3.3.5-3

[fool2fish]

3.3.5-3

[fool2fish]

From the question:

without an intervening */, unless it is inside double-quotes

so, this means:

\/\*([^*"]*|".*"|\*+[^/])*\*\/
             ^^^^
             any comment in quotes is regarded as normal text

[misaka-10032]

That's what I mean. That RegExp can handle /"xxx"/ , but cannot handle /"xx/ or "/xx/" . In fact, the first and second should be regarded as comments, but the third shouldn't, right?

[fool2fish]

This answer do implies that: quotes in comment must be closed. Follow your thought, a stricter answer could be (in JavaScript，and you can ignore the space):

\/\*  (  [^*"]*  |  ".*"  |  "[^"]*?(?=\*\/)  |  \*+[^/]  )  *\*\/

I hope you don't intend to ask me: how about /*"*/\"*/, an such like.

Actually, in practice the comment rule is often more simple:

\/\*[\s\S]*?\*\/

But the third is another matter, it should be considered as a string rather than a comment. We can get some lex rules below:

\/\/.*$              {return COMMENT}
\/\*[\s\S]*?\*\/   {return COMMENT}
"(\\"|[^"])*"       {return STRING}

[misaka-10032]

I would prefer \/\*[\s\S]*?\*\/ :D. I get your solution for "comments in quotes" then. Actually, what I want to ask is, if it is beyond the ability of regExp to describe the pattern "comments not in quotes" in a single line? And we can only resort to other approaches to solve the problem?

p.s. I am a beginner, and your patient explanation does mean a lot of help to me. Thank you very much. :-)

[fool2fish]

It is completely normal that a lex analyser comprise several lex rules(similar to regExp, and there is an rule example, plus a hand made lex analyser example). Thus, if you want to recognize both comment and string, you will have to use two(or three) rules. But if you only want to recognize comment, then "/*...*/" is illegal.

I'm not sure if I explained clearly, you'd better try to write a lex analyser by hand that only recognize comment and string, it may give you some stimulation.

[misaka-10032]

Got it! Thx :-)