Flex string recognition "unrecognised rule" error - c

Question

Im trying to create a string recognition rule to run in flex,the string can consist of escape characters(\n , \t , \r , \ , " , '), symbols( -, +, *, /, :, _, $, !, #, #, &, ~, ^, (, ) ) and a-zA-Z0-9 characters,i have tried many variations of the code below,but i keep getting the same error mentioned above.
ESCAPECHAR [\n] | [\t] | [\r] | [\] | ['] | ["]
SYMBOLS [-+*/:_$!##&~^()]
CHARACTERS [0-9a-zA-Z]
STRING ("({ESCAPECHAR} | {SYMBOLS} | {CHARACTERS})*") | ('({ESCAPECHAR} | {SYMBOLS} | {CHARACTERS})*')

Answer 1

You would do well to read the Flex manual chapter on patterns syntax. It is not very long, and it gives a complete description of the syntax of Flex patterns.
Here are a few of the errors you have made:
Flex patterns cannot include unquoted whitespace (unless you put them inside of a subexpression marked with the x flag). So
[\n] | [\t] | [\r] | [\] | ['] | ["]
is invalid.
Also, the \ is used to indicate that:
the following letter is a code for a control character (so that \n is a newline character), or
the following punctuation symbol should not be given special significance.
So in [\], the \ indicates that the following ] should be treated as an ordinary character, instead of being the end of a character class, which means that the character class will continue up to the next ]. Space characters inside a character class are considered to be quoted, so the character class consists of the characters ], space, |, [ and '. (Flex lets you repeat characters inside a character class, so it won't complain about the fact that there are two space characters.) You probably meant [\\].
Anyway, you should write character classes in the same way you wrote the other character classes, as a series of characters or escaped codes inside [ and ]:
[\n\t\r\\ '"]
Flex lets you quote characters by surrounding them with quotation marks, so that `"({ESCAPECHAR} | {SYMBOLS} | {CHARACTERS})*" is treated as a single literal string, which must be matched literally in the text. You probably intended the quotation marks to be ordinary characters, so you should have escaped them or put them into a single-character character class:
["]({ESCAPECHAR}|{SYMBOLS}|{CHARACTERS})*["]
Again, it is necessary to remove the whitespace from the pattern.
I assume that your intention was to allow "escape characters" to appear in a string only if they are actually escaped. Your {ESCAPECHAR} macro expands to a collection of actual characters, so that it includes newline, tab and carriage return characters. It also includes quote and apostrophe, which really should be reserved for terminating the string literal. Probably, what you meant was to allow escape codes if they are preceded with a \ (as with C or, as mentioned above, flex itself). In that case, what you really need to write is
ESCAPECHAR \\[ntr'"]
(That is, a \\, followed by exactly on of the characters n, t, r, ', ".) Even that is not precise, though: It does not allow the use of \\ to indicate a single \, and it forces the user to write "Don\'t just copy code." and '\"', both of which would normally be written without the backslash escapes.