Understand metacharacters and escape character

[StephenArk30]

In most languages, the escape character is '\'. Sequences start with an escape character have special meaning, eg: '\n', '\t'.

Metacharacters are escape characters, but often match a group of characters, eg: '\s' matches any whitespace like '\t', '\n'.

So, when we are matching single comment lines in C, we need to use:

\/\/[^\n]*,

instead of

\/\/[^\\n]*

which won't match a line break but a "\n" string.

[kokke]

Hi @StephenArk30 and thanks for your comment

I'm not really sure what you mean. Could you please have a look at https://github.com/kokke/tiny-regex-c/blob/master/tests/test1.c
... and write a few test cases so I'm absolutely sure what we're talking about.

I'm trying to mimic the Python implementation, but the code is in C which just handles escapes like that. Or maybe I'm misunderstanding you?

[marler8997]

@kokke what he's asking for is this:

  { OK, "\\n", "\n",  (char*) 1 },
  { OK, "\\t", "\t",  (char*) 1 },

[StephenArk30]

Sorry for the late reply. This is just a note for some beginners who thinks themselves "smart" like me. Not asking for anything.

So the case was I was trying to matching single comment lines in C like "// this is a comment", and I used

\/\/[^\\n]*

because I thought I should keep the '\'. But it didn't work so I created a pr here trying to match "\n". But after some research, I realize it should be

\/\/[^\n]*

instead. The "metacharacters" mentioned in the pr are some characters like '\s' which matches many different characters, so '\n' should not be a metacharacter.

This might be a common mistake beginners will make so I put it here. Closed it if needed.

[marler8997]

I think this would be a common mistake. The first reason being that most regex implementations support standard escape sequences like \n and \t etc. The second being the confusion that comes with including regex escape characters in C String literals which have to escape the regex escape character, \. Note that tiny-regex-c can support these escape sequences as well with a few lines of changes so it's up to @kokke as to whether it should be supported. The python implementation supports it so that's one thing to consider.