字符串输入流创建一个AbnfParser实例
AbnfParser parser = AbnfParserFactory.newInstance(input);
try {
// 调用测试函数
tester.test(parser);
// 如果执行到这里(意味着测试函数没有抛出异常),则测试不通过
Assert.fail();
} catch (MatchException me) {
// 如果捕捉到匹配异常,则继续检查输入流的指针位置是否正确
Assert.assertEquals(expectedPos, parser.getInputStream().getPos());
Assert.assertEquals(expectedLine, parser.getInputStream().getLine());
} catch (IOException e) {
// 其他异常表明测试不通过
Assert.fail();
} catch (CollisionException ce) {
// 其他异常表明测试不通过
Assert.fail();
}
}
// 断言要抛出冲突异常(流程类似匹配异常,主要针对ABNF规则名重名的情况)
public static void assertCollisionException(String input, Tester tester, int expectedPos, int expectedLine) {
AbnfParser parser = AbnfParserFactory.newInstance(input);
try {
tester.test(parser);
Assert.fail();
} catch (MatchException me) {
Assert.fail();
} catch (IOException e) {
Assert.fail();
} catch (CollisionException ce) {
Assert.assertEquals(expectedPos, parser.getInputStream().getPos());
Assert.assertEquals(expectedLine, parser.getInputStream().getLine());
}
}
}
这里再补充说明一下AbnfParserFactory工厂类,这个类主要是为了方便将字符串转换为输入流并创建AbnfParser实例而写的,代码很简单:
/*
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/>.
*/
import java.io.ByteArrayInputStream;
public class AbnfParserFactory {
public static AbnfParser newInstance(String prefix, String input) {
return new AbnfParser(prefix, new ByteArrayInputStream(input.getBytes()));
}
public static AbnfParser newInstance(String input) {
return new AbnfParser("", new ByteArrayInputStream(input.getBytes()));
}
public static AbnfParser newInstance(String prefix, char[] input) {
return new AbnfParser(prefix, new ByteArrayInputStream(String.valueOf(input).getBytes()));
}
public static AbnfParser newInstance(char[] input) {
return new AbnfParser("", new ByteArrayInputStream(String.valueOf(input).getBytes()));
}
}
有了这个工厂类,单元测试的时候就方便多了,因为我们的测试用例可能是字符串、字符数组,这个工厂类能够帮助我们快速创建AbnfParser实例。
通过单元测试可以发现很多我们意想不到的情况,例如:
// elements = alternation *c-wsp
@Test
public void testElements() throws Exception {
Tester<Elements> tester = new Tester<Elements>() {
@Override
public Elements test(AbnfParser parser) throws MatchException, IOException {
return parser.elements();
}
};
String input;
input = "A/B/C";
Assertion.assertMatch(input, tester, AbnfParserFactory.newInstance(input).elements(), 6, 1);
// TODO
input = "A/B/C ";
Assertion.assertMatchException(input, tester, 8, 1);
}
原来这个预测解析器无法处理对于输入的Alternation为“A/B/C ”即后面多一个空格的情形,具体的原因我们在后面的帖子里再慢慢分析吧。
本系列文章索引:基于预测的ABNF文法分析器