 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Java Forum / First Aid / March 2008Write a program that reads a Java source-code file and displays all the comments.
I am trying to do exercise 17 on page 546 of Bruce Eckel's Thinking In Java (4th edition): "Write a program that reads a Java source-code file (you provide the file name on the command line) and displays all the comments." This is at the end of a section about regular expressions. We have just learnt how to use appendReplacement(). I am having great difficulty dealing with comment-markers (// or /*) inside string literals, comments that contain quotation marks, and backslashes before quotation marks. For example: System.out.println("// This is not a comment."); System.out.println("Nor is any \" // of this."); // All of this line is a comment, including "this". I found a proposed answer at http://greggordon.org/java/tij4/solutions.htm. But it is completely useless. It does not even cope with comments using /* and */ that go over two lines. Does anyone have a solution to this exercise? or any hints? (I would also like solutions to the following two exercises: write a program that reads a Java source-code file and displays all the string literals; and write a program that examines Java source code and produces all the class names used in a particular program.) > I am trying to do exercise 17 on page 546 of Bruce Eckel's Thinking In > Java (4th edition): [quoted text clipped - 4 lines] > > Does anyone have a solution to this exercise? or any hints? The way I would do it would be to create a Reader on the file, read each character and perform simple lexical analysis there, like so: boolean inEOLComment = false, inCComment = false, inString = false; for each character in stream: if inEOLComment: print character if character is newline, inEOLComment = false else if inCComment: if character is * and next is /, inCComment = false else print character else if inString: if character is \, skip next character else if character is ", inString = false else if character is /: if next character is /, inEOLComment = true else if next character is *, inCComment = true else if character is ", inString = true else, do nothing (writing the actual Java code is left as an exercise to the reader) > (I would also like solutions to the following two exercises: write a > program that reads a Java source-code file and displays all the string > literals; and write a program that examines Java source code and > produces all the class names used in a particular program.) The first should be a trivial modification of the previous code, and the latter requires some more complex modification. Look up lexical analysis and parsing for more details. Note: The code I provided does not provide for preprocessing of Unicode escapes. If you need to handle that, the easiest way would be to wrap an input stream.  SignatureBeware of bugs in the above code; I have only proved it correct, not tried it. -- Donald E. Knuth anon36@yahoo.com wrote: >> I am trying to do exercise 17 on page 546 of Bruce Eckel's Thinking In >> Java (4th edition): [quoted text clipped - 4 lines] >> >> Does anyone have a solution to this exercise? or any hints? > The way I would do it would be to create a Reader on the file, read each > character and perform simple lexical analysis there, like so: [quoted text clipped - 17 lines] > > (writing the actual Java code is left as an exercise to the reader) Another approach is to borrow from the LEX / YACC approach, and have a "lexer" extract tokens from the input, along with a token-type enum identifying it as "String literal", "identifier/keyword", "punctuation", etc. The output of the lexer becomes the input to the parser, which examines each token and its identifier, and operates a state machine with, say, states of IN_COMMENT and NOT_IN_COMMENT. You run the parser in a loop, with the interpretation of each token depending on the current state. So, for example, if you hit IN_SINGLE_LINE_COMMENT state, you ignore each token up until the LINE_END token. If you reach IN_MULTI_LINE_COMMENT state, you ignore each token until you reach the END_COMMENT token ("*/"). Strings inside the comment do not trigger a false END_COMMENT because you've already lexed such strings into tokens. The parser will not see an string-embedded "*/", it'll only see a STRING_LITERAL token. SignatureLew > I am trying to do exercise 17 on page 546 of Bruce Eckel's Thinking > In Java (4th edition): [quoted text clipped - 6 lines] > inside string literals, comments that contain quotation marks, and > backslashes before quotation marks. One option is to use a state machine that knows when the current file position is within a comment or string and when the next character has been escaped with a backslash. There are many web pages and Usenet posts on this technique. Regards, Patrick ------------------------------------------------------------------------ S P Engineering, Inc. | Large scale, mission-critical, distributed OO | systems design and implementation. pjm@spe.com | (C++, Java, Common Lisp, Jini, middleware, SOA) Sat, 23 Feb 2008 08:50:14 -0800 (PST), /anon36@yahoo.com/: > I am trying to do exercise 17 on page 546 of Bruce Eckel's Thinking In > Java (4th edition): [quoted text clipped - 3 lines] > just learnt how to use appendReplacement(). > [...] You may take a look at the java.io.StreamTokenizer [1] class which will break the source text into tokens (comments, string literals, etc.) properly. Then you could build your parsing logic on top of it. [1] http://java.sun.com/j2se/1.5.0/docs/api/java/io/StreamTokenizer.html SignatureStanimir Sun, 24 Feb 2008 00:23:25 +0200, /Stanimir Stamenkov/: > Sat, 23 Feb 2008 08:50:14 -0800 (PST), /anon36@yahoo.com/: > [quoted text clipped - 11 lines] > > [1] http://java.sun.com/j2se/1.5.0/docs/api/java/io/StreamTokenizer.html Just tried the StreamTokenizer won't help you with the comments as it just discards them when it recognizes them. SignatureStanimir >"Write a program that reads a Java source-code file (you provide the >file name on the command line) and displays all the comments." >This is at the end of a section about regular expressions. We have >just learnt how to use appendReplacement(). see http://mindprod.com/jgloss/finitestate.html http://mindprod.com/jgloss/parser.html -- Roedy Green Canadian Mind Products The Java Glossary http://mindprod.com > I am trying to do exercise 17 on page 546 of Bruce Eckel's Thinking In > Java (4th edition): [quoted text clipped - 21 lines] > literals; and write a program that examines Java source code and > produces all the class names used in a particular program.) We can't give you an outright solution since it is likely to be a homework problem. But, we can give hints. You might have to assume the Java in the source file is syntactically correct. If you don't get a clean compile, you could conceivably have a source code file in which you (a human) can't tell where a string literal ends and a comment begins. But, I digress ... It might help to have read the file character-by-character. You could have some support methods, such as "readString" that does nothing but read characters until it reaches the end of a string. You could have one or more other methods that do nothing but read and print characters until the end of a comment is reached. Maybe "readCommentA", "readCommentB" to account for different types of comments. Thank you all. So it seems that the simplest way is just to parse it one character at a time. I did it like Joshua suggested, and it works fine. For what it is worth, I have a feeling that Eckel didn't intend this. For one thing, the preceding text in the book is about replace operations and appendReplacement(StringBuffer sbuf, String replacement). I am fairly sure the exercises are intended to use appendReplacement. What's more, like the last person said, I think you have to assume the Java source file is syntactically correct, otherwise it might be actually impossible to do. I think the author probably intended readers to assume that there are no //s inside quotation marks, no quotation marks inside comments, etc. ... which would make it all a lot easier. But he didn't say that. Thanks anyway! Free MagazinesGet these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...
|
Reply to this Message
|
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2008 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.